This commit adds a node that fills every element of an attribute
with the same value. Currently it supports float, vector, and color
attributes. An immediate use case is for "billboard" scattering.
Currently people are using the same input to a Random Attribute node's
min and max input to fill every element of a vector with the same value,
which is an unintuitive way to accomplish the same thing.
Differential Revision: https://developer.blender.org/D9790
Since the initial merge of the geometry nodes project, the modifyPointCloud
function already was already modifying a geometry set. The function wasn't
renamed back then, because then the merge would have touched many
more files.
Ref T83357.
Caused by rB4212b6528af.
outlineColor is computed by the vertex shader, so not a uniform.
So outlineColor was undefined.
note: it was still possible to run into the situation that a selected UV
is drawn ontop of a selected pinned UV [you had to disable sticky
selection for this], now also make sure selected-pinned are drawn
topmost, then selected, then unselected UVs.
Maniphest Tasks: T83361
Differential Revision: https://developer.blender.org/D9786
Previously, the transformation of recursive instances did not work as
on would expect. Second-level instances would detach from first-level
instances when the object was moved.
This was returning an empty allocated string, however almost
all callers checked if the return value was NULL before freeing,
making for misunderstandings on the intended use of this function.
BCAnimationSampler::initialize_curves for example detected the
f-curves animation type to 'bone' based on a non-NULL return value
which never failed.
Also fixes two leaks where the the result of BLI_str_quoted_substrN
wasn't freed.
BLI_str_quoted_substrN didn't unescape the resulting string,
yet all callers were using this with animation paths which are
created using BLI_str_escape().
- Fix BLI_str_quoted_substrN use with escaped strings by calling
BLI_str_unescape().
Note that it's possible we want an a version of this function that
keeps escape characters. This could be added if it's required.
- Fix end quote detection using BLI_str_escape_find_quote
checking for `\"` isn't reliable since an even number of back-slashes
before a quote means it's not escaped.
The issue is caused by stale data on the Mesh Node which is not cleared
during synchronizing since the socket API refactor so that we can detect
changes. However, synchronization only updates the sockets of the Mesh,
so other properties were left with outdated values.
This caused an underflow when computing attribute size for undisplaced
coordinates as it was using the current number of vertices minus the
previous count of subdivision vertices, which at this point should be 0.
Added a simple method to clear non socket data. Also modified
Mesh.add_undisplaced to always use an ATTR_PRIM_GEOMETRY as the data is
not subdivided yet and it avoids any further issues regarding computing
attribute sizes.
Testing equality of vectors can't be done properly if they are implicitly
converted to float before-hand. This adds a lot of boilerplate code to
the node, maybe that can made more elegant in the future, but it's not
so complicated anyway.
For the attribute comparison node, it makes sense to read both input
attributes implicitly converted to the same data type. But the choice
of which of the two data types to read with is somewhat arbitrary.
This commit adds a utility function to choose the least "complex" data
type as a lowest common denominator, to make that choice less arbitrary
, and to support other needs in the future.
With the previous helper function in this file you had to know the data
type you wanted for the input attribute before requesting it. This adds
another utility function for getting the data type without building the
attribute. The attribute's can be requested later after doing some
logic to figure out the right data type to read.
No functional changes in this commit.
This adds deformation types to snake hook and the elastic deformation
type. This mode deforms the mesh using a kelvinlet instead of applying
the displacement directly inside the brush radius, which is great for
stylized shapes sketching.
Changes in rake rotation when using elastic are too strong when set
to 1, so I'll add a nicer way to support rake rotations with smoother
transitions in the future.
Reviewed By: sergey, JulienKaspar
Differential Revision: https://developer.blender.org/D9560
relation immediately
Texture and ParticleSettings have a DEG relation, but
`DEG_relations_tag_update` was not called when the texture changed.
This lead to no updates when e.g. texture size changes, relation only
went into full effect after save/reload or adding/removing keyframes.
Two places were additional relation tagging is needed:
- ParticleSettings `active_texture` changes
- ParticleSettingsTextureSlot (basically any TextureSlots') texture
changes
Maniphest Tasks: T82242
Differential Revision: https://developer.blender.org/D9393
The change to match socket color and category was already done, but it was missing
versioning code to update the theme on load.
Fixes T83500 (already closed as invalid, but this would solve the non-matching colors)
Reviewed by Hans Goudey (HooglyBoogly)
The SVM AO node calls "scene_intersect_local" with a NULL pointer for the intersection
information, which caused a crash with OptiX since it was not checking for this case and
always dereferencing this pointer. This fixes that by checking whether any hit information
was requested first (like is done in the BVH2 intersection routines).
This is a non-functional change. The functionality introduced in this commit
is not used in master yet. It is used by nodes that are being developed in
other branches though.
Now the gizmo is drawn only when the eventstate located in
`wm->winactive->eventstate` has not changed.
So it doesn't matter if it's "selected" or not.
This commit also removes the use of the private header "wm.h"
Reviewed By: campbellbarton
Differential Revision: https://developer.blender.org/D9539
This is a workaround for T80804.
There's a startup crash that happens on 2.91.0 on Windows, an `EXCEPTION_ACCESS_VIOLATION`
on `atio6axx.dll`. It is triggered by `glClear` on the `detect_mip_render_workaround`
function. The workaround moves the function after the device/driver workaround section and
sets the flag to the affected one to avoid running the check.
It is deprecated hardware that has not meet the minimum requirements since 2.79, but is
still usable and this extends its usability a bit before the cards are finally blacklisted.
Reviewed By: Jeroen Bakker
Differential Revision: https://developer.blender.org/D9667
This patch contains 2 scripts that will help with LTS releases
create_download_urls.py
This python script is used to generate the download urls which we can
copy-paste directly into the CMS of www.blender.org.
Usage: create_download_urls.py --version 2.83.7
Arguments:
--version VERSION Version string in the form of {major}.{minor}.{build}
(eg 2.83.7)
The resulting html will be printed to the console.
create_release_notes.py
=======================
This python script is used to generate the release notes which we can
copy-paste directly into the CMS of www.blender.org and stores.
Usage: ./create_release_notes.py --task=T77348 --version=2.83.7
Arguments:
--version VERSION Version string in the form of {major}.{minor}.{build}
(e.g. 2.83.7)
--task TASK Phabricator ticket that is contains the release notes
information (e.g. T77348)
--format FORMAT Format the result in `text`, `steam`, `wiki` or `html`
Requirements
============
* Python 3.8 or later
* Python phabricator client version 0.7.0
https://pypi.org/project/phabricator/
For convenience the python modules can be installed using pip.
pip3 install -r ./requirements.txt
Differential Revision: https://developer.blender.org/D9055
This reverts commit cbae82ba96.
This change introduced the following problems:
- We could no longer reliably duplicate or use an existing
custom property names.
- We could no longer assume a bone or ID name
can be used in a custom-property.
- Importers that support custom properties (such as FBX)
could fail with an exception creating custom properties.
This was added when Linux's audio support
often needed to be manually configured.
Further 3rd party libraries have their own environment variables & docs
these need not be part of Blender's help message unless they're likely
to be needed to properly run Blender.
Implement improvement from T73139 for merging along edges.
It is now called "Connected" mode, while the default is called "All".
With the recent performance improvement, the Connected Mode is in some
cases only double the speed than the usual merge all strategy but in
other cases it may be even faster. The bottleneck is somewhere further
down the line of merging geometry.
The motivation for this patch came from T80897, because the merging in
complex solidify is making it very slow.
Now merging can be removed from solidify without greater consequences,
as this is just a quicker and more advanced algorithm to do the same
thing that solidify currently does slowly.
Reviewed by: mano-wii, campbellbarton
Ref D8966
When opening a temporary File Browser, we have to make sure the file selection
parameters are refreshed. When opening it in a new Window that would always be
the case, if the File Browser uses a maximized window (as set in the
Preferences), it might reuse space-data from a previous use. So we have to
force the refresh.
Also renamed the relevant function to be more clear about what it's doing.
Mistake in 95b3c4c966.
This just makes sure that the mesh coming out of the subdivision
surface node has correct normals. Ideally, we would lazily compute
derived data like normals in more cases, but that can be solved later.
Correctness is more important right now.
In order to solve this better, it would be nice if functions like
`BKE_mesh_ensure_normals` would take a `const Mesh *`.
The mesh could be considered to be logically const, because
normals are derived data that is only cached for the current
state of the mesh.
Buffer strokes weren't being the excluded from depth only draw calls
so were being included in depth tests. They are now excluded by
bypassing the creation of the buffer strokes.
Reviewed By: fclem
Differential Revision: https://developer.blender.org/D9742
OptiX support is not in fact experimental anymore, so it is time for that notice to go.
All Cycles features that are currently supported on the GPU do work now when OptiX is selected.
This enables support for baking when OptiX is active, but uses CUDA for that behind the scenes, since
the way baking is currently implemented does not work well with OptiX.
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D9784
* Avoid direct access to `SpaceFile.params`, use a getter instead. This matters
because once the asset-browser changes are in, there will be an alternative
selection parameter object. The getter can return the correct one.
* Rename the function to ensure the parameters. The old name
`ED_fileselect_get_params()` wasn't a mere getter, it would create the
parameters if necessary. Now we have an actual getter, so better be clear.
* In some instances, I replaced the old "get" function with the new mere
getter. So the ensure logic is called less often. However, in these cases we
should be able to assume the selection parameters were created already as
part of the editor creation routine.
The term "active" in the new function names may seem a bit odd in the current
context, but that is a preparation for the Asset Browser merge as well. Like
said, there will be two file selection parameter objects in the space.
Caused problems in the Asset Browser branch when passing an empty path. The
function shouldn't blindly add a slash but sanity-check the input parameters
first.
Also makes NLA tracks and strips overridable.
User can either edit existing strips in existing NLA tracks (but not add or remove them), and/or add new NLA tracks after those comming from the linked data.
Most of the work was as usual checking operators and adding protections against illegal operations in override context.
Note that since we can only rely on indices to deal with local added tracks, we forbid any local track being before any linked/original track.
Maniphest Tasks: T72629
Differential Revision: https://developer.blender.org/D9611
This patch expands on the `is_property_set` doc-string,
detailing the purpose of the `ghost` argument added in
d3bcbe10c2.
Reviewed By: sybren
Ref D9780
It's easier to read and less 'weird' to check that an item is non-local
in a liboverride data-block, than the other way around. Thanks to
@sybren for noticing it.
Updating Industry Compatible keymap for Property filtering and new Modifier behavior.
Differential Revision: https://developer.blender.org/D9765
Reviewed by Julian Eisel
Small improvements to the layout of the Tooltips section in Preferences.
Differential Revision: https://developer.blender.org/D9772
Reviewed by Pablo Vazquez
The transform modes `shrinkfatten` and `seq_slide` have a special way of
handling events.
They use modal events in a different way than expected.
Therefore, this commit adds special event handles for these modes and
removes the keymodal tips from the status bar.
These effects are already described in the header anyway.
If the stroke had zero points the pointer returned NULL.
Also replaced for loop by FOREACH macro.
This is a corner case of "empty" strokes without points.
Change the default for the Show Channel Group Colors preference to OFF.
This option in the user preferences was introduced in rBad85256e7108. It
moved a per-file-per-editor option to the user preferences. As this
option would be frequently turned off by animators, this would now have
to happen only once. This commit takes this one step further, and turns
it off by default, as it can cause major readability issues of the
animation channel list.
RNA collections that support insertion of new items in liboverride
data-block need a special way to distiguish between locale and
orig-from-linked items (since some operations are allowed on the forer,
but no the latter).
In future we want a proper solution to abstract that at the
`BKE_lib_override` level, but for now we need to add some code for each
case.
Note that this commit also fixes a few potential issues with GPencil
modifiers, and constraints, regarding their handling of local overrides.
Fix the Assign Automatic Weights operator in weight paint mode when the
Vertex Groups X option is enabled. This issue was probably introduced in
rB5502517c3c12 where `me->editflag & ME_EDIT_MIRROR_X` was replaced by
either `me->editflag & ME_EDIT_VERTEX_GROUPS_X_SYMMETRY` or
`me->symmetry & ME_SYMMETRY_X`. In this case, the former wasn't working,
so I replaced it with the latter.
Continuation of the work started with 249e4df110. After all display modes
were ported to this new design, this commit starts the (more complex) work on
the individual tree-element types. More concretely it ports animation
tree-elements (action data-blocks, drivers and NLA data).
The commit above explains motivations. In short, we need a better design that's
easier to reason about and better testable.
Changes done here are pretty straight forward and introduce similar class
hierarchy and building patterns as introduced for the display modes already.
I.e. an abstract base class, `AbstractTreeElement` with derived classes for the
concrete types, and a C-API with a switch to create the needed objects from a
type enum. The latter should be replacable with something nicer later on (RAII
based, and type-safer through meta-programming).
Each tree-element type has its own class, with an own header and source file
(okay some closely related types can share a header and source file, like the
NLA ones).
I added some further temporary bits for the transition to the new design, such
as the `TreeElement.type`. It should entirely replace `TreeElement` eventually,
just as `outliner_add_element()` should be quite small by then and easily
replacable by a `TreeBuilder` helper.
Mistake in 35a5dee2ef.
Code would simply assume that the element under the cursor is an Object if it
was an ID (but not a collection).
This wouldn't cause any issues in current code, since the only other ID that
set the direct-data were collections, which are special in that they don't have
a `TreeStoreElem.type` of 0, which is already being checked for here. And if
there was no direct-data set, the object lookup in the View-Layer would
correctly fail too.
I carefully checked and am quite sure for `TSE_ANIMATION_DATA` and
`TSE_NLA_ACTION` the direct-data isn't used at all. This data is stored and
accessed in unsafe ways based on conventions anyway (`void *`). We should aim
for a safer design but it will be difficult to get there. Any removed
complexity will help.
Non-ID tree-elements would cast their data pointer to `ID *` and take the name
and ID-Code from there. The name would actually be overridden a few lines
later, so that didn't cause issues. But the ID-Code stored inside the tree
element kept an undefined value. In practice that probably didn't cause many
issues either, since it's just an undefined value that was very unlikely to
take a valid 16-bit ID-code value, meaning ID-Code checks would simply fail as
they should. Further there typically are other checks to see if the element
actually represents an ID.
However, in theory this may have caused a few random crashes or
data-corruptions.
Modernize loops by using the `for(type variable : container)` syntax.
Some loops were trivial to fix, whereas others required more attention
to avoid semantic changes. I couldn't address all old-style loops, so
this commit doesn't enable the `modernize-loop-convert` rule.
Although Clang-Tidy's auto-fixer prefers to use `auto` for the loop
variable declaration, I made as many declarations as possible explicit.
To me this increases local readability, as you don't need to fully
understand the container in order to understand the loop variable type.
No functional changes.
Blender has now the place to store the Cryptomatte settings. This patch
migrates Cycles to use the new settings.
Reviewed By: Brecht van Lommel
Differential Revision: https://developer.blender.org/D9746
Shared helper function to create a split layout with an alert icon for popup dialogs.
Differential Revision: https://developer.blender.org/D9486
Reviewed by Julian Eisel
New layout for the 'About' dialog featuring the full version of the Blender logo.
Differential Revision: https://developer.blender.org/D9507
Reviewed by Hans Goudey
This feature is intended to generate Microsoft Compiled HTML Help files
which is not something we use/need.
This also fixes an error in generation because the executable cannot be
found.
I didn't actually confirm this is caused by invisible Outliners. But I'm pretty
sure the error happens with Outliners that aren't initialized (so were open in
an area before, but another editor is active there currently).
In that case, the runtime data may not be set yet and that is fine.
Fixes T83420.
No functional changes. This is a few minor cleanups to the remaining C
code for building the outliner tree after parts have been moved to C++.
Differential Revision: https://developer.blender.org/D9741
No functional changes. Code is ported to C++ with additional cleanups to
the logic and variable names.
Differential Revision: https://developer.blender.org/D9741
This way Outliner internal data stays internal, non-Outliner code will not be
able to access and mess with this. Further it allows us to use the real type
(rather than `void *`), change the type to a C++ container if needed and
slightly reduces the size for every Outliner stored in files.
Slightly changed how we set the `SO_TREESTORE_REBUILD` for this, but it should
effectively behave the same way as before.
I could use a 16 bit atomic fetch + AND for D9719. The alternative would be to
turn a `short` into a `int` in DNA, which isn't a nice workaround.
Also adds tests for the new functions.
This patch builds clang-extra-tools on macOS for the
clang-tidy binary. The script "run-clang-tidy.py" is
also harvested because using the `CMAKE_C[XX]_CLANG_TIDY`
option can miss out some files (like makesrna), and using the
script is faster as it does not compile the files.
Thanks to `@LazyDodo` for the base patch D8502.
Reviewed By: LazyDodo, sebbas, #platform_macos
Differential Revision: https://developer.blender.org/D9450
Technically it shouldn't have any effect on these editors.
The key tips in the header can be misleading.
The effect it previously had was not intended.
This involves re-implementing some of Blender-defined helpers in utils,
we keep debug code to ensure those are still matching on
behavior/results sides.
This will allow to get more i18n tools independent from blender
executable.
With rBc0677b662f4b, we try to track all modal events in order to detect the
one corresponding to the release.
But modifier keys can mask the modal event and thus confirm realease ends up
being skipped.
This resulted in the T83387.
With this commit we now read the actual key drop value in the modal event.
This fixes T83387
Support for the AO and bevel shader nodes requires calling "optixTrace" from within the shading
VM, which is only allowed from inlined functions to the raygen program or callables. This patch
therefore converts the shading VM to use direct callables to make it work. To prevent performance
regressions a separate kernel module is compiled and used for this purpose.
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D9733
Replace `typedef` with `using` in C++ code.
In the case of `typedef struct SomeName { ... } SomeName;` I removed the
`typedef` altogether, as this is unnecessary in C++. Such cases have been
rewritten to `struct SomeName { ... };`
No functional changes.
In ad85256e71 the "Show Group Colors" option was changed from a
per-editor option to a user preference. Since so many people wanted to
turn this option off, this makes sense. However, this move caused some
confusion because the option was just gone from the menu.
This commit adds a dummy menu item. It's disabled, and the tooltip
explains that the option can now be found in Preferences.
T83390 was created to track the removal of these hints.
Reviewed by: Severin
Differential Revision: https://developer.blender.org/D9735
Cryptomatte is a standard to efficiently create mattes for compositing. The
renderer outputs the required render passes, which can then be used in the
compositor to create masks for specified objects. Unlike the Material and Object
Index passes, the objects to isolate are selected in compositing, and mattes
will be anti-aliased.
Cryptomatte was already available in Cycles this patch adds it to the EEVEE
render engine. Original specification can be found at
https://raw.githubusercontent.com/Psyop/Cryptomatte/master/specification/IDmattes_poster.pdf
**Accurate mode**
Following Cycles, there are two accuracy modes. The difference between the two
modes is the number of render samples they take into account to create the
render passes. When accurate mode is off the number of levels is used. When
accuracy mode is active, the number of render samples is used.
**Deviation from standard**
Cryptomatte specification is based on a path trace approach where samples and
coverage are calculated at the same time. In EEVEE a sample is an exact match on
top of a prepared depth buffer. Coverage is at that moment always 1. By sampling
multiple times the number of surface hits decides the actual surface coverage
for a matte per pixel.
**Implementation Overview**
When drawing to the cryptomatte GPU buffer the depth of the fragment is matched
to the active depth buffer. The hashes of each cryptomatte layer is written in
the GPU buffer. The exact layout depends on the active cryptomatte layers. The
GPU buffer is downloaded and integrated into an accumulation buffer (stored in
CPU RAM).
The accumulation buffer stores the hashes + weights for a number of levels,
layers per pixel. When a hash already exists the weight will be increased. When
the hash doesn't exists it will be added to the buffer.
After all the samples have been calculated the accumulation buffer is processed.
During this phase the total pixel weights of each layer is mapped to be in a
range between 0 and 1. The hashes are also sorted (highest weight first).
Blender Kernel now has a `BKE_cryptomatte` header that access to common
functions for cryptomatte. This will in the future be used by the API.
* Alpha blended materials aren't supported. Alpha blended materials support in
render passes needs research how to implement it in a maintainable way for any
render pass.
This is a list of tasks that needs to be done for the same release that this
patch lands on (Blender 2.92)
* T82571 Add render tests.
* T82572 Documentation.
* T82573 Store hashes + Object names in the render result header.
* T82574 Use threading to increase performance in accumulation and post
processing.
* T82575 Merge the cycles and EEVEE settings as they are identical.
* T82576 Add RNA to extract the cryptomatte hashes to use in python scripts.
Reviewed By: Clément Foucault
Maniphest Tasks: T81058
Differential Revision: https://developer.blender.org/D9165
This patch adds support for AOVs in EEVEE. AOV Outputs can be defined in the
render pass tab and used in shader materials. Both Object and World based
shaders are supported. The AOV can be previewed in the viewport using the
renderpass selector in the shading popover.
AOV names that conflict with other AOVs are automatically corrected. AOV
conflicts with render passes get a warning icon. The reason behind this is that
changing render engines/passes can change the conflict, but you might not notice
it. Changing this automatically would also make the materials incorrect, so best
to leave this to the user.
**Implementation**
The patch adds a copies the AOV structures of Cycles into Blender. The goal is
that the Cycles will use Blenders AOV defintions. In the Blender kernel
(`layer.c`) the logic of these structures are implemented.
The GLSL shader of any GPUMaterial can hold multiple outputs (the main output
and the AOV outputs) based on the renderPassUBO the right output is selected.
This selection uses an hash that encodes the AOV structure. The full AOV needed
to be encoded when actually drawing the material pass as the AOV type changes
the behavior of the AOV. This isn't known yet when the GLSL is compiled.
**Future Developments**
* The AOV definitions in the render layer panel isn't shared with Cycles.
Cycles should be migrated to use the same viewlayer aovs. During a previous
attempt this failed as the AOV validation in cycles and in Blender have
implementation differences what made it crash when an aov name was invalid.
This could be fixed by extending the external render engine API.
* Add support to Cycles to render AOVs in the 3d viewport.
* Use a drop down list for selecting AOVs in the AOV Output node.
* Give user feedback when multiple AOV output nodes with the same AOV name
exists in the same shader.
* Fix viewing single channel images in the image editor [T83314]
* Reduce viewport render time by only render needed draw passes. [T83316]
Reviewed By: Brecht van Lommel, Clément Foucault
Differential Revision: https://developer.blender.org/D7010
This node reads two input attributes as floats and allows for basic
element-wise comparison operations. Like the attribute math node,
it's also possible to switch to using floats as inputs. This node can
be combined with the point separate node for more flexibility on which
points to separate out.
I used the same abstractions that the attribute math node uses in this
commit, assuming that makes sense.
Improvements to the layout of the Keymaps section of Preferences by removing unneeded Decorator columns.
Differential Revision: https://developer.blender.org/D9726
Reviewed by Hans Goudey
The threshold input is unecessary with an "attribute compare"
node before this one, or with attribute rounding. So to remove some
complexity it is removed, and a boolean input attribute is used instead.
For the implementation, the existing code becomes a bit more specific
to point cloud data, because I'm realizing that mesh separate is quite a
different problem, and it doesn't make sense to force the same code
to be able to solve that too.
This adds a new property to the sculpt vertex color paint brush to limit
the area of the brush that is going to be used to sample the wet paint
color. This is exactly the same concept as normal radius and area radius
that exist for sculpting brushes for sampling the surface depth and
orientation.
When working near color hard edges, this allows to prevent the color
from the other side of the edge to blend into the wet paint.
With 1.0 (the previous default) wet paint radius, as soon as the brush touches
one vertex of the other color, the wet paint mix color changes, making it
impossible to maintain the border between the two colors.
Reviewed By: sergey, dbystedt, JulienKaspar
Differential Revision: https://developer.blender.org/D9587
When inverting erase displacement the filter can increase the
displacement over the limit surface. After using apply base, this can be
used as an alternative intensify details as it usually gives better
results.
This is the same concept as smoothing inverting to intensify details.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D9679
This was missing from D9623. Now the same sculpt scene stats are also
displayed in the status bar.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D9700
This enables the use of clang-tidy in the VS IDE.
To use it:
1 - Enable WITH_CLANG_TIDY in your cmake configuration
2 - From the Analyse pull down menu select Run Code Analysis on...
The analyser is currently not enabled by default on build
given it is quite slow and there are quite a few problems
it reports that we still need to deal with.
When change the order of the stroke using the arrange operator, the arrange was done, but the viewport did not display the result until you refresh the viewport.
These operators existed since 2.83, but the menu was hidden by error.
Also the operators have been cleanup and make multiframe compatible.
Reviewed By: mendio
Differential Revision: https://developer.blender.org/D9671
NanoVDB includes "assert.h" and makes use of "assert" in several places and since the compile
pipeline for CUDA/OptiX kernels does not define "NDEBUG" for release builds, those debug
checks were always added. This is not intended, so this patch disables "assert" for CUDA/OptiX
by defining "NDEBUG" before including NanoVDB headers.
This also fixes a warning about unknown pragmas in NanoVDB thrown by the CUDA compiler.
This change solves a bottleneck which was caused by attempt to cache
postprocessed search areas used for tracking. It was a single cache
used by all threads, which required to have some synchronization
mechanism. This synchronization turned out to be making all threads
to idle while one thread is accessing the cache. The access was not
cheap, so the multi-threading did not provide expected speedup.
Current solution is to remove the cache of search areas. This avoids
any threading synchronization overhead because there is no need for
it anymore. The downside is that for certain configurations tracking
became slower when comparing to master branch. There is no expected
slowdown compared to 2.91 release.
The slowdown is mainly experienced when using big search area and
keyframe matching strategy. Other cases should still be within a
ballpark of performance of single-threaded code prior to this change.
The reason why is it so is because while this change makes it so the
image accessors needs to process images multiple times the complexity
of this process is almost the same as all the overhead for the cache
lookup and maintenance.
Here are Some numbers gained on different configurations.
CPU: Intel Xeom CPU E5-2699 v4
OS: Linux
Footage: Old_Factory MVI_4005.mov from the first part of Track Match
Blend training which can be found on the Blender Cloud.
Tracking 443 markers across 250 frames. The unit is seconds.
File: F9433209
2.91: 401.520874
before: 358.650055
after: 14.966302
Tracking single marker across 250 frames. The unit is seconds.
File: F9433211
2.91 before after
Big keyframe 1.307203 1.005324 1.227300
Big previous frame 1.144055 0.881139 0.944044
Small keyframe 0.434015 0.197760 0.224982
Small previous frame 0.463207 0.218058 0.234172
All at once 2.338268 1.481220 1.518060
Add Custom Space to the list of space conversions for constraints.
Constraints can use World Space, Local Space, Pose Space, Local with
Parent, and now also Custom Space with a custom object to define the
evaluation space.
The Custom Space option uses the Local Space of an other
object/bone/vertex group. If selected on owner or target it will show a
box for object selection. If an armature is selected, then it will also
show a box for bone selection. If a mesh object is selected it will show
the option for using the local space of a vertex group.
Reviewed By: #animation_rigging, sybren, Severin, angavrilov
Differential Revision: https://developer.blender.org/D7437
Fix linker errors by explicitly linking PugiXML when OpenImageIO is
used.
PugiXML has been separated from OpenImageIO (D8628). This means that any
time OpenImageIO libraries are linked, the PugiXML libraries have to be
linked as well. This was set up correctly for Cycles, but not for imbuf.
Because of this, building without Cycles but with OIIO would produce
linker errors.
These comments are used to generate Py API docs.
Most of the comments are just copies of the mesh operator counterparts.
More improvements can be made here in the future.
Thanks to Jon Denning on twitter:
https://twitter.com/gfxcoder/status/1334178566993555459
The icons for collapsed children already draw highlighted when hovered.
Because the item icons are now select targets (for outliner to properties
editor tab switching) this adds highlights on hover for all outliner
element icons.
During the development of the new nodes in the `geometry-nodes` branch
the color of the new nodes wasn't considered, so all of the nodes ended
up red, the color for "input" nodes. This patch introduces two new
colors, one for "Geometry" and one for "Attributes". There are only two
attribute nodes currently, but the next sprint will add two more,
attribute mix, and sample from texture. The attribute nodes are
conceptually different enough from the nodes that modify the geometry
that they deserve their own color.
Differential Revision: https://developer.blender.org/D9682
Until there is a icon made specially for this, the nodetree icon is up
for grabs. Using it in the nodegroup + modifier + editor helps the users
to make a connection on where to edit those modifiers.
The point cloud object is the only one that will support instancing at
first. So we can expose it as a regular object.
It is limited since it has no edit mode. But this is not different than
the volume object.
This commit adds functions to set and get the object's active
modifier, which is stored as a flag in the ModifierData struct,
similar to constraints. This will be used to set the context in
the node editor. There are no visible changes in this commit.
Similar to how the node editor context works for materials, this commit
makes the node group displayed in the node editor depend on the active
object and its active modifier. To keep the node group from changing,
just pin the node group in the header.
* Shortcuts performed while there is an active modifier will affect
only that modifier (the exception is the A to expand the modifiers).
* Clicking anywhere on the empty space in a modifier's panel will make it active.
These changes require some refactoring of object modifier code. First
is splitting up the modifier property invoke callback, which now needs
to be able to get the active modifier separately from the hovered
modifier for the different operators.
Second is a change to removing modifiers, where there is now a separate
function to remove a modifier from an object's list, in order to handle
changing the active.
Finally, the panel handler needs a small tweak so that this "click in panel"
event can be handled afterwards.
The handling of muted nodes is handled at the derived node tree
level now. This is also where expanding node groups is handled.
Muted nodes are relinked and removed from the derived tree
during construction. The geometry node evaluation code does
not have to know about muted nodes this way.
Objects can evaluate to a geometry set instead of a single ID (only point cloud
objects for now). In the depsgraph object iterator, the evaluated geometry
components are expanded into temporary objects.
It's important to note that instanced objects can also contain geometry
components. Therefore, they have to be split up into multiple objects
as well in some cases.
At a high level the iterator works like so:
```
for object in depsgraph:
for component in object:
yield object_from_component(component)
for dupli in make_duplis_list(object):
for component in dupli:
yield object_from_component(component)
```
DEG_iterator_objects_next has been cleaned up, to make this structure
a bit more apparent.
This should not change anything for objects that are not point clouds.
This is the initial merge from the geometry-nodes branch.
Nodes:
* Attribute Math
* Boolean
* Edge Split
* Float Compare
* Object Info
* Point Distribute
* Point Instance
* Random Attribute
* Random Float
* Subdivision Surface
* Transform
* Triangulate
It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier.
Notes on the Generic attribute access API
The API adds an indirection for attribute access. That has the following benefits:
* Most code does not have to care about how an attribute is stored internally.
This is mainly necessary, because we have to deal with "legacy" attributes
such as vertex weights and attributes that are embedded into other structs
such as vertex positions.
* When reading from an attribute, we generally don't care what domain the
attribute is stored on. So we want to abstract away the interpolation that
that adapts attributes from one domain to another domain (this is not
actually implemented yet).
Other possible improvements for later iterations include:
* Actually implement interpolation between domains.
* Don't use inheritance for the different attribute types. A single class for read
access and one for write access might be enough, because we know all the ways
in which attributes are stored internally. We don't want more different internal
structures in the future. On the contrary, ideally we can consolidate the different
storage formats in the future to reduce the need for this indirection.
* Remove the need for heap allocations when creating attribute accessors.
It includes commits from:
* Dalai Felinto
* Hans Goudey
* Jacques Lucke
* Léo Depoix
Note: This also changes the Shader socket color, to match "Shading" in the Outliner.
Theme update for shader nodes will be committed separately.
Ref T82689.
This commit expands the Windows-specific code in rBdca9aa0053f7 and Linux-specific code in rB33b7d53df08a.
It also fixes a capitalization issue in FindPugiXML.cmake
Since Blender 2.91 the TeraScale 2 based cards crash during startup.
This patch will show the user a screen that the platform they are using
isn't supported.
The GPUs have been carefully handpicked from dozens of reports. T83124,
T83127, T83103, T83091, T83045, T83065, T82750, T82889, T82925, T82640,
T82429, T82436, T82446.
This fixes T68521, T82334.
The object are not properly mirrored in object mode.
For mirroring an object that has an arbitrary rotation along the X axis
this is the procedure:
1. mirror x location.
2. negate x scale of the
3. negate y and z component of the rotation (independent of which rotation mode is used).
This knowledge applies now for all angles and axes to finally make the
mirror operation work in object mode.
The new mirror function has the downside that it can not (in this form)
be used with proportional editing.
This is no problem since the old behavior can still be replicated by
scaling with -1 along any axis.
The solution to get perfect mirrors can not work for scaling in general,
because in that case there could be scew created.
Reviewed By: mano-wii
Differential Revision: https://developer.blender.org/D9625
Caused by rBba97da21acf2.
For non-topology mirror hidden verts were never respected/included in
EDBM_verts_mirror_cache_begin (they were excluded from the kdtree).
Prior to said commit, hidden mirrored verts that were still in the map
would have been excluded in a separate loop over vertices in
'editmesh_mirror_data_calc()' by checking BM_ELEM_HIDDEN. Due to the new
nature of this function this check was now moved to
EDBM_verts_mirror_cache_begin.
Maniphest Tasks: T83119
Differential Revision: https://developer.blender.org/D9673
The Vertex Weight Edit Modifier already got the Custom Curve, there was no
real reason for the proximity not to have it as well.
With some fixes by Bastien Montagne (@mont29).
Reviewed By: mont29
Differential Revision: https://developer.blender.org/D9594
This infinite loop is caused by a conflict between the volume mesh
creation which unintentionally clears the shaders before early exiting
when no grid is found, and the Blender exporter which adds back the
shaders causing us to reupdate as the shaders changed.
To fix this simply preserve the shaders on the Volume node.
Use RNA_api_sequences() for SequenceEditor and MetaSequence
sequences member.
Defines pair of dispatch functions rna_Sequences_editing_* and
rna_Sequences_meta_* that pass pointer to seqbase to
rna_Sequences_* function.
Downside of this implementation is, that it defines 2 seemingly
different RNA collections - SequencesMeta and SequencesTopLevel
Reviewed By: mont29
Differential Revision: https://developer.blender.org/D9601
Make more of the logic generic so that it can more easily work on more
component types in the future. Basically anything that handles attributes
can be generic.
I'm not quite sure creating the "radius" attribute in here is necessary
when any other attribute can be created later, but without this the
original radius attribute didn't pass to the new point cloud components
properly.
This should be investigated though, because the "radius" attribute
of pointclouds currently exists in some "semi-built-in" space that
doesn't really make sense.
This node takes an input geometry and ouputs two geometries, moving
some points to the first and other points to the second depending on a
mask attribute input and a threshold.
The implementation tries to be "attribute component agnostic", meaning
it should be relatively easy to add "separate" functionality for more
component types in the future when they need to support it. (Then the
node could be renamed "Separate Geometry").
There is still an error carrying over the values for some attributes to
the output geometries.
During the development of the new nodes in the `geometry-nodes` branch
the color of the new nodes wasn't considered, so all of the nodes ended
up red, the color for "input" nodes. This patch introduces two new
colors, one for "Geometry" and one for "Attributes". There are only two
attribute nodes currently, but the next sprint will add two more,
attribute mix, and sample from texture. The attribute nodes are
conceptually different enough from the nodes that modify the geometry
that they deserve their own color.
Differential Revision: https://developer.blender.org/D9682
PugiXML was historically shipped hidden embedded into OIIO, the Grease
Pencil team had a requirement for an XML library recently so pugi seems
like a natural choice since it's not really a 'new' library, we just
turn an implicit dependency into an explicit one.
This commit expands the Windows-specific code in rBdca9aa0053f7 to
include Linux. macOS support will be handled in a later commit.
NOTE: run `cmake -U'*PUGIXML*' .` in the build directory to ensure CMake
finds PugiXML in the new location.
For details see D8628
The idea is to avoid any synchronization needed in the worker threads
and make them to operate on a local data. From implementation detail
this is achieved by keeping track of "wavefront" of markers which are
to be tracked and the tracking result. Insertion of results to the
AutoTrack context happens from main thread, which avoids need in the
lock when accessing AutoTrack.
This change makes tracking of many (300+) about 10% faster on the
Xeon) CPU E5-2699 v4. More speedup will be gained by minimizing
threading overhead in the frame cache.
Another important aspect of this change is that it fixes non-thread
safe access which was often causing crashes. Quite surprising the
crash was never reported.
Log to verbosity level 1 rather than INFO severity.
Avoids a lot of overhead coming from construction of the INFO stream
and improves performance and threadability of code which uses logging.
This makes tracking of 250 frames of a track of default settings to
drop down from 0.6sec to 0.4sec.
Makes it more clear whether object is allowed or not allowed to be NULL.
Also, avoids possible access to the different object mode enumerator.
Should be no functional changes.
These functions were only checking the X axis for flipping the
displacement for a symmetry area depending on the initial position of
the pivot.
This affects transform and any other tools that transform vertices and
applies symmetry based on areas (the pose brush, for example).
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D9654
This adds the vertex and face count info to the scene stats in sculpt
mode. These stats count the active vertices and faces in the
sculptsession for the active object. This has the following advantages:
- It is possible to know how many vertices the sculptsession has active
comparted to the vertex count of the entire scene from sculpt mode
- When sculpting with constructive modifiers, these stats will report the
number of vertices that you can actually sculpt with, instead of the
vertex count of the modified mesh and the entire scene.
Reviewed By: sergey, dbystedt
Differential Revision: https://developer.blender.org/D9623
Adjust the threshold for switching from the base case to trace > 0,
based on very similar example code from www.euclideanspace.com to
avoid float precision issues when trace is close to -1.
Also, remove conversions to and from double, because using double
here doesn't really have benefit, especially with the new threshold.
Finally, add quaternion-matrix-quaternion round trip tests with
full coverage for all 4 branches.
Differential Revision: https://developer.blender.org/D9675
Extract initialisation code of box selection into separate functions.
No functional changes.
Reviewed By: zeddb, sybren
Differential Revision: https://developer.blender.org/D9196
Fix various problems in the Action Group rearranging code. All fixes are
necessary to resolve the bug.
- Before groups are rearranged, the channels are moved into their
respective groups (so no longer referenced by `action->channels`). A
temporary group is made for ungrouped channels. The code made
assumptions about the channels being in the same order as the groups;
that assumption has been removed.
- Looping over channels in an Action Group should stop when reaching the
last channel, and not until `NULL`.
- After all the reshuffling is done, the `action->channels` linked list
wasn't terminated properly. Now `first.prev` and `last.next` are set
to `NULL` to avoid infinite loops.
Previously image accessor was sharing array pointer for tracks access.
Now it is possible to pass a temporary array valid only during the
initialization process.
Should be no functional changes.
This is something not-so-trivial to see from just reading code, which
shown an important of proper comments and clear naming.
Main source of confusion was that it is not immediately clear that
AutoTrack context is to see all tracks, but tracking to only operate on
selected ones.
The modifier was called "Empty", which reflects the fact that the modifier
doesn't do anything when you add it. However, that is not true after you
change the modifier's node group, and the fact that the modifier uses
nodes is the one thing that distinguishes it at this point. This commit
changes the name to "Geometry Nodes", which is simple and consistent
with the "Geometry Node Editor" and the naming of the project so far.
The windows build leaves a convenience helper script
in the build folder called, rebuild.cmd. This change
passes any parameters you give rebuild.cmd to ninja
so you can easily pass it additional parameters without
having to edit the batch file manually.
This operators reset the vertex color information of the strokes.
If nothing is selected, all strokes are reset. If any is selected, only selected strokes are reset.
Also added a new menu Paint in Vertex Color mode.
Differential Revision: https://developer.blender.org/D9647
The event queue can contain events from before pointer warping, ignore those
now. This is an old issue, but became more common now that we disabled event
coalescing and started using the event mouse location rather than the current
mouse location.
Thanks to Yevgeny Makarov and Nicholas Rishel for helping solve this.
Ref D9662
Blacklist a bunch of tests on the GPU, which are known to currently have
differences with CPU. These can be enabled as they are fixed or the test
is modified to give compatible results when there are known limitations.
OSL tests were also moved to their own directory since those are not
supported on the GPU.
Ref T82193
CYCLES_TEST_DEVICES is a list of devices (CPU, CUDA, OPTIX, OPENCL). It is set
to CPU only by default.
Test output is now writen to build/tests/cycles/<device>, and the HTML report
has separate report pages for the different devices, with option to compare
between CPU and GPU renders.
Various GPU tests are still failing due to CPU/GPU differences, these are to be
fixed or blacklisted still.
Ref T82193
Allows to use mutex, scoped_lock, and conditional_variable from within
the libmv namespace.
Implementation is coming from C++11. Other configurations are easy to
implement, but currently C++11 is the way to go.
During revision of {D8952} one of the comments was to make a function that converts the render percentage to a factor. This to avoid code duplication. However the duplicated code was already all over the compositor code. So in order to avoid this code duplication for {D8952} I propose to first cleanup the duplicated code and build patch {D8952} based on this clean up.
The method that converts the render percentage to a factor is put in the CompositorContext. Why? The CompositorContext keeps DNA information like the renderdata. DNA, and thus the CompositorContext, keeps the size of the render resolution in percentage (user oriented). The compositor needs the size of the render resolution as a factor. So the CompositorContext seems like the obvious place to have this conversion method.
Why not in de NodeBase? The method could've been added to the nodebase, but I wanted to keep the nodebase as clean as possible and not put simple "conversion" methods into this base class. Also I didn't really like the call flow: you'd always have to get the renderdata size from the context and then convert.
Putting it in the CompositorContext avoids this extra invoke of a call.
Why not in the Converter? See nodebase. And the Converter seems more like a class for "structural" and complex node tree conversions. Not the simple conversions.
Reviewed By: Sergey Sharybin
Differential Revision: https://developer.blender.org/D9566
This commit replaces while loops and for loops with the equivalent
macro. This results in much more readable code in some places,
and it's now more apparent when the situation is more complicated
than just iterating through a linked list.
As proposed in T74432 and already implemented in several commits,
"region" is the preferred name for `ARegion` variables, rather than
any variant of "ar". This commit changes a few "ar" variables that have
popped up over time and also adjusted names of variants like "arnew".
This separates out PugiXML that was previously
bundled by OIIO.
As this linux/mac libs are not available
this commit only contains the builder and windows
changes, and the option to enable pugixml is
guarded by a platform if, this can be removed
once all platforms have committed the svn libs.
For details see D8628
By checking if a cell has already been processed in the finding patch
component code, an enormous speedup happens. This only will be
noticeable if there are lots of patches, and some cells with a
large number of patches.
In case where there are coplanar instersections where
each part has a lot of triangles, the finding-cells algorithm was
very inefficient. This uses a Set instead of a Vector to keep track
of a cell's patches, avoids going through all patch x patch combinations,
avoids going through all patches to renumber after a merge, and
merges smaller patch-sixe cells into larger ones.
All this reduces the time to find cells in the cited case by a factor of 10.
The handling of muted nodes is handled at the derived node tree
level now. This is also where expanding node groups is handled.
Muted nodes are relinked and removed from the derived tree
during construction. The geometry node evaluation code does
not have to know about muted nodes this way.
Note: This also changes the Shader socket color, to match "Shading" in the Outliner.
Theme update for shader nodes will be committed separately.
Related to T82689
Ensure consistent order of pose bones. Now it should always match the
one from bones in armature obdata (as exposed by e.g. RNA, i.e.
children-first).
Previously when some pose bones would need to be added or removed from a
pose due to changes in the bone armature, or if bones in armature were
re-ordered, the bones order in pose would not match anymore the one from
armature, and could even become different between e.g. a proxy and its
linked source.
This was not really nice, but not a big issue before either. But with
diffing process of override, consistent order of items between reference
linked collection and local override one is crucial.
Reviewed By: #animation_rigging, sybren
Maniphest Tasks: T82758
Differential Revision: https://developer.blender.org/D9646
The change was intentional so that the orientation matrices match
(`rv3d->viewinv` becomes equal to the orientation matrix).
But, although in a projection matrix the Z axis is negative, this
should not be so from the user's point of view.
So the solution here is to negate the Z axis when the View orientation
is chosen.
This affects all modes, but is only evident for rotation.
---
Another change here is to use the final rotation value (`values_final`)
for the gizmo drawing since this value can be changed by the mode.
Avoid the evaluation of constraints when computing the parent-inverse
matrix.
Constraints are meant to be evaluated last; object transforms are
computed this order:
1. `parent->obmat` (the parent object's world matrix)
2. `ob->parentinv` (the object's parent-inverse matrix)
3. Object's loc/rot/scale
4. Object's constraint evaluation
When the constraints are used to compute the parent-inverse matrix,
their effect is moved from step 4 to step 2 in this list, potentially
rotating or scaling the object's local transform. This causes unwanted
movement as reported in T82156.
Reviewed By: looch
Differential Revision: https://developer.blender.org/D9413
Move the logic for determining if the item at a given x position is an
icon into the function. This is used for determining selection over an
icon, and will be used in a later commit for checking for hover over an
icon. No functional changes.
This adds support for incremental updates in the sculpt transform
code. Now tools can define if they need the displacement applied
for the original coordinates or incrementally.
This is needed for features like elastic transform or cloth deformation
target in the transform tool.
No functional changes.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D9547
Before, attributes could only be accessed via setter and getter functions.
While this works well, it can be cumbersome and slow.
It would be better if most code could just operate on the underlying
attribute array directly. The issue is that some attributes are currently
not stored in plain arrays (e.g. vertex positions, vertex groups, ...).
This patch implements a solution where these "special" attributes are
converted into a temporary array for easy access and then copied
back into their correct form after all modifications are done.
Attribute accessors now have a `get_span` and `apply_span` method,
which can be used exactly for that purpose. In the case of attributes
that are plain arrays internally, only small constant overhead is added.
This was reported in the form of the eyedropper of the 'Parent' property
creating a custom property 'parent' if self was picked.
Problem arises when certain checks for setting rna pointer properties
failed (for example: the PROP_ID_SELF_CHECK check) and then a different
code path was entered (which was only meant for IDProperties).
Problem was introduced in rBa7b3047cefcb.
To solve, now first enter the branch for rna-based pointer properties,
then perform the sanity-checks (and if these fail: dont enter the other
unrelated codepath but instead do nothing)
Maniphest Tasks: T83055
Differential Revision: https://developer.blender.org/D9652
dynamic paint modifiers
Looks like prior to the introduction of mantaflow, the former
SmokeModifierData always had a domain initialized (even if its type was
set to None). Since mataflow, the FluidModifierData type needs to be set
to MOD_FLUID_TYPE_DOMAIN (otherwise domain is invalid)
Maniphest Tasks: T82729
Differential Revision: https://developer.blender.org/D9644
Code was actually not applying any override operation over linked data.
Reasonn behind that was that if library file is saved with latest
override applied then this is not needed, since data saved for the
override in the lib file is already up to date.
But this is actually fully breaking in case someone update the lib file
of the lib file, without re-saving the libfile itself.
So now we alwaya apply overrides also on linked data.
Note that this will not fix the case where a resync is needed.
The panel title text intersected any buttons in the header because
the label offset retrieved from the layout code was not scaled by
the block's zoom level. Error in rB0d93bd8d63980.
The transform modifiers are confirmed by releasing any button.
Thus, the operation can be falsely confirmed if the button that launched
the operation is released after the modifier has been activated.
Previously the events that confirmed the modifiers were hardcoded.
An option to fix this would be to add custom confirmation keyitens for
specific modifiers. But this can be a bit confusing and would make the
modal keymap even bigger.
So the solution here is to skip the button that launched the operation
when confirming the modifier.
Material indices and smooth states are not destroyed anymore.
If the two incoming meshes use different materials, the materials of
the first input will be used. This is obviously not ideal, because the
second input might change its material.
In the future we want to handle that differently and keep both materials,
but for now this is consistent with the boolean node.
Orthogonalize the constraint target's matrix before decomposing it into
Euler angles. This removes sheer, and is actually a requirement for
correct decomposition.
It's conceivable that someone has used the incorrect behaviour in a rig.
As the shear caused unpredictable flipping of the constrained object,
this is unlikely.
Reviewed By: angavrilov, sybren
Differential Revision: https://developer.blender.org/D8915
OpenImageIO's ustring makes assumptions about the interals of
std::string which aren't true on macOS/ARM. A patch for OpenImageIO
addresses this for macOS/ARM builds only, at the expense of potentially
duplicated allocations of strings.
The build script for macOS dependencies expects to find bison provided
by Homebrew under /usr/local. Since the default install location for
Homebrew changes on macOS/ARM, the script now looks at the host
architecture and switches to /opt/homebrew on Macs with Apple Silicon.
This update introduces two improvements from the Mantaflow repository:
(1) Improved particle sampling:
- Liquid and secondary particles are sampled more predictably. With all parameters being equal, baked particles will be computed at the exact same position during every bake.
- Before, this was not guaranteed.
(2) Sparse grid caching:
- While saving grid data to disk, grids will from now on be saved in a sparse structure whenever possible (e.g. density, flame but not levelsets).
- With the sparse optimization grid cells with a value under the 'Empty Space' value (already present in domain settings) will not be cached.
- The main benefits of this optimization are: Smaller cache sizes and faster playback of simulation data in the viewport.
- This optimization works 'out-of-the-box'. There is no option in the UI to enable it.
- For now, only smoke simulation grids will take advantage of this optimization.
Previously, the alternate row color in the Video Sequence Editor was
just a shaded version of the editor's background color. This makes it
theme-able just like in the file browser and outliner, although the
default color is very slightly different.
Differential Revision: https://developer.blender.org/D9634
If the cursor is not over the mesh, the operator was still using the
last Face Set ID updated by the drawing cursor code when the cursor was
over the mesh.
This now cancels the operator instead of modifying a Face Set that will
look random to the user.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D9545
The original name was chosen because symmetry was a property of
the Paint struct. Now symmetry is a property of meshes, which
also makes the name of the property easier to understand.
Reviewed By: sergey, dbystedt
Differential Revision: https://developer.blender.org/D9618
This commit makes the way the selector in the middle of the node editor
header interacts with the active modifier much more intuitive.
- With no active modifier, or when the active modifier is a non-node
modifier, the "new" operator adds a new nodes modifier.
- With an active node modifier, the drop-down affects the modifier's
selected node group.
- With no active object, there is no drop-down in the header.
- The node editor's node group always updates, even when there is no
active modifier.
- Pinning disables the drop-down and keeps the same node group
selected, regardless of the context.
This is designed to work similarly to materials and material slots,
which should be familiar to any Blender user. One workflow not allowed
by this design is editing any geometry node group, regardless of the
context, but if that turns out to be an issue it should be solved in a
similar way for all node editors.
Since D6133 fluid particle code uses thread local storage to collect
springs created during a time step before adding them to the actual
spring array.
Prior to the switch to TBB there was a single finalize callback which
was called on the main thread, so it could use psys_sph_flush_springs
and insert the new entries into the final buffer. However in D7394 it
was replaced with a reduce callback, which is supposed to be thread
safe and have no side effects. This means that the only thing it can
safely do is copy entries to the other temporary buffer.
In addition, careful checking reveals that the 'classical' solver
doesn't actually add springs, so reduce isn't needed there.
Differential Revision: https://developer.blender.org/D9632
The roll value may not be as expected when a matrix is not orthogonal
or has a negative determinant.
This can lead to confusion as seen in T82930.
Therefore, make it clear that this is a limitation and that a value for
the roll is somewhat indeterminable in these cases.
This fixes T82930
This adds "modifier" as a new context variable in the property editor.
It is used instead of getting the active modifier directly. This way it
displays in the breadcrumbs panel, and the variable can be set only
in the modifier tab, meaning it won't work from the shortcut even
when the modifiers aren't visible.
While we can support different domains already in this node, there is not
really anything you could do with that data yet. It can be added back when
there is a use case for it.
Currently, implementing a node that supports math operations requires
you to write a long switch statement. The goal of this abstraction
is to reduce the number of these switch statements.
Note, this commit adds new such long switch cases. However, in the upcoming
commits existing math nodes are updated to use this abstraction, resulting in
less code overall.
In `calc_curve_deform` a factor is calculated without checking if
the divisior is zero or close to zero. This patch adds the missing
checks and sets the factor to zero if the division shouldn't be
computed.
Reviewed By: mont29
Differential Revision: https://developer.blender.org/D9645
Blender uses 64bit atomics to manipulate SessionUUID, and these atomics
were not defined on any of 32bit platforms.
While official support is limited to 64bit platforms only, the code
should not make assumptions about bitness or endianess, in terms that
there should be codepaths and fallback (or provision of them) for 32bit
platforms.
This change makes 64bit atomic functions defined for all platforms.
The atomic_test was compiled and successfully tested on i686 and armv7l
platforms. The rest of compilation process of Blender will be very
tedious, so that was not done.
This change is essential, but not necessarily enough to make Blender
compilable on i686 (ability to compile Blender on 32bit platforms was
lost during the 2.91 development).
This is a functional part of original fix done by Brecht in D9577.
There is a special defines block needed for ARM on Linux. Move it from
public header to an implementation file.
No functional changes.
This is a non-functional part of original fix done by Brecht in D9577.
The logic was duplicated.
Should be no functional changes. The modified function is expected
to give same exact results for all inputs.
On the "caching last-used track" topic. The code was using last_marker
to allow faster lookup of marker closest to the frame. With this
change it is still the case since the BKE_tracking_marker_get() does
cache last used marker.
Based on discussion with the team, summarized in T82691, it makes sense
to simplify the shortcut behavior and remove the way the shortcuts
currently activate on hover if there is no active modifier. This commit
also removes the behavior of clearing the active modifier when you
click on blank space in the modifier tab.
Selecting an object by clicking on its instances only worked,
when the object itself is visible. However, it is possible to hide
the object and still keep the instances visible.
The solution is to give every object the correct `select_id` in the
depsgraph object iterator right before rendering.
Reviewers: fclem, brecht
Differential Revision: https://developer.blender.org/D9640
This change will use the image engine to draw the backdrop of the compositor. With this patch the alpha blending will be done in Linear Scene Reference space and shows pure emissive colors.
See differential for an example image.
**Technical changes**
As only the backdrop drawing is done using the draw manager there are some technical changes.
1. The overlay buffer is partly drawn outside the draw manager. When drawing the backdrop image the overlay buffer needs to be masked to simulate premultiplied alpha under.
2. The backdrop of the node editor is done in region pixel space. A `DRWView` is constructed with this space.
3. UDIM textures uses world position to generate the UV coordinates. This has been implemented more strict by the `IMAGE_DRAW_FLAG_USE_WORLD_POS`. When the flag isn't used the local coordinates are used to generate the UV coordinates what is image space.
4. The draw manager now checks the actual `eSpaceType` of the space data to use different code paths. In the future the movie clip editor will be added.
NOTE: The preview images in nodes are drawn in display space and cannot show pure emissive colors. As preview images are used on more locations it is best to fix this in a separate patch.
Reviewed By: Clément Foucault
Differential Revision: https://developer.blender.org/D9451
The spikes were caused by non-initialized tangent matrix used during
smoothing process. The reason tangent matrix was not initialized was
because wrong usage of API: n-gons should pass corner of 0 to the
matrix construction function.
Corrected usage of the API and added assert() to help catching such
kind of non-initialized issues easier.
When clicking on an already-selected keyframe, mark it as active if the
click caused the previously-active keyframe to become deselected.
When clicking on a key in the graph editor, it is selected and all other
keys are deselected. If that key was already selected before the click,
it would not become the active keyframe. This is now fixed.
Reviewed by: Severin
Differential Revision: https://developer.blender.org/D9639
Regression introduced by {rB042143440d76}. The deform group lookup was
performed on the wrong object. Before the lookup was performed on
the lattice object. This patch changes this back.
Reviewed By: Bastien Montagne
Differential Revision: https://developer.blender.org/D9638
According to feedback the outliner to properties editor tab switching
was annoying when it always changed tabs on selection, especially for
selecting individual objects. This limits the tab switching behavior to
only when the icons in the outliner are selected.
This commit includes these improvements to operators and how they
relate to the active modifier:
- Shortcuts performed while there is an active modifier will affect
only that modifier.
- Clicking anywhere on the empty space in a modifier's panel will make
it active.
- Clicking anywhere in the empty properties space will clear the
active.
These changes require some refactoring of object modifier code. First
is splitting up the modifier property invoke callback, which now needs
to be able to get the active modifier separately from the hovered
modifier for the different operators. Second is a change to removing
modifiers, where there is now a separate function to remove a modifier
from an object's list, in order to handle changing the active. Finally
the panel handler neeeds a small tweak so that this "click in panel"
event can be handled afterwards.
Adds smooth scrolling with the trackpad for popovers. Also fixes the position of the scroll arrows on high-DPI.
Differential Revision: https://developer.blender.org/D9533
Reviewed by Brecht Van Lommel
Layout block safety sizes were not scaled correctly with interface scale.
Differential Revision: https://developer.blender.org/D9569
Reviewed by Brecht Van Lommel
The existing hash function didn't work well with Set's method of
masking to the lower bits, because many verts have zeros in the
lower bits.
Also, replaced VectorSet with Set for Vert deduping.
This provide a simple implementation for basic math operations. Add,
subtract, multiple, and divide are implemented so far. Domain
interpolation is not implemented yet either.
This commit also adds two "Type" options to the node for choosing
explicitly whether to use an attribute or a float value as one of the
inputs. It is not possible to use two floats as inputs, because that
would be a regular math node.
Some cleanup to avoid some code duplication will come in a future
commit.
This node creates a new geometry based on two incoming geometries.
In the future the node ui can be extended to support more than two inputs.
The output geometry will have all the attributes of the inputs.
If both inputs have non-matching set of attributes, missing data is either
interpolated, type cast or zeroed.
Ref T82786.
* It now knows its own component type.
* Add virtual method to check if the component is empty.
* Add virtual methods to find all attribute names.
* Support adding existing components to a geometry set.
Recent changes introduced `acc` parameter into the texture read
functions. When nanovdb isn't enabled this leads to compilation errors
as the `acc` variable wasn't defined. OpenCL only compiles needed
features what made it more prominent.
Reviewed By: Patrick Mours
Differential Revision: https://developer.blender.org/D9629
ROCm 3.9 already defined `NULL`. This patch will first check if it was
already defined to remove compilation warnings.
NOTE: This doesn't add official support for ROCm as it still fails to
render correctly (crashes with default cube).
Reviewed By: Brecht van Lommel
Differential Revision: https://developer.blender.org/D9610
Split some of the code of `graph_edit.c` into:
* `graph_view.c`: preview range, view all, view selected etc.
* `graph_slider_ops.c`: the decimate modal operator code.
The latter file will be extended later with more slider-based operators.
Maniphest Tasks: T81785
Reviewed By: sybren
Differential Revision: https://developer.blender.org/D9312
This new discontinuity filter performs actions on the entire Euler
rotation, rather than only on the individual X/Y/Z channels. This makes
it fix a wider range of discontinuities, for example those in T52744.
The filter now runs twice on the selected channels, in this order:
- New: Convert X+Y+Z rotation to matrix, then back to Euler angles.
- Old: Add/remove factors of 360° to minimize jumps.
The messaging is streamlined; it now reports how many channels were
filtered, and only warns (instead of errors) when there was an actual
problem with the selected channels (like selecting three or more
channels, but without X/Y/Z triplet).
A new kernel function `BKE_fcurve_keyframe_move_value_with_handles()` is
introduced, to make it possible to move a keyframe's value and move its
handles at the same time.
Manifest Task: T52744
Reviewed By: looch
Differential Revision: https://developer.blender.org/D9602
The caller is still responsible for allocating list of markers,
but the track allocation and initialization can now be reused.
Currently no functional changes, preparing for an upcoming
development.
Having up to nine levels of indentation make this function hard to
follow. Instead of indenting the rest of the loop for a simple special
case, just continue.
By using floating point filters, the speed improves by a factor of 2 to 10.
This will help speed up some cases of the Exact Boolean modifier.
Changed the interface of mpq2::isect_seg_seg to not return mu, as it was
not needed and not calculating it saved 15% time.
This changes the meaning of the order of the two geometry input sockets
to match the existing boolean modifier. The first socket is now the base
mesh and the second socket is the "cutter" geometry.
Similar to how the node editor context works for materials, this commit
makes the node group displayed in the node editor depend on the active
object and its active modifier. To keep the node group from changing,
just pin the node group in the header.
There are still missing updates when switching the modifier's node group.
This exposes the operator to set the active modifier as the type icon
in the header of the modifier panels. Tweaks to the widget drawing
code were necessary to use the red alert for non-emboss operator
buttons. Then, the panel for the active modifier gets a border around it
(which currently uses the property search theme color), requiring
an "active property" field in the PanelType struct.
Previously, it was always outputting an empty geometry in that case.
This was confusing because e.g. the output of a union operation
with a single non-empty mesh is also non-empty.
This will probably not stay exactly like this, because it looks a bit ugly.
Adding this now, because it increases flexibility a lot and allows me
to test the attribute api a bit more.
Originally, there were technical reasons for why two different domains
were used. Those reasons have been lost unfortunately, and there
don't seem to be any issues with combining the domains now.
This reduces confusion when users have to select the domain type
in enums.
When lowering the wireframe opacity with sculpt overlays enabled, the
wireframe overlay was creating white artifacts along the edges.
Reviewed By: fclem
Differential Revision: https://developer.blender.org/D9607
Attributes can now be accessed through a GeometryComponent.
MeshComponent and PointCloudComponent implement the specific
attribute access mechanisms using virtual methods.
This simplifies code that should work on different geometry components
quite a bit, because it only has to be written once without using templates.
The Random Attribute node benefits from that a lot.
This adds an invert toggle for the outliner object state filters.
There are some cases where we want a filter for invertable states (Selected,
Unselected) and having a single toggle to invert the filter reduces the
number of separate filter types needed. This removes the "Hidden" filter
which can now be replicated with an inverted "Visible" filter.
Differential Revision: https://developer.blender.org/D9598
Although disabling this option is equivalent to setting the edge angle
to 180 degrees, there is no great reason to change the settings from what
is exposed in the modifier.
Differential Revision: https://developer.blender.org/D9564
In case we do not use names, code adding new insert operations in
collections was broken.
Not a proble in practice so far, since this case was not yet in use, but
will be soon with NLA overrides.
Since add-ons now unregister on exit
(as of fa566157a5)
clearing functions in `bpy.app.handlers` caused an error on exit.
Resolve by restoring handlers before exiting.
When an projecting onto a plane that is orthogonal to the views Z axis,
project onto a view aligned axis then map it back to the original plane.
see: ED_view3d_win_to_3d_on_plane_with_fallback
Since the depth can't be properly visualized in 3D, display a 2D so
it's possible to to tell the depth from the cursor motion.
When using surface orientation and nothing is under the mouse-cursor,
pick the axis that's closest to the view's Z axis.
Now by default default, drawing into empty spaces wont use a plane
orthogonal to the view.
Generating the thumbnail left the view matrices set to values
that couldn't be used for cursor drawing.
Backup/restore the matrices for off-screen drawing.
The Normal Map node was falling back to (0, 0, 0) when it was missing
the required attributes to calculate a new normal.
(0, 0, 0) is not a valid normal and can lead to NaNs when it is
normalized later in the shader. Instead, we now return sd->N,
the unperturbed surface normal.
The random attribute node currently doesn't support every attribute
domain or data type. With a bit of boilerplate code we can remove these
unsupported enum items from the UI for now, and also get a system to
remove unsupported items from the UI in the future. If all attribute
creation nodes end up always supporting all attibute types and domains,
this can be removed.
There will be an operator to access this property instead, which provides
a better name and tooltip to expose in the UI, and makes it more clear
that the property is dependent on the other modifiers.
This commit adds functions to set and get the object's active
modifier, which is stored as a flag in the ModifierData struct,
similar to constraints. This will be used to set the context in
the node editor. There are no visible changes in this commit.
The issue was that the same point cache was read by multiple
threads at the same time (the same object was evaluated for
render and for the viewport).
Both threads incremented PTCacheMem->cur which lead to the crash.
The fix is to remove the PTCacheMem->cur and store it on the
stack instead. This way every thread has its own cur.
Reviewers: brecht
Differential Revision: https://developer.blender.org/D9606
Cover all atomic functions with unit tests.
The tests are quite simple, nothing special so far. The goal is to:
- Make sure implementation exists.
- Implementation behaves the same way on all platforms
(We had issue when MSVC and GCC were behaving differently in the
past).
- Proper bitness is used for implementation and non-fixed-size
function implementation.
The tests can be extended further to make sure, for example, that
CAS operations do not cast arguments to a more narrow type for
comparison. Considering it a possible further improvement, as it is
better be done being focused on that specific task.
There is an annoying ifdef around 64bit implementation, which uses
same internal ifdef as the header does. This check is aimed to be
removed, so is easier to simply accept such duplication for now.
The tests seems somewhat duplicate for signed/unsigned variants and
things like this. The reason for that is to keep test code as simple
as possible: attempting to do something smart/tricky in the test code
often causes the test code to be a subject of being covered with its
own unit tests.
Differential Revision: https://developer.blender.org/D9590
Previously the lock-free tests were actually testing guarded allocator
because the main entry point of tests was switching allocator to the
guarded one.
There seems to be no allocations happening between the initialization
sequence and the fixture's SetUp(), so easiest seems to be just to
switch to lockfree implementation in the fixture's SetUp().
The test are passing locally, so the "should work" has high chance
of actually being truth :)
Differential Revision: https://developer.blender.org/D9584
Previously the only way to use lockfree implementation was to start
executable and never switch to guarded allocator.
Surely, it is not possible to switch implementation once any allocation
did happen, but some tests are desired to test lock-free implementation
of the allocator. Those tests did not operate properly because the main
entry point of tests are forcing guarded allocator to help catching
bugs.
This change makes it possible for those tests to ensure they do operate
on lock-free implementation.
There is no functional changes here, preparing boilerplate for an
upcoming work on the allocator tests themselves.
Don't refresh the list of sockets, so that when the .blend file is restored the
links remain valid. Also display such nodes in red to indicate an error, same
as when the node type info is missing.
Some DNA headers already did this, most did not. Even though many of them would
be included in C++ files and thus compiled as C++. This would be confusing and
developers may think they have to add `extern "C"` too a whole lot of
(indirect) includes to be able to use a C header in C++.
However, this is a misconception.
`extern "C"` does not cause code to be compiled with C rather than C++! It only
causes the linker to not use C++ function name mangling. See
https://stackoverflow.com/a/1041880.
Because extern DNA headers don't have function declarations, using `extern "C"`
actually should not have any effect. On the other hand, adding it causes no
harm and avoids confusion. So let's just have it consistently in C header
files.
Differential Revision: https://developer.blender.org/D9578
Reviewed by: Bastien Montagne, Sybren Stüvel
Don't allocate StripElem for movieclip, scene and mask strips. This
struct is not handled in seq_dupli function. This caused field to be
uninitialized in COW datablock.
StripElem is not allocated when adding strip with operator and it is
not needed for these strip types.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D9600
Split up the Euler filter function into two more functions. The Euler
filter operator works in two stages (find channels that define the X/Y/Z
Euler rotations, and perform filtering on those channels), and each
stage now has its own function. This makes it clearer which data are
used in which part of the code, and makes future improvements easier.
No functional changes.
This is actually a user interface issue, introduced on 0688309988.
but->tip is checked for NULL in other parts, and rightly so, since here
it crashes Blender.
On a side note I don't know how to reliably reproduce the original bug
whose fix introduced this issue. That got on the way of seeing whether
this is the correct fix, if it happens outside greasepencil operators.
Maybe it is an operator that is missing a tooltip? Impossible to tell
without more information from said commit. That said since this happened
during the end of bcon3 and we are not in bcon4 I'm going ahead with
this NULL check commit.
Patch reviewed by Philipp Oeser and Sergey Sharybin.
Code was rewritten to work on per-sequence basis returning reference to
created strip.
There should be no functional changes.
Selection logic is left as is for now. I could simplify it, but it
belongs to operator, not split logic.
Reviewed By: sergey
Differential Revision: developer.blender.org/D9592
My last cleanup commit for this function missed this case. It likely
happens because the panel's block size doesn't update properly somewhere.
Short of investigating that right now, it makes sense to return early
in this case anyway.
This option joins any stroke with an end near the actual stroke. Now it is not limited to the last stroke, any stroke in the same layer for the actual frame can be joined. The join can join two strokes drawing a third stroke.
If the end and the start of the result stroke are very small, the stroke is changed to be cyclic automatically.
There is a limit distance to join the stroke, if the distance is greater than this value, the strokes are not joined. Actually, a constant, threshold distance is used, but we could expose
as a parameter in the UI in the future.
The tool can be used with freehand drawing or with primitives.
Note: Great part of the patch is just a refactor of the old code to make it accessible and to keep code organized.
Reviewed By: mendio
Maniphest Tasks: T82377
Differential Revision: https://developer.blender.org/D9440
The existing panel drawing function was a bit convoluted with dependent
conditions in different scopes, redundant and unecessary computations,
and un-helpful naming.
This commit separates the function into two parts, the backdrop and the
widgets. It also improves naming and uses const where possible, and in
general cleans up the code.
There are some slight visual changes, mostly with the placement of the
drag icon, which moves a bit downward to be centered with the triangle
icon. The black rectangle displayed while dragging is also removed.
The logic for separator-spacers (used here for right-alignment) didn't take
region scaling into account. Usually that's not an issue because they are
otherwise only used in headers which can't zoom.
The logic for separator-spacers (used here for right-alignment) didn't take
region scaling into account. Usually that's not an issue because they are
otherwise only used in headers which can't zoom.
During some operators like rotate in grease pencil edit mode the
hierarchy lines in the outliner would draw twice as thick. Set the width
before drawing the lines in the outliner.
Differential Revision: https://developer.blender.org/D9589
This is an addendum to previous boolean fix, where the object
transformation was "cleaned". Now the operand one is too.
This fixes the issue shown in the video in T82301 when you move
a column around the XY plane with the top and bottom faces
supposedly coplanar with a cube. The transformation matrix when
you do that has a tiny offset in the z component.
Scaling of forces needs more work. Before making changes to them it would be nice to have a setup, that works physically correct across multiple modifiers (cloth, rigid bodies, fluid).
This will be a to do for 2.92.
Active tile could be NULL when it was on the second tile before
switching back and forth between the Image/UDIM.
In the future we might also check that the active_tile_index is always
valid.
This adds some ui boilerplate code for the Random Attribute node and
provides an initial implementation.
Note, while the implementation can already randomize attributes, it might
not behave as expected under all circumstances yet. It's still work in progress.
The root cause of this bug is that the function that updates the PBVH
normals is drw_sculpt_generate_calls. As now both the overlays and
mesh can be drawn without using pbvh drawing, the normals were not
updating. This patch forces a normals updates also in the no PBVH
drawing code path of the overlays. This was affecting both shading and
sculpt surface sampling in both flat and smooth shading modes.
Having the sculpt normals being updated by the drawing code is a wrong
design which also causes other issues like:
Brushes that sample the surface and do multiple stroke steps between
redraws will sample invalid normals, creating artifacts during the
stroke clearly visible in some brushes.
Brushes that do not need to sample the surface update the normals on
each redraw. This affects performance a lot as in some cases, updating the
normals takes more time than doing the brush deformation. If flat shading
is being used, this is only necessary to do once after the stroke ends.
Reviewed By: fclem
Differential Revision: https://developer.blender.org/D9535
I will have to do a couple more iterations on this api in the upcoming weeks,
but for now it should be good enough.
The API adds an indirection for attribute access. That has the following benefits:
* Most code does not have to care about how an attribute is stored internally.
This is mainly necessary, because we have to deal with "legacy" attributes
such as vertex weights and attributes that are embedded into other structs
such as vertex positions.
* When reading from an attribute, we generally don't care what domain the
attribute is stored on. So we want to abstract away the interpolation that
that adapts attributes from one domain to another domain (this is not
actually implemented yet).
Accessing attributes through this indirection does have a performance penalty.
In later iterations of this API I want to reduce this penalty and extend the API
so that performance critical code does not have to go through the indirection
for every attribute access.
Other possible improvements for later iterations include:
* Actually implement interpolation between domains.
* Don't use inheritance for the different attribute types. A single class for read
access and one for write access might be enough, because we know all the ways
in which attributes are stored internally. We don't want more different internal
structures in the future. On the contrary, ideally we can consolidate the different
storage formats in the future to reduce the need for this indirection.
* Remove the need for heap allocations when creating attribute accessors.
Pie menu had inconsistent behavior for dragging & releasing the key
compared to clicking on the button.
This was caused by the `onfree` argument being set to true,
preventing the button from running callbacks such as setting
up undo data & auto-keyframe
This argument should only be used when freeing the button,
set this to false as is done for regular menus.
Until there is a icon made specially for this, the nodetree icon is up
for grabs. Using it in the nodegroup + modifier + editor helps the users
to make a connection on where to edit those modifiers.
The point cloud object is the only one that will support instancing at
first. So we can expose it as a regular object.
It is limited since it has no edit mode. But this is not different than
the volume object.
As part of preparing this work for master, with an initial subset of
working functionalities the Geometry Nodes Editor can be exposed for
everyone.
(it also helps people that are testing the branch)
This brush needs to be disabled for dyntopo as it stores its custom data
and deforms from original coordiantes.
Reviewed By: sergey
Maniphest Tasks: T82542
Differential Revision: https://developer.blender.org/D9516
Panels for active uiBlocks always have a type, because the process that
makes them uses the types. Add an assert just to make it clear that the
assumption is purposeful.
Regression introduced by {b17cca6966}. When centralizing the gpu texture
premultiplication setting it was assumed that generated images
(`IMA_TYPE_UV_TEST`) were stored as premultiplied. That assumption was
totally wrong as the alpha association is determined by the existing of
the float/byte buffer.
NOTE: This change will render generated images with pure emissive
colors (show colors when alpha=0.0) what might add more reports. Any
reports could be merged in the next report {T82790}.
Reviewed By: Clément Foucault, Philipp Oeser
Differential Revision: https://developer.blender.org/D9585
After recent changes to the context panel layout (rB187cc5e26d28b1a8),
there has been an error printed when running propery search:
> Error: separator_spacer() not supported in popups.
The layout code thinks it's drawing in a menu because region->visible
isn't properly set for the other tab searches. This patch sets that field
for the temporary searching region, but it also disables searching in the
context breadcrumbs panel, because at best this will just give results
for the names of the active object, etc. This isn't helpful since
those labels are mostly in every tab anyway.
Differential Revision: https://developer.blender.org/D9425
Logic was incorrect, mistake in f3b8792b96.
Updated comment to make intent more clear.
Same as fd78f8699e, but needed conflict resolution as the panel-type
flags where renamed in master.
Edges with 3 or more connected UV's caused UV pack to fail.
Instead of using functions from uvedit_parametrizer.c which are intended
specifically for ABF/LSCM unwrapping, use a simpler method for packing
which stores arrays of BMesh faces.
The clone tool in the image editor can show a second texture on top
of the image. This wasn't ported and now results into alpha and depth
issues. This fix adds the clone tool drawing to the overlay engine.
Reviewed By: Clément Foucault
Differential Revision: https://developer.blender.org/D9352
Remove redundant call to `ofstream::close()` from `~PSStrokeRenderer`
and `~TextStrokeRenderer`. ofstream will be destructed automatically.
- For `~Depsgraph`, `std::function`'s constructor can throw.
- Passing throwing statements in the lambda will not be detected by
clang-tidy.
Fix these issues by using lambda as function argument.
Reviewed By: sergey, sybren
Differential Revision: https://developer.blender.org/D9497
In 8d1978a8e0 bmain entry is updated when prefetching is started,
but this must be done before seq_prefetch_update_context(). Otherwise
created cache keys will be incorrect.
This happens immediately after prefetch is started. When it's refreshed
problem goes away.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D9581
Some RNA structs, like operators or keymaps, are not allowed to have ID
pointer properties. now this check will ignore those, and report an
error message in the console.
Related to T82597.
Notes: While a bit more involved than rBf39fbb3e6046, this commit
remains fairly localized and non-intrusive.
It relies on some rather obscure and weird behaviors of our RNA code
though, a cleaner solution could be e.g. to add a tye to
`StructOrFunctionRNA`, so that we could properly 'rebuild' (re-cast) the
pointer to either `StructRNA` or `FunctionRNA` when needed in internal
code...
When offsetting strips to the left, to make space when syncing strip
length, move the preceding NLA strips instead of the succeeding strips.
The bug seems to be from a copy/pasting typo.
Reviewed By: sybren
Differential Revision: https://developer.blender.org/D9583
Select only editable keys on selected curves when "Only Show Selected
Curves" is enabled in the Graph Editor's View menu.
This prevents selecting invisible keys with box, circle, and lasso
select in the Graph Editor.
Reviewed By: looch
Differential Revision: https://developer.blender.org/D9580
Crash is related to the definition of the GL_MAX_TEXTURE_SIZE. OpenGL does not clearly
defined `GL_MAX_TEXTURE_SIZE` exactly means. Both on AMD and NVIDIA we have issues with
huge textures that they don't get created even if they are smaller. (See {D9530} for
research).
This patch will try to create the texture in a smaller size when the texture creation
failed.
Final implementation by: Clément Foucault
We should create a solution that doesn't need downscaling. For this specific case ARB_sparse_texture might help to create cleaner code, but you still have to commit the whole image what introduces several draw calls. Other improvement is to optimize the scaling; current implementation isn't optimized for performance.
Reviewed By: Clément Foucault
Differential Revision: https://developer.blender.org/D9524
This is was caused by incorrectly set preview_render_size in VSE
rendering context. Value was set to SEQ_PROXY_RENDER_SIZE_FULL, but
it should be SEQ_PROXY_RENDER_SIZE_SCENE as scene render size is
being used.
This is same fix as 0d7036b40e, but I did not checked openGL
render pipeline.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D9562
Use either scene render size or fixed preview scale factor.
Previously scene render size was used as baseline value for text size
correction. This is incorrect.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D9563
The icons originally chosen for the collection colors were selected
during development and had a few issues with contrast in the light
theme, and the gray color was not a good choice against the default gray
backgrounds.
The new colors are more readable in both default Blender themes. Gray
was replaced with pink.
Differential Revision: https://developer.blender.org/D9504
Coalescing on macOS overwrites a singular unprocessed mouse event. To
receive all mouse and tablet events coalescing is disabled.
Disabling coalescing for macOS disables coalescing for trackpad
gestures. Repeat trackpad events are unnecessary and found to
negatively impact performance thus are re-coalesced in Window Manager.
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D9574
This was introduced by rBdb7d8281c5a2.
The color needs to be premultiplied as there is no blend mode and
the output color is replacing the framebuffer color & alpha.
The OpenVDB data structure can store voxel data in leaf nodes or tiles
when all the nodes in a given region have a constant value. However,
Cycles is using the leaf nodes to generate the acceleration structure
for computing volume intersections which did not include constant tiles.
To fix this, we simply voxelize all the active tiles prior to generating
the volume bounding mesh. As we are using a MaskGrid, this will not
allocate actual voxel buffers for each leaf, so the memory usage will be
kept low.
Differential Revision: https://developer.blender.org/D9557
Reviewed by: brecht, JacquesLucke
X axis panel dragging traces back to Blender versions before 2.5,
where panels could be aligned horizontally. But for many years now
panels have been vertically aligned. Considering this, keeping the
X axis dragging around is a bit odd. It makes interaction confusing,
or at least more complicated. It also looks bad, since any part of
the panel outside the region is cropped.
Differential Revision: https://developer.blender.org/D9549
This reverts commit f39fbb3e60.
Code is not valid, `DefRNA.laststruct` does not always point to the
proper struct when defined from Python, need to be done differently.
The original code for viewlayer collection flag syncing across moves
from D9158 didn't consider the case where the collection could no longer
be found in its original view layer (moving a collections betwen scenes).
The fix is to just check if the collection starts in the same scene as
it will be moved to before trying to do the flag syncing. I thought about
this for a while and tried a couple other solutions, but I couldn't come
up with a proper way to support syncing the layer collection flags across
scenes without making too many changes.
Differential Revision: https://developer.blender.org/D9568
Root of the issue is that point caches are added to the object's list
on-demand, which often ends up with them being added only during
depsgraph evaluation, i.e. on COW objects.
This could result in having 'orig' data caches with invalid/unset stack
index at some points (e.g. when reading a file and applying
liboverrides), leading to discarding valid existing disk cache files.
Fact that one of those index is signed, and the other not, does not
help...
While this is very weak, fixing broken PointCache code is out of the
scope of a bug fix, so this patch merely:
* Simplifies and factorizes the code generating the 'extension' part of
caches filenames;
* Ensures `BKE_object_insert_ptcache` is called when needed so that we
always have a valid stack index to generate that filename extension.
This is only a bandaid, but it is simple and should be safe enough for
now.
Related to T82503.
Some RNA structs, like operators or keymaps, are not allowed to have ID
pointer properties. now this check will ignore those, and report an
error message in the console.
Related to T82597.
Remove `return` from for-loop which blocked the recalculation of driven
values when it found the first driver.
Reviewed By: sybren, sergey
Differential Revision: https://developer.blender.org/D9515
Previously, GeometrySets and GeometryComponents has reference
counters and could be shared. This commit changes it so that
only GeometryComponents are shared. A GeometrySet is a fairly
small type that is cheap to copy.
A lot of code simplifies when we can assume that GeometrySet
is cheap to copy.
Previously the return value of `ufopen` wasn't checked and if it failed,
`NULL` was passed into `fclose()` which resulted in a crash. This patch
avoids this by returning from `BLI_gzopen` when the file cannot be created.
Reviewed By: sebbas, iss
Differential Revision: https://developer.blender.org/D9576
is present
Caused by rB4212b6528afb.
'updateGLSLCurveMapping()' compares cacheIDs and in certain scenarios,
these are the same when they should not.
- whenever we had multiple viewports that are colormanaged with
curvemappings this worked right (cacheIDs were different)
- for example, this also worked right when the ImageEditor displays a
Render Result or a Compositor Viewer
- but it worked wrong when the Image Editor displays any other Image (or
no Image at all)
- it also worked right if there were multiple Image Editors [and one of
them displays a Render Result e.g]
Now why is this so?
For comparison, the curve mapping's pointer/address is used.
- update_glsl_display_processor frees the curve_mapping, see
BKE_curvemapping_free(global_glsl_state.curve_mapping)
- similar, update_glsl_display_processor creates a new curvemapping, see
BKE_curvemapping_copy(view_settings->curve_mapping)
- now for the situation that a viewport with curvemapping and a viewport
without curvemapping is present and you make changes to the curvemapping
the following happens:
-- curve_mapping_settings->cache_id is set once [to the memory address
of curvemapping before change]
-- change happens
-- viewport 1 frees curvemapping
-- viewport 2 duplicates using BKE_curvemapping_copy, but this one gets
the same address like before the change
-- this means we have different data on the same address with the same
cacheID...
Solution: to really make the cache ID unique we can combine the pointer
with its 'changed_timestamp' [which increases on every change].
Reviewers: jbakker
Maniphest Tasks: T82460
Differential Revision: https://developer.blender.org/D9559
Brecht and I decided to move the GeometrySet classes
to the global namespace for now. This way we can use
the same type in C and C++ and don't have to use
reinterpret_cast as often.
The issues was that Blender was trying to refresh the
splash screen region. However, opening the file browser
closed the splash screen and freed the region.
The fix is to simply not refresh the region.
Has been approved in T81817.
First detect the other selected object, then check it can be written to.
Otherwise the target object could be the first one found when
looping over objects which is random from the user perspective.
Move the type check to the operator, which also checks the data isn't
library data which was being ignored.
Move most functions into separate files.
File sequencer.c is reserved for alloc, free and duplicating functions.
There should be no functional changes.
Two problems were fixed. One, the code for dissolving vertices
left a face around if dissolving a vertex would leave less than
three vertices. Instead, the face should be deleted.
Two, with transformations like "rotate 180 degrees", this should
be no problem with exact, but the current transformation matrix
has very small non-zero entries where it shouldn't. Cleaning the
transformation matrix makes it more likely that user expectations
about coplanar faces will be fulfilled.
More than output for the modifier's node group is fine, because the
node group may be used in other situations, but the modifier will
not work if there is no geometry output socket.
It should not be possible to set the scene collection's color tag
through rna. Also adds a missing notifier for setting the collection
color tag from python.
After rB452a1c7b3838 there were still a few cases where the old
collection icon was used in the interface. Replace these with the new
filled collection icon.
Bypassing the macro that creates modifiers to allow the enum to be
referring to the Empty modifier, but the internal struct in the code to
be Nodes.
Also fix the name of the new created modifiers to be "Empty".
Fix was built with rBb458ea6b23381a9acb90dbbd73ced678e1d404c5.
Leave the internal DNA data struct as "Nodes", and only rename the
modifier and user interface to Empty.
See 4a2734a835 for the original take and the reasoning
for renaming from Nodes to Empty.
This diff remove "Combine Strings" and "Group Instance ID" nodes from the UI (but keep them in the source code to be able to bring them back if needed).
This is part of "Cleanup 1st Sprint Nodes" (https://developer.blender.org/T82370).
Differential Revision: https://developer.blender.org/D9558
This introduces a regression. Users need to re-create their modifiers
and set their values again. The NodeGroups are intact so all they need
is to create new "Empty" modifiers and set the node group.
---
Original problem: Eventually most modifiers will be node-based, so naming
this new modifier "Nodes" will get outdated pretty quickly.
Calling it "Empty" it's fairly descriptive, since the node tree simply connects
the modifier Input to an Output, without any effect.
There is a potential problem, that it could be associated with the
Empty object type. Like some sort of Hook.
This is the task T82700, and the design task T82537.
Differential Revision: https://developer.blender.org/D8660
This patch is the result of the GSoC 2020 "Editing Grease Pencil Strokes
Using Curves" project. It adds a submode to greasepencil edit mode that
allows for the transformation of greasepencil strokes using bezier
curves. More information about the project can be found
here: https://wiki.blender.org/wiki/User:Filedescriptor/GSoC_2020.
The overlap with the `Panel` flags that start with "PNL" was quite
confusing because wasn't clear which enum a flag was from. The
new names are a bit longer, but the clarity is worth it.
Instead to use the ID of the object, now the parameter is an Enum with Selected object or New.
If use selected mode, the first grease pencil object selected is used. If none of the selected objects is a grease pencil object, a new object is created.
Small cleanup changes to the original patch.
Differential Revision: https://developer.blender.org/D9529
LibOverrides only support a small sub-set of PointCache features for now
(one cannot add new caches, baking in memory is not supported...).
Part of first step of T82503: support disk cache in liboverrides.
This was done as some sort of safety, but should not actually be needed,
and including tags like `ID_RECALC_POINT_CACHE` e.g. makes usage of
point caches impossible with liboverrides (since it would systematically
invalidate all cache on file load).
In theory we should not have to tag anything here in fact, RNA accessors
are supposed to take care of it, but for now we keep the
`ID_RECALC_COPY_ON_WRITE` one.
Part of first step of T82503: support disk cache in liboverrides.
Note that due to convoluted layout of point caches in RNA (active one
also storing list of all available ones), we'll often have the
pointcache overrides rules twice. Should not be a huge problem,
practically speaking.
Part of first step of T82503: support disk cache in liboverrides.
We already were using one of earlier RC of the library, so there is no
expected big changes. Just making the update official, using official
version and stating it in the readme file.
Becomes rather annoying to duplicate them across C/C++ GCC/Clang sets,
almost as if the test should test both C and C++, and to do it for all
compilers.
Solves strict warning in the upstream of Ceres library.
This reverts commit 9d172f007e.
Got a second thought and remembered why it was not done in the first place.
The issue here is that the server needs to communicate codesign result back
and that must happen within the new protocol. So if the client talks old
protocol it is possible to receieve data from it, but is not possible to
communicate result back to it.
Image format code checked the file type against an enum except for
zero which is used when the format can't be detected.
Also add doc-strings to some of the image file type callbacks.
the function vec_roll_to_mat3_normalized() has a bug as described in T82455. This Differential is only for refactoring the code such that it becomes more clear what the function does and how the bug can be fixed. This differential is supposed to not introduce any functional changes.
Reviewed By: sybren
Differential Revision: https://developer.blender.org/D9410
HDRI preview should have resolution dependent on dpi, viewport scale and HDRI gizmo size.
This patch uses a LOD to render a more round sphere.
Reviewed By: Jeroen Bakker
Differential Revision: https://developer.blender.org/D9382
This would allow python script to access `lineWidth` uniform when drawing lines
without using `glLineWidth`.
Reviewed By: Jeroen Bakker
Differential Revision: https://developer.blender.org/D9518
Standard deviation formula wasn't being applied correctly when selecting
R G B cases. Issue is there since Blender 2.64 as it was incorrectly
ported over from the previous compositor.
Reviewed By: Sergey Sharybin, Jeroen Bakker
Differential Revision: https://developer.blender.org/D9384
This patch adds an opacity slider to the wireframe overlay. The previous
wireframe in dense geometry scenes could be too dark and sometimes the
user just wants an impression of the geometry during modelling.
Reviewed By: Jeroen Bakker
Differential Revision: https://developer.blender.org/D7622
This commit moves some of the logic around so that the logic in
panel_activate_state is clearly separated by the state being activated.
There are fewer nested and redundant checks, and it's easier to see
the progression of interaction with the panel handler.
This callback made some sense before moving the file-type information
from a bit-flag to an enum: e142ae77ca
Since then, we can compare the type value directly.
Also replace loops over file types with IMB_file_type_from_{ibuf/ftype}.
This commit makes the property editor switch to the physics tab instead
of the modifier tab when selecting physics modifiers. Since the modifier
isn't visible then, it's confusing to change the expansion, so this commit
also disables the modifier expansion for these modifiers.
Differential Revision: https://developer.blender.org/D9544
Now that the Blender sync mechanism deletes nodes from the scene, we need to
ensure scene update is stopped before we do this.
Also add some more early out in scene geometry update to ensure we do not
continue working on incomplete geometry data, though that was not the cause of
this crash.
Two issues:
* Automatic deduplication of OpenVDB grid data was failing when Cycles had
already cleared the OpenVDB grid, causing an empty grid. Instead rely on
Blender return the same OpenVDB grid pointer when deduplication is possible.
* The volume bounds mesh was not properly cleared when the OpenVDB grid was
empty, causing a mismatch between mesh and voxel data.
The function vec_roll_to_mat3_normalized() basically has to handle 3 scenarios:
- When a bone is oriented along the negative Y axis
- When a bone is very close to the negative Y axis
- All other cases
The tests in the Differential make sure that all 3 situations are covered.
Reviewed By: sybren, mont29
Differential Revision: https://developer.blender.org/D9525
This implements a workaround for the issue that (for historical reasons)
the names of vertex groups are stored on the object while the actual vertex
group data is stored on the mesh.
The solution is to copy the vertex group names from the object into
the `MeshComponent` so that the information is not lost, when the
object cannot be accessed.
Previously, the execution function of a geometry node has three parameters.
Now it has only one. This makes it easier to pass more information to the
execution function, that might only be used by a few nodes, because we
don't have to add more parameters that are unused in most cases.
Only a single DEG operation node `POSE_SPLINE_IK_SOLVER` should
be added in this case [ see `build_splineik_pose`, same is already done
for overlapping IK in `build_ik_pose`]
ref T82347.
Reviewers: sybren
Maniphest Tasks: T82347
Differential Revision: https://developer.blender.org/D9471
* Support vertex weights to control density.
* O(n) performance instead of O(n^2).
* More stable when density weights are changed.
The vertex group has to be specified using an index for now.
This is a technical limitation that will resolved a bit later.
Objects can evaluate to a geometry set instead of a single ID (only point cloud
objects for now). In the depsgraph object iterator, those geometry components
are expanded into temporary objects.
It's important to note that instanced objects can also contain geometry
components. Therefore, they have to be split up into multiple objects
as well in some cases.
At a high level the iterator works like so:
```
for object in depsgraph:
for component in object:
yield object_from_component(component)
for dupli in make_duplis_list(object):
for component in dupli:
yield object_from_component(component)
```
DEG_iterator_objects_next has been cleaned up, to make this structure
a bit more apparent.
This changes the signature of the modifyPointCloud function.
I'm doing that instead of making a new callback, because it requires
changes to significantly fewer files. Eventually it would be good
combine modifyMesh, modifyHair, modifyPointCloud and modifyVolume
into one modifyGeometrySet.
I temporarily disabled the displacement only modifiers for point clouds.
Support can be added back a bit later. I assume those have not been
used anywhere anyway.
The output of point cloud modifiers can not only be another point cloud,
but also a mesh and/or some instances. I added a new geometry_set_eval
field to Object_Runtime. For point cloud objects, the final geometry is
now referenced by that pointer instead of data_eval. The data_eval field
is still initialized after modifier evaluation to make some other code happy.
The evaluated geometry set is not yet passed to the renderer, so a point
cloud is currently rendered empty.
This caused warnings when e.g. building the lite profile because NanoVDB was not disabled, but
OpenVDB was. This Fixes this by setting the "WITH_NANOVDB" flag too.
This node takes an object as input and outputs its location, rotation,
scale and geometry. Right now the loc/rot/scale are extracted from the
objects `obmat`. The geometry is just the mesh in the local space of the
source object.
We will likely need some more control over space transformations
using enums in the node, but those can be added a bit later.
The fundamental difference between object sockets and the
other existing data sockets is that an object is an ID data block.
Changing the value of an object socket also changes the depsgraph.
The modifier has to analyse the node tree to figure out which other
objects it depends on. Currently, this is done very simply by just
looping over all sockets and collecting the objects. In the future
this can be improved by also figuring out what components of
an object are needed.
Instead of passing object pointers around in the node tree, we actually
use a handle. This handle is just a number internally that identifies
a specific object. The conversion between handles and object pointers
is done using a map that is provided by the modifier.
This approach has a couple of benefits. It protects us a bit from passing
around pointers that are not known to the modifier and therefore are
not in the depsgraph. Furthermore, the object pointer can change
while the handle stays the same. This is not important right now, but
is not unlikely to become useful in the future.
The API for how nodes access object pointers is not ideal yet and
will be improved in the future.
Add a new property `co_ui` to Keyframes, the modification of which will
apply to the keyframe itself as well as its Bézier handles.
Dragging the "Keyframe" slider in the properties panel now maintains the
deltas between the keyframe and its handles, just like moving the key in
the graph editor would.
Reviewed by @sybren in T81813.
This commits implements multiple changes:
* Adds a simple C API.
* Improves the ownership handling by introducing GeometryOwnershipType.
* Adds an InstancesComponent that stores positions and Object pointers.
This is more in line to other eyedropper usages throughout blender.
Affected operators:
- Sample Dyntopo detail
- Extract Face Set (as reported in T82615)
ref T82615
Maniphest Tasks: T82615
Differential Revision: https://developer.blender.org/D9531
Pass codesign errors (if any) from codesign buildbot server to the
buildbot worker, so that the latter one can abort build process if
the error happens. This solves issues when non-properly-notarized
DMG package gets uploaded to the buildbot website.
When creating an armature from the skin modifier, resulting bones would
always be flagged BONE_CONNECTED.
Those bones cannot be transformed (just rotated).
Now only flag bones that really have a parent BONE_CONNECTED.
Maniphest Tasks: T82624
Differential Revision: https://developer.blender.org/D9534
Issue exposed by rB4c7b1766a7f1.
Main idea is that non-memfile first undo step should check into previous
memfile and tag the ID it is editing as `future_changed`.
That way, when we go back and undo to the memfile, said IDs are properly
detected as changed and re-read from the memfile.
Otherwise, undo system sees them as unchanged, and just re-use the
current data instead.
Note that currently only Sculpt mode seems affected (probably because it
is storing the mode switch itself as a Sculpt undo step instead of a
memfile one), but similar action might be needed in some other cases
too.
Maniphest Tasks: T82388
Differential Revision: https://developer.blender.org/D9510
Pass codesign errors (if any) from codesign buildbot server to the
buildbot worker, so that the latter one can abort build process if
the error happens. This solves issues when non-properly-notarized
DMG package gets uploaded to the buildbot website.
When painting in the image editor on data images (Non-color, Raw) the
color mismatched between the sampled color and the actual effect that
the painting has on the image. The root cause is that the sampling is
color managed, but the painting still uses a fixed color management
pipeline with a lot of assumptions. Due to recent changes the drawing
of the image editor is color managed, but the painting isn't what made
these changes show up.
This patch is a work-a-round so that the sampled colors and the effect
the paint has on the texture matches. This isn't the correct solution
as that would be to migrate all the painting tools to use proper color
management.
Reviewed By: Pablo Dobarro
Differential Revision: https://developer.blender.org/D9411
There are two implementations of the Sample Color operation. One is used
by the paint texture and one by the image editor. The image editor
variant sampled from the ibuf directly, but the paint texture variant
was sampling from the screen front buffer. This can lead into incorrect
samples due to color pipeline.
This patch will use the image editor variant when sampling a color for
2d texture painting
Reviewed By: Pablo Dobarro
Differential Revision: https://developer.blender.org/D9408
Previously the header was a fixed size and assumed to be zeroed.
Now read in bytes up to `HEADER_SIZE`, pass the number or bytes
read to the callback which must not read past those bytes.
We haven't supported 32bit mac builds for a while so this should be safe to remove.
Reviewed By: #platform_macos, brecht
Differential Revision: https://developer.blender.org/D9489
Needed a better normal to for plane to offset into when there are
non in-plane edges between two beveled edges. It was using the vertex
normal, which is just wrong.
Differential Revision: https://developer.blender.org/D9508
These are generally only one or two word phrases and are not sentences.
This change slightly improves readability.
Note, the check when display labels:
```
Tip Label (only for buttons not already showing the label).
```
Could be improved here because there are a lot of false positives.
See https://developer.blender.org/D9499.
It's odd to include a C++ header (".hh") in C code, we should avoid that. All
of the Outliner code should be moved to C++, I don't expect this C header to
stay for long.
See https://developer.blender.org/D9499.
"View" leads to weird names like `TreeViewViewLayer` and after all they are
specific to what we call a "display mode", so "display" is more appropriate.
Also add, update and correct comments.
See https://developer.blender.org/D9499.
* Use C++17 nested namespaces.
* Use `_` suffix rather than prefix for private member variables.
Also: Simplify code visually in `tree_view.cc` with `using namespace`.
See https://developer.blender.org/D9499.
* Turn functions into member functions (makes API for a type more obvious &
local, allows implicitly sharing data through member variables, enables order
independend definition of functions, allows more natural language for
function names because of the obvious context).
* Prefer references over pointers for passing by reference (makes clear that
NULL is not a valid value and that the current scope is not the owner).
* Reduce indentation levels, use `continue` in loops to ensure preconditions
are met.
* Add asserts for sanity checks.
See https://developer.blender.org/D9499.
Also:
* Add `space_outliner/tree/common.cc` for functions shared between display
modes.
* I had to add a cast to `ListBaseWrapper` to make it work with ID lists.
* Cleanup: Remove internal `Tree` alias for `ListBase`. That was more confusing
than helpful.
See https://developer.blender.org/D9499.
* Add comments to explain the design ideas better.
* Follow code style guide for class layout.
* Avoid uninitialized value after construction (general good practice).
See https://developer.blender.org/D9499.
* Use `blender::Map` over `GHash`
* Use `blender::Vector` over allocated `ListBase *`
Benefits:
* Significantly reduces the amount of heap allocations in large trees (e.g.
from O(n) to O(log(n)), where n is number of objects).
* Higher type safety (no `void *`, virtually no casts).
* More optimized (e.g. small buffer optimization).
* More practicable, const-correct APIs with well-defined exception behavior.
Code generally becomes more readable (less lines of code, less boilerplate,
more logic-focused APIs because of greater language flexibility).
See https://developer.blender.org/D9499.
* Turn functions into member functions (makes API for a type more obvious &
local, allows implicitly sharing data through member variables, enables order
independend definition of functions, allows more natural language for
function names because of the obvious context).
* Move important variables to classes rather than passing around all the time
(shorter, more task-focused code, localizes important data names).
* Add helper class for adding object children sub-trees (smaller, more focused
units are easier to reason about, have higher coherence, better testability,
can manage own resources easily with RAII).
* Use C++ iterators over C-macros (arguably more readable, less macros is
generally preferred)
* Add doxygen groups (visually emphasizes the coherence of code sections,
provide place for higher level comments on sections).
* Prefer references over pointers for passing by reference (makes clear that
NULL is not a valid value and that the current scope is not the owner).
This introduces a new C++ abstraction "tree-display" (in this commit named
tree-view, renamed in a followup) to help constructing and managing the tree
for the different display types (View Layer, Scene, Blender file, etc.).
See https://developer.blender.org/D9499 for more context. Other developers
approved this rather significantly different design approach there.
----
Motivation
General problems with current design:
* The Outliner tree building code is messy and hard to follow.
* Hard-coded display mode checks are scattered over many places.
* Data is passed around in rather unsafe ways (e.g. lots of `void *`).
* There are no individually testable units.
* Data-structure use is inefficient.
The current Outliner code needs quite some untangling, the tree building seems
like a good place to start. This and the followup commits tackle that.
----
Design Idea
Idea is to have an abstract base class (`AbstractTreeDisplay`), and then
sub-classes with the implementation for each display type (e.g.
`TreeDisplayViewLayer`, `TreeDisplayDataAPI`, etc). The tree-display is kept
alive until tree-rebuild as runtime data of the space, so that further queries
based on the display type can be executed (e.g. "does the display support
selection syncing?", "does it support restriction toggle columns?", etc.).
New files are in a new `space_outliner/tree` sub-directory.
With the new design, display modes become proper units, making them more
maintainable, safer and testable. It should also be easier now to add new
display modes.
- scrolling would be restricted (usually, if the object to be renamed is
in view, this prevents scrolling away without finishing the rename
operation)
- renaming by typing and confirming with Enter was not possible (you
would have to escape, scroll to the object and use F2 again)
- other shortcuts like A and H are still active instead of being handled
as text input
Avoid all these issue by forcing the item into view using
outliner_show_active / outliner_scroll_view.
Maniphest Tasks: T82553
Differential Revision: https://developer.blender.org/D9521
Scrolling to an item after opening relevant parents can go wrong if said
parent e.g. the last in the list [as in: then the Outliner does not
scroll down all the way]
It stems from the fact that 'region->v2d.tot.ymin' is not up-to-date in
outliner_scroll_view after outliner_show_active opens up parents, 'tot'
will only update on a redraw.
Now calculate the trees height on the fly using
'outliner_tree_dimensions()'.
ref D9521
ref T82553
Maniphest Tasks: T82553
Differential Revision: https://developer.blender.org/D9523
Selection should be possible from the left gutter in object mode. When
in other modes the mode column displays icon buttons which should be
prioritized for selection only in those modes.
Introduced in rB2110af20f5e6.
While `tp_print` was deprecated, Python 3.8+ uses this for
'tp_vectorcall_offset' which wasn't stated in the comment from
efd71aad4f.
Instead of suppressing clang-tidy, use preprocessor a check since
this properly represents the difference between Python versions.
For blender we disable VCPKG to prevent it from picking
up the wrong libraries from VCPKG rather than our lib folder
some of the cycles tests needed this to link correctly.
reported by @alef on chat
The names of the parameters are based on those of those of the sockets, so they also need to be updated. This was forgotten about in the previous commit (rBa284e559b90e).
Ref T82561.
There were some changes to the NanoVDB API that broke the way Cycles was previously using it.
With these changes it compiles successfully again and also still compiles with the NanoVDB revision
that is currently part of the Blender dependencies. Ref T81454.
Some operations, like remapping and ID (Object) to another, can lead to
having the same object in more than one base.
While this is not a valid state, this is being taken care of by the
`BKE_layer_collection_sync` call, so the object-to-base GHash generation
itself should be resilient to such issue.
Note: another way to fix this would be to make remapping post-process
code check explicitely for such doublons, but I would rather avoid
adding even more 'specialized' code there, it already has to deal with
too many of those corner cases.
If a define of NOMINMAX was made before BLI_task.hh was included,
the compiler would emit a
warning C4005: 'NOMINMAX': macro redefinition
warning, to work around this only define it if it is not already
defined, and only undefine it if we were the ones that made the
define earlier.
For the `plane_depth` of type `PLACE_DEPTH_SURFACE` to take effect, a `snap_context` is needed.
But, even if a temporary `snap_context` was created for `plane_orient == PLACE_ORIENT_SURFACE`,
this context was not used for `plane_depth`.
So, use a temporary `snap_context` for `PLACE_DEPTH_SURFACE` (as it is done for `PLACE_ORIENT_SURFACE`).
Differential Revision: https://developer.blender.org/D9345
Differential Revision: https://developer.blender.org/D9435
The "type" sockets on shader nodes were renamed in rB31a620b9420cab to
avoid clashes with the `NodeType type` member from the Node base class,
but the OSL shader compilation was missing those changes.
Expand unit test for `BKE_fcurve_active_keyframe_index()` to test edge
cases better.
This also introduces a new test macro `EXPECT_BLI_ASSERT()`, which can be
used to test that an assertion fails successfully.
No functional changes to actual Blender code.
Add `IMB_ispic_type_from_memory` so we can detect the file format
of in-memory images.
This removes `is_a_filepath` callback and uses a magic check for
photo-shop files that's compatible with OIIO.
Even though OIIO doesn't support packed images, we can still use the
file magic for detecting the format.
This change allows D9500 (a fix for unpacking images),
to be implemented without a significant performance penalty,
although the actual performance cost would depend heavily on the
blend file.
Reviewed By: dfelinto, sergey
Ref D9517
Perform grease pencil rendering delayed in this case, as there are no render
buffers available for compositing. This keeps memory usage lower, but does
involve multiple depsgraph evaluation. This seems in line with the intent of
the save buffers feature, to use minimal memory.
The skin modifier did not output consistent results, causing failure to find
corresponding vertices across frames.
Remove unnecessary use of GSet, all we need is an array.
Previous code was flipping the bits on a 32-bit number and doing a zero extension to cast to 64-bit, so mark the constant as long to begin with.
This would also erase previously set bits in this part the flag.
forceinline attribute is only applicable for function which are
marked inline. Interestingly, it can be used for class methods
without explicit inline statement. But for functions it is another
story.
And remove Blender preference, which was expected to be set to match the system
preference for correct behavior. Instead just handle this automatically.
Differential Revision: https://developer.blender.org/D9402
This should not change any functionality.
After talking to Brecht, we agreed that it might be good
not to have a class called Geometry for now. In the future
we might want to use a Geometry class as base class for
meshes, curves, etc.
This commit renames the Geometry class to GeometrySet,
because it is essentially a container that can contain
multiple geometries of different types.
Make post-bake cleanup of the Bake Action operator optional, and disable
by default.
Previously Bake Action would do two things:
- Bake the Action
- Clean up the FCurves
It is now possible (and even the default) to only perform the baking
operation.
Reviewed By: #animation_rigging, looch, sybren
Maniphest Tasks: T82210
Differential Revision: https://developer.blender.org/D9453
Volumes using tricubic sampling were producing different results with NanoVDB compared
to dense textures. This fixes that by using the same tricubic sampling algorithm in both
cases. It also fixes some remaining offset issues and some minor things that broke OpenCL
kernel compilation on NVIDIA.
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D9491
Move the "Show Group Colors" toggle from a per-editor option to a single
user preference in the Animation preferences. The Grease Pencil
animation channel side panel allows picking a channel color; this now
shows a message when channel colors are disabled.
The old "Show Group Colors" toggle had to be set per editor, and was on
by default. This meant that disabling group colors would require an
action for every file, for every editor. It is very hard to select a
color that works both as bone color in the 3D Viewport (needs to be
bright there) as well as the channel list (needs to be dark there), most
animators turn channel list colors off.
Differential Revision: https://developer.blender.org/D9391
Remove `static int clear_scene_in_allseqs_fn(...)`. It was a utility
function for `BKE_sequencer_clear_scene_in_allseqs()`, which was removed
in c063813c30.
No functional changes.
This reverts commit 4987b7d347.
This introduced a slight change in curve direction at end-points
(while correct), it caused tests to fail.
Keep this change for 2.92, revert for 2.91.
The code was trying to ignore hidden geometry when doing boolean,
which is correct when used as a tool, but not when a modifier.
Added a "keep_hidden" argument to bmesh_boolean to distinguish the
two cases.
Also fixed a bug when the tool is used with hidden geometry that
is attached to unhidden geometry that is deleted by the operation.
The code was trying to ignore hidden geometry when doing boolean,
which is correct when used as a tool, but not when a modifier.
Added a "keep_hidden" argument to bmesh_boolean to distinguish the
two cases.
Also fixed a bug when the tool is used with hidden geometry that
is attached to unhidden geometry that is deleted by the operation.
The code for determining coplanar clusters had a bug where it would
miss some triangles. The fix for now is to just put triangles in
the cluster if their bounding boxes overlap. This works but maybe
makes clusters bigger then they have to be. I'll follow this commit
with work on making the CDT routine faster when using exact arithmetic.
Also removed a lot of unused code, and added some new intersect
performance tests.
When using a trackpad Zoom to Mouse Position would always zoom to center of canvas.
Differential Revision: https://developer.blender.org/D8683
Reviewed by Brecht Van Lommel
Ensure that Zoom does not crash on Mac Trackpad by checking for existence of Continuous Zoom timer.
Differential Revision: https://developer.blender.org/D8682
Reviewed by Julian Eisel
Stop with an error when the Clang-Tidy executable cannot be found.
Without this check, CMake will happily report "Found Clang-Tidy" but with
the fallback version (0, 0, 0), when `CLANG_TIDY_EXECUTABLE` points to a
non-existing executable.
Nicer appearance for the progress bar that is drawn over the application icon during long processes on macOS.
Differential Revision: https://developer.blender.org/D9398
Reviewed by Brecht Van Lommel
When using Optimal Display, some edges are not flagged `ME_EDGEDRAW` |
`ME_EDGERENDER`.
When the modifier is applied through the UI in the modifier stack this is
not an issue because the `modifyMesh` callback is run with
`MOD_APPLY_TO_BASE_MESH` (this will effectively turn of Optimal
Display).
When converting to mesh though, this will just get an evaluated mesh
(where the edge flags are still the same as with the subdivision
modifier).
Now ensure every edge is flagged to draw after conversion.
Maniphest Tasks: T81997
Differential Revision: https://developer.blender.org/D9331
Caused by rBb077de086e14.
Not entirely sure why this was rebuilding the tree prior to above
commit, but sending an ND_OB_SHADING notifier is appropriate (and also
what the Outliners listener listens to).
Maniphest Tasks: T82251
Differential Revision: https://developer.blender.org/D9396
This is caused by the TAA being reset after the init phase, leading to
1 sample being kept as valid when it is clearly not.
To fix this, we run the lookdev validation before TAA init.
Reviewed By: Jeroen Bakker
Differential Revision: https://developer.blender.org/D9452
This adds a Box option to the Text strip's style properties, plus related Box Margin value:
{F9208309}
When enabled the text is placed on top of a solid-filled rectangle of a chosen color, as shown below:
{F9208324}
When the box option is disabled the text strip works the same as it does now. When the box option is enabled the meaning of the Shadow option changes to provide a drop-shadow on the rectangle (and not on the text itself). The latter made more sense to me.
The box margin is specified as a fraction of the image width. The offset of the drop-down box shadow is fixed to a specific fraction of the image width as well.
I tested this feature on a movie of a couple of minutes containing dozens of text strips (all with box background), edge cases like multi-line strings and text overlapping the image edges.
Reviewed By: ISS
Differential Revision: https://developer.blender.org/D9468
Proxies are expected to be fast to read. Storing them in cache has
little to no effect on performance.
This change also allows to omit invalidation of cache when user switch
between proxies and original media.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D9473
Variables renaned:
- cfra -> timeline_frame
- nr -> frame index
- cfra_over -> overlap_frame
Function seq_give_stripelem_index was renamed to seq_give_frame_index.
Use bool return type where possible instead of int (the return values from fluid object are already boolean instead of int).
Also removed several if guards in API functions. If one of the arguments is in fact invalid / nullptr (should not happen though), it better to catch them directly where they failed and not silently escape them.
Refactor material assignment code such that:
- `build_mat_map()` just returns the built map (instead of relying on
modifying a map passed as parameter),
- `LISTBASE_FOREACH` is used to loop over a `ListBase` (instead of a
hand-crafted for-loop),
- just `return` when not enough material slots can be created (instead
of setting a boolean to false, then doing some useless work, then
checking the boolean),
- reorder some code for clarity, and
- rename `mat_map` to `matname_to_material` so that the semantics are
clearer.
No functional changes.
So a keyframed e.g. location slider would stay yellow/green even if its
corresponding channel was removed.
Needs a appropriate notifier so the listeners (e.g.
buttons_area_listener, view3d_buttons_region_listener) would cause a
redraw.
Maniphest Tasks: T82364
Differential Revision: https://developer.blender.org/D9438
This fix makes sure new files save `wmWindow.global_areas` under a different
name, so old Blender versions don't recognize and 0-initialize it.
Since enabling global area writing (ef4aa42ea4), loading a file in old
Blender versions would cause `wmWindow.global_areas` to be read, because there
was already reading code for it and `ScrAreaMap` was in SDNA.
However the `ScrArea.global` of the global areas would be NULL, because it was
*not* in SDNA (`ScrGlobalAreaData` was excluded).
Now, issue is that the code assumes that areas in the global area-map have a
valid ScrArea.global pointer.
Think this was a mistake in rB5f6c45498c92. We should have cleared all this data
on reading, until the global area writing was enabled.
Differential Revision: https://developer.blender.org/D9442
Reviewed by: Brecht Van Lommel
accelerations
Caused by rB45dbc38a8b15.
Above commit would place parentheses surrounding a block until the next
operator was found.
For velocities and accelerations though, the '/' in 'm/s' or 'ft/s'
should not be considered an operator.
Maniphest Tasks: T82407
Differential Revision: https://developer.blender.org/D9467
When editbones were selected from the Outliner (and they were connected
to a parent) with the 'Sync Selection' option turned ON, they could not
get duplicated.
For duplication to work, the (connected) parent bone's tip also has to
be selected [which was not the case when selection is done from the
Outliner under above circumstances]. The reason being that
armature_duplicate_selected_exec ->
ED_armature_edit_sync_selection clears the BONE_ROOTSEL flag if the
parent bone's BONE_TIPSEL is not set.
Caused by rB71eb65328078 btw.
The correct "parent-tip-selection" would actually happen in activation
- `tree_element_active_ebone`
-- `tree_element_active_ebone__sel`
but for 'Sync Selection' this happens [also] in
- `outliner_sync_selection_from_outliner`
-- `outliner_select_sync_to_edit_bone`
which did not do the "flushing" to the parent bone's tip
Now use existing dedicated function for this.
ref. T82347
Reviewers: Zachman
Differential Revision: https://developer.blender.org/D9470
Using configue_file(..) would have avoided the breakage from
1daa3c3f0a, caused by buildinfo not properly escaping quotes.
Rely on CMake to escaping strings instead using configure_file().
Use common prefix as this collided with existing API's (eg BLI_voronoi).
Also expand some non-obvious abbreviations:
- 'g' -> 'generic'
- 'vl' -> 'variable_lacunarity'
- 'V' -> 'v3'
The draw face sets brush uses the poly center when used in meshes to increase
its precision when working in low poly geometry. For this to work with deformed
meshes, the deformed coordinates from the PBVH should be used instead.
Reviewed By: sergey
Maniphest Tasks: T81915
Differential Revision: https://developer.blender.org/D9424
BKE_mesh_free() seems to not free the meshes correctly, so using BKE_id_free() instead.
The looptri array was also not freed.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D9426
Used the sampled cursor normal when available instead of the raycast face normal.
This makes the preview match the previous orientation of the cursor.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D9460
"Camera Parent Lock" can be useful when rigging cameras, but it is not
intuitive, and has also generated a lot of confusion (bug reports).
This is because it breaks the fundamental parent <-> child relationship
conventions in Blender, and there is no indication that it's intended
without diving into the preferences.
This commit moves the setting to the object level, and exposes it in
the relations panel in the property editor. It is exposed for every
object type because any object type can be "View Locked" in the 3D view.
The property description is also updated to reflect this change and be
more specific without getting too long.
In the future this could become a more general feature of the transform
system, but for now it is limited to "Lock Camera to View".
Differential Revision: https://developer.blender.org/D9239
Commit rBf5080c82dd915db6c7b9dd68a52aaaccf2600137
accidentally remove the Shift modifier key from
the `AUTOCONSTRAINPLANE` shortcut.
Differential Revision: https://developer.blender.org/D9480
Since it is possible to have multiple draw callbacks, (some of which
use bgl and others gpu), check and force the reset of the drawing status
at the end of each callback.
Differential Revision: https://developer.blender.org/D9476
The two entries `TFM_MODAL_AUTOCONSTRAINT` and
`TFM_MODAL_AUTOCONSTRAINTPLANE` had the same name
displayed in the UI. The latter is now includes
"plane" in it's name.
Reviewed By: mano-wii
Differential Revision: https://developer.blender.org/D9474
This commit corrects the maximum resolution field in Fluid objects. The field should be set to the maximum possible resolution, i.e. it should not adjust with adaptive domains and stay constant all the time.
The reason for this is that this resolution value will be used to scale gravity. And this gravity should be constant for adaptive domains too.
It remains to be shown if this issue was the only reason for line-artifacts as seen in T74559.
Fixes 18 misspellings of 'predefined', 'Look Up', 'Lookup', and 'No One'.
Differential Revision: https://developer.blender.org/D9466
Reviewed by Hans Goudey
Map frame for cached raw images to strip input media frame range. This
means that static images or extended frame range of movies will only
generate one cache entry.
timeline_frame is stored in cache key as a reference - on what frame
was this entry created, so we don't have to reverse lookup frame range.
Since each media frame corresponds to one cache frame with same frame
index key, there is no need to invalidate raw cache when changing time
remapping properties like use_reverse_frames or strobe
No changes are needed for disk cache, since invalidating raw entry
assumes all data will be invalidated.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D9462
This commit corrects the maximum resolution field in Fluid objects. The field should be set to the maximum possible resolution, i.e. it should not adjust with adaptive domains and stay constant all the time.
The reason for this is that this resolution value will be used to scale gravity. And this gravity should be constant for adaptive domains too.
It remains to be shown if this issue was the only reason for line-artifacts as seen in T74559.
Replace BKE_sequencer wirh SEQ_render or SEQ_proxy prefixes.
In cases where function is very generic, only SEQ prefix is used.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D9439
Move functions closely related to rendering images and proxies into
render.c and proxy.c files. render.h and proxy.h are created for
functions used internally.
There should be no functional changes.
Keymaps have previously been exported with an encoding dependent
on the current system locale. This caused issues when the
keymap contained non-ASCII characters, for instance in a string
property for an operator.
Reviewed By: campbellbarton
Differential Revision: https://developer.blender.org/D9449
Re-enables support for menus to have items without identifier or name that can be used to separate sections.
Differential Revision: https://developer.blender.org/D9463
Reviewed by Hans Goudey
This adds a new runtime flag for panels that is set during the entire
drag and animation operation. The flag is set recursively so that
sub-panels know to draw on top too.
Note that this also replaces most of the fixes in 1960b8a361 and
8e08d80e52 (D7462) with a more "built-in" solution.
These functions were not used elsewhere, and the handling for the panel
tabs should be kept local to this file where possible. Also remove
another unused function and removed an unecessary "_ex" function.
The current layout gave too little space for the full "Search" string inside
the button.
Fix this by making sure radio-buttons have their text center aligned by default
in pop-ups too, like they do anywhere else.
This does affect a few other cases, e.g. the "RGB"/"HSV"/"Hex" radio-toggles
for color pickers. But this should be fine, I don't think they were ever
intentionally using left-aligned text (while similar buttons outside of pop-ups
didn't).
Use "Rotation" in place of "Rot" for the rotation header text.
Cleanup various RNA titles and tooltips.
Differential Revision: https://developer.blender.org/D9457
This function for replacing a button pointer in button groups is
needed by the block update from old code, so it cannot be removed
like its layout equivalent.
Even after the last commit to fix this assert, it still fails in the
case where a button was added before there was a button group
added to the block. Another fix for this would be to always create
a button group in UI_block_begin, but this assert has no particular
purpose, so it's simpler to just remove it.
Set step in RNA property definition.
There was a hardcoded setting of a1 for this specific button, which
used to be the variable to store the step size of number buttons until
rBe6f0b60c2e91. hardcoded value is removed in rBe29206f86a5f.
Reviewed By: Severin
Differential Revision: https://developer.blender.org/D9277
Use a more reliable method to check the availability of the flag than
compiler versions. Some compilers have different behaviors for
C and C++.
Reviewed By: campbellbarton, ChrisLend
Differential Revision: https://developer.blender.org/D9446
If the annotation draw operator is used, enable the annotations in the current area.
Before this change, some editors had the annotation flag set to OFF, but this could be solved with a versioning code, but this did not solve the root problem. The user can disable annotation visibility in the overlay or side panel, depending on the editor. If the user uses the annotation tool and this flag is OFF, the annotation is not visible, and this is not correct. With this patch, every time the user uses the tool, the annotation visibility flag is set to ON to ensure the annotation is visible.
This solves the problem of T82273, T79578 and T80294
Maniphest Tasks: T82273
Differential Revision: https://developer.blender.org/D9409
The NanoVDB sampling implementation behaves different from dense texture sampling, so this
adds a small offset to the voxel indices to correct for that.
Also removes the need to modify the sampling coordinates by moving all the necessary
transformations into the image transform. See also T81454.
This patch will apply the view transform when a movie clip is used as
camera background image. It does this by rendering the image in the
color buffer when it needs the view transform. For other images it uses
the overlay buffer.
Reviewed By: Clément Foucault
Differential Revision: https://developer.blender.org/D7067
Issue was that the `tris_per_mat` are not created when the first batch is drawn
during select operator and then is not created when needed by the workbench pass
since they are not tracked by mesh_buffer_cache_create_requested.
This change will create the `tris_per_mat` just in case they are needed later.
Solution by Clément Foucault
Differential Revision: https://developer.blender.org/D9430
The issue is that the shaders are stolen from the original Geometry by
the temporary Geometry used to accumulate data, but the main thread
still needs them for syncing the attributes.
So make a copy of the shader array to preserve the data on the original
Geometry.
The issue was that the volume for the current frame
might not have been loaded already by the time the
modifier runs.
The solution is simply to make sure that the volume
is loaded. This is similar to the Volume Displace modifier.
Keymaps have previously been exported with an encoding dependent
on the current system locale. This caused issues when the
keymap contained non-ASCII characters, for instance in a string
property for an operator.
Reviewed By: campbellbarton
Differential Revision: https://developer.blender.org/D9449
This encapsulates Node socket members behind a set of specific methods;
as such it is no longer possible to directly access Node class members
from exporters and parts of Cycles.
The methods are defined via the NODE_SOCKET_API macros in `graph/
node.h`, and are for getting or setting a specific socket's value, as
well as querying or modifying the state of its update flag.
The setters will check whether the value has changed and tag the socket
as modified appropriately. This will let us know how a Node has changed
and what to update, which is the first concrete step toward a more
granular scene update system.
Since the setters will tag the Node sockets as modified when passed
different data, this patch also removes the various modified methods
on Nodes in favor of Node::is_modified which checks the sockets'
update flags status.
Reviewed By: brecht
Maniphest Tasks: T79174
Differential Revision: https://developer.blender.org/D8544
When there was an active button in the "old" block from the last redraw,
this code tried to replace its pointer in the new block's button groups.
But in cases like the outliner or file browser, there are no groups
because the block doesn't use the layout system at all. This commit
just tweaks the assert to check whether there are any button groups.
When there was an active button in the "old" block from the last redraw,
this code tried to replace its pointer in the new block's button groups.
But in cases like the outliner or file browser, there are no groups
because the block doesn't use the layout system at all. This commit
just tweaks the assert to check whether there are any button groups.
This simply makes the panel a bit nicer given how things are layed out--
the items with larger visual weight are grouped at the top.
Differential Revision: https://developer.blender.org/D9366
This enables self intersections in the sculpt trimming tools boolean operations.
This should fix wrong booleans results after using the operator to add new disconnected
geometry with the join mode.
Reviewed By: sergey
Maniphest Tasks: T81799
Differential Revision: https://developer.blender.org/D9423
The cloth brush fixed simulation areas are created using the initial_location variable in the
StrokeCache. This variable was not being updated by tiling symmetry, so all symmetry passes
were using the same simulation area location.
Reviewed By: sergey
Maniphest Tasks: T81842
Differential Revision: https://developer.blender.org/D9421
When extracting new objects from a mesh in sculpt mode, it makes sense to keep
the current symmetry options and settings in the new mesh. In previous versions symmetry
options were stored in the tool settings, so this bug was not that obvious.
This also preserves the remember settings in the new object, which is also the desired
behaviour.
Reviewed By: sergey, mont29
Differential Revision: https://developer.blender.org/D9417
If the new object contains the mask it always needs to be cleared before starting
sculpting on it.
This fix was also committed before in the mask extract operator.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D9416
The strength of this brush needs to take alpha into account
Reviewed By: sergey
Maniphest Tasks: T82297
Differential Revision: https://developer.blender.org/D9419
The list of buttons in the button group needs to be updated to take into
account the old button inserted into the button list for the new block.
Differential Revision: https://developer.blender.org/D9428
What I thought was an "optimization" was really a bug. The "use search
for expansion" value needs to be set for every panel, even panels in
other tabs. Otherwise it won't be properly set when switching back to
a tab that was visited during search.
Differential Revision: https://developer.blender.org/D9427
Setting the button's step size like this wouldn't work anymore after
e6f0b60c2e, which is reported in T81794. Instead, the step size should be set
for the RNA property, as proposed in D9277. That will be committed separately
as bug fix.
Until it is decided whether to ship JACK with pre-compiled libraries,
search for the same in system directories.
Ref T79261
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D9436
This avoids recomputing the BVH for geometries that do not have changes in topology but whose vertices are modified (like a simple character animation), and gives up to 40% speedup for BVH building.
This is only available for viewport renders at the moment.
Reviewed By: pmoursnv, brecht
Differential Revision: https://developer.blender.org/D9353
The tooltip implies that this option should be off per default instead
of on. Some talking with the animators at the studio and in the
animation module confirmed that this seems to be the most logical
default.
The links where added to the socket one after the other. However,
the virtual socket had a link limit of 1, so whenever a new link was
added, the previously added one was removed.
There is not really a reason for why the link limit should be 1 instead
of something higher. I'm setting it to the max value: `0xFFF`.
I'm also setting the `input_link_limit` to that value. Blender does not
need this currently, but addons might have input sockets that allow
more than one incident link.
Default backbuffers needs not to be bound with sRGB encoding
enabled. This works when using `GPU_framebuffer_restore` but
using `GPU_framebuffer_bind` would trigger the wrong behavior.
This fix T81969 UI turns whiteish when playing video sequence
based on a scene and moving in the image editor after saving
This patch allows the user to type a property name into the
Attribute node, which will then output the value of the property
for each individual object, allowing to e.g. customize shaders
by object without duplicating the shader.
In order to make supporting this easier for Eevee, it is necessary
to explicitly choose whether the attribute is varying or uniform
via a dropdown option of the Attribute node. The dropdown also
allows choosing whether instancing should be taken into account.
The Cycles design treats all attributes as one common namespace,
so the Blender interface converts the enum to a name prefix that
can't be entered using keyboard.
In Eevee, the attributes are provided to the shader via a UBO indexed
with resource_id, similar to the existing Object Info data. Unlike it,
however, it is necessary to maintain a separate buffer for every
requested combination of attributes.
This is done using a hash table with the attribute set as the key,
as it is expected that technically different but similar materials
may use the same set of attributes. In addition, in order to minimize
wasted memory, a sparse UBO pool is implemented, so that chunks that
don't contain any data don't have to be allocated.
The back-end Cycles code is already refactored and committed by Brecht.
Differential Revision: https://developer.blender.org/D2057
Uniform attributes require immediate access to the shader list
in object update code, so setting the field can't be deferred
to a background task. This required adding a parameter to the
clear method of Geometry.
Ref D2057
Add a new Alpha socket to the Attribute node that outputs the
fourth component of the attribute. Currently the only such
attribute is vertex color, but there may be more in the future.
If the attribute has no alpha channel, the expected value is 1.
The Cycles code is already refactored and committed by Brecht.
Ref D2057
Add code preserving scene's toolsettings accross undo.
IDPointers are dealt with special care, we try to keep existing ones for
some (like brushes) when possible.
Note that this covers ToolSettings, Brushes and Palettes currently.
I'm not especially happy about how this new code mixes with existing
'foreach_id' one, in particular in scene. But cannot think of a better,
more generic way to do it currently.
Maniphest Tasks: T71759
Differential Revision: https://developer.blender.org/D9311
This is essentially adding that new callback, and using it only for already
existing Scene's 3DCursor.
Note that the place where this is called has been moved again, after all
have been lib-linked, such that those callbacks may also work on ID pointers.
Maniphest Tasks: T71759
Differential Revision: https://developer.blender.org/D9237
Surface deform weight calculation assigned weights in a non-uniform
way that caused vertices to deform upon binding.
This was caused by the face-corner angle being used in
calculations which where squared & scaled.
Causing a triangle fan of many thin faces to have a much greater
influence compared to the same shape made from a single triangle.
Change the calculation of the weight so each face-corner is scaled
by it's angle.
Ensure that When checking "Hide in Viewport" option for a collection
that child objects are drawn grayed out for consistency with the
"Disable in Viewports" toggle.
For checking an object visibility in the viewport the flag
BASE_VISIBLE_VIEWLAYER should be used instead of BASE_VISIBLE_DEPSGRAPH
because the latter ignores viewport visibility.
Manifest Task: T77161
Differential Revision: https://developer.blender.org/D7904
This addition to the filters allows the user to enable the
outliner tree to restrict the listing to only Selectable objects.
Differential Revision: https://developer.blender.org/D7310
In some situations where two beveled edges were very close to in-line
but not quite straight, bevel would build a miter when it shouldn't.
The code that chose whether to use a miter at each vertex was slightly
incorrect.
For outer miters there is a check for 3 or more selected edges, but an
inner miter can still be useful with only two beveled edges at a vertex,
so we can't use that here. Instead I changed the check for in-line edges
to run before determining whether the angle is reflex or not. The logic
ends up a bit more straightforward as well. This doesn't completely
remove the rather strange looking triangle vertex meshes at each corner,
but it does make it stable when locations are slightly adjusted.
The only other place this `edges_angle_kind` function was used is for
profile=1.0 vertex meshes. I tested and made sure that still works well.
Differential Revision: https://developer.blender.org/D9420
WM_POINTERLEAVE occurs when the pen goes out of range or when a
hovering pen leaves the window's boundary. When leaving the window
boundary the xy position is invalid for some Wacom devices.
This change removes creation of GHOST_EventCursor during
WM_POINTERLEAVE events. This prevents unexpected jumping behavior
during continuous grab.
`WM_operator_properties_filesel()` allows C operators to set a display or sort
type for the File Browser to use. But the File Browser would always override
that because of an invalid `_is_set()` check. (The operators don't actually set
the value, they only set the property's default value.)
The only operator affected by this is "Recover Auto Save". It is supposed to
show a vertical list ordered chronologically. It used settings from the
previous File Browser usage before this patch.
Operators using the File Browser should generally use
`FILE_DEFAULTDISPLAY`/`FILE_SORT_DEFAULT` now, except if they have a reason not
to. See comments at their definition.
----
This makes it so operators that set a different display or sort type
don't change the sort or display type for the next File Browser operation.
So using "Recover Auto Save" entirely isolates display and sort type from other
operations.
Differential Revision: https://developer.blender.org/D8598
Reviewed by: Bastien Montagne
I think there wasn't actually any issue currently, but only by luck. We still
passed around and NULL-checked a pointer to freed memory (the file operator,
`SpaceFile.op`) which is easy to break and should be avoided.
Noticed while testing D8598.
This patch changes behavior of strip transform and crop feature.
Purpose of this change is to allow display arbitrary portion of input
image, simplify user interface and workflow.
Offset and Crop values in old files are converted in versioning.
Offset animation is also converted. Crop animation and animation of
crop or offset enable properties is not taken into account
Changes in behavior and interface:
- If image is added to timeline it is scaled to fit inside preview area
while maintaining aspect ratio. Image is centered. This is considered
as a baseline for further transformation.
- Scale and rotation was added, so it is possible to transform image at
it's original resolution.
- Crop will not affect image transformation (does not move image).
- Values of Crop and Transform Position are in pixels, these values are
corrected if preview is fraction of project resolution.
- Transform and Mirror panel has been removed and new Transform panel
and Crop panel is moved to Adjust panel. Mirror is now part of new
Transform panel.
Technical changes:
- Preprocessing stage must work on duplicated image, because original is
cached. Previously Crop and Offset could run at once and required only
one duplication of image. This is not the case with new algorithms, so
duplication on demand is implemented. Transformation can read original
image and will output new image that is safe to modify. It should be
possible to add crop step to transform algorithm, so that Crop won't
require previous duplication though.
- Use Crop and Use Translation checkboxes were removed. Individual
values are compared to default values to check if image needs to be
processed. In case of transform this will be done also if resolution of
source.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8393
This patch changes behavior of strip transform and crop feature.
Purpose of this change is to allow display arbitrary portion of input
image, simplify user interface and workflow.
Offset and Crop values in old files are converted in versioning.
Offset animation is also converted. Crop animation and animation of
crop or offset enable properties is not taken into account
Changes in behavior and interface:
- If image is added to timeline it is scaled to fit inside preview area
while maintaining aspect ratio. Image is centered. This is considered
as a baseline for further transformation.
- Scale and rotation was added, so it is possible to transform image at
it's original resolution.
- Crop will not affect image transformation (does not move image).
- Values of Crop and Transform Position are in pixels, these values are
corrected if preview is fraction of project resolution.
- Transform and Mirror panel has been removed and new Transform panel
and Crop panel is moved to Adjust panel. Mirror is now part of new
Transform panel.
Technical changes:
- Preprocessing stage must work on duplicated image, because original is
cached. Previously Crop and Offset could run at once and required only
one duplication of image. This is not the case with new algorithms, so
duplication on demand is implemented. Transformation can read original
image and will output new image that is safe to modify. It should be
possible to add crop step to transform algorithm, so that Crop won't
require previous duplication though.
- Use Crop and Use Translation checkboxes were removed. Individual
values are compared to default values to check if image needs to be
processed. In case of transform this will be done also if resolution of
source.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8393
The behavior of the incremental snap did not take into account the
relative dimensions of the window, which resulted in a different behavior
if the area height was greater than the width.
There should not be much user visible here (other than T73668 being addressed).
I added the writing code already for the initial implementation of workspaces,
but we decided to keep it disabled until the top-bar design is more clear. It
was never planned to keep this disabled for so long.
Fixes T73668.
Cycles defines some basic integer types since it cannot use the standard headers when
compiling with NVRTC. NanoVDB however only does this when the "__CUDACC_RTC__" define
is set and otherwise includes the standard "stdint.h" header which clashes with those typedefs.
So for compatibility do the same thing in the Cycles kernel headers. See also T81454.
This fixes critical bug with liboverride when soe add-ons add some
RNA ID Pointer properties.
ID pointers should **never** have ownership of their ID when defined
from python.
(As a reminder, RNA properties owning their ID pointers are extremely
rare even from C code, only embedded IDs (root node trees, master
collections) and the shape keys snowflakes are concerned.)
Add a comment to the declaration of the `BKE_object_where_is_calc...()`
functions to explain where the result of the calculation is stored.
No functional changes.
Consistently return `false` from `ED_object_parent_set()` when parenting
is not possible. Before, when parent and child were the same object, the
function would return `true` even though the parent-child relation was
not made.
Just returning `false` in the `parent == child` case would break the
parenting operator, as `false` stops its loop over all selected objects.
This tight coupling caused T82312. The loop now has its own check for
this, so that it properly continues, and the implementation of
`ED_object_parent_set()` is decoupled from its surrounding code.
No functional changes.
The issue was in `buildinfo.c`:
char build_c[xx]flags[] = BUILD_C[XX]FLAGS;
Non-escaped double-quotes were terminating the string early, and
causing the compile error. So use single-quotes.
This change removes the user-specific information from
macros like `__FILE__` and keeps it relative to top level
source or build (for generated files) directory.
It makes traces concise.
Added option `WITH_COMPILER_SHORT_FILE_MACRO` enabled by default.
Reviewed By: campbellbarton
Differential Revision: https://developer.blender.org/D9386
This was added when Python was initially bundled so any problems
finding Python could be investigated.
Move this to use logging so we can show this information when needed.
Tests files are based on test from D8393
Test files should be in `lib\tests\sequence_editing`
These are files, I will add few more tests including animation test.
{F9155273}
Using generic tool to compare rendered vs reference image as other render engines.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D9394
This reverts commit 6527a14cd2.
Some 3rd party scripts made use of this so scripts can work with
2.90 to 2.92, although eventually this can still be removed.
For the fast solver, there was an optimization carried over
from the non-collection case for empty meshes which did not
work in the chained boolean code for collection operands.
Removed that optimization in the collection case.
For the fast solver, there was an optimization carried over
from the non-collection case for empty meshes which did not
work in the chained boolean code for collection operands.
Removed that optimization in the collection case.
The code that decided to use a faster double version of plane
side testing forgot to take an absolute value, so half the time
the exact code was being used when it was unnecessary.
The code that decided to use a faster double version of plane
side testing forgot to take an absolute value, so half the time
the exact code was being used when it was unnecessary.
This node is similar to the subdivision surface modifier. It subdivides
the mesh component of a geometry. The options have the same meaning
as in the modifier.
Differential Revision: https://developer.blender.org/D9364
Naming based on supported geometry components only works in some cases.
For nodes that work with multiple data types, it's better to categorize
them based on their functionality.
While Cycles already supports using both CPU and GPU at the same time, there
currently is a large problem with it: Since the CPU grabs one tile per thread,
at the end of the render the GPU runs out of new work but the CPU still needs
quite some time to finish its current times.
Having smaller tiles helps somewhat, but especially OpenCL rendering tends to
lose performance with smaller tiles.
Therefore, this commit adds support for tile stealing: When a GPU device runs
out of new tiles, it can signal the CPU to release one of its tiles.
This way, at the end of the render, the GPU quickly finishes the remaining
tiles instead of having to wait for the CPU.
Thanks to AMD for sponsoring this work!
Differential Revision: https://developer.blender.org/D9324
Previously Wintab was handled by saving the most recent tablet pressure and tilt information and deferred appending tablet infromation to Windows mouse events. This caused synchronization issues evident at the beginning and ending of strokes where pressure and tilt were either ahead or behind in time from mouse button up or down events. This also dicarded swaths of data which resulted in blockly grease pencil lines most apparent when a context switch resulted in several coalesced mouse events.
This patch changes the behavior of Wintab to instead rely entirely on Wintab information for pressure, tilt, position, and button input.
Wintab has several design decisions and conventions which complicate relying soley on it's input while retaining current workflows reliant on non-API behavior. For example, many device optionally allow the user to map barrel buttons to non-mouse actions. These mappings may or may not modify the intended behavior when touching the stylus down, such as scroll vs alt mappings. This behavior is not exposed in the Wintab API, but Wintab will continue updating button state sans this necessary context.
To work around the problem, this refactor synchronizations tablet input to Windows mouse down and up events, this captures events which should result in pen input while allowing events such as pen scrolling. Until a Windows mouse down event fires Wintab input is left unprocessed; when a Windows up event occurs Wintab is processed until a corresponding button up event is found.
Wintab allows for either button state or changes to be reported, but not both. An earlier refactor tried to use button changes to let state to be managed by Wintab. This was replaced when it was found that button change events were unreliable at corner cases such as switching windows. It was also found that with Wacom drivers Wintab peek functions would modify events in the queue causing errant and loss of button events.
For the latter stated reason this patch opts to read all Wintab events into a queue as they arrive, removing events as they become stale. This was chosen over using Wintab peek functions due to the afformentioned issue. As a bonus this seems to work better as it prevents the queue in Wintab from filling, thus neither a flood of events need to be handled when Wintab processing begins and a Wintab implementation need not be trusted to overwrite old events in it's queue.
Maniphest Tasks: T75566
Differential Revision: https://developer.blender.org/D7840
event to the button down location as this should be a more accurate point
of contact than the last mouse move event.
Signed-off-by: Nicholas Rishel <rishel.nick@gmail.com>
events until one is found. This prevents errant cursor moves that occur
before the Wintab button event is reported. We need to skip these events
because if no button event exists, we generate one assuming it will either
arrive later in the Wintab queue or that the button was from a non-Wintab
device. For the case that this was generated by a non-wintab device, such
as buttons mapped to mouse on the tablet pad, these cursor move events can
significantly move the cursor from the intended click position.
Signed-off-by: Nicholas Rishel <rishel.nick@gmail.com>
the issue with Wintab button events are more significant than simply
setting what buttons should receive button up/down events during context
initialization.
Signed-off-by: Nicholas Rishel <rishel.nick@gmail.com>
handling mouse input. This Wintab to mouse synchronization issues, and
likely prevents queue exhaustion for some Wintab implmenetations.
Signed-off-by: Nicholas Rishel <rishel.nick@gmail.com>
window intitialization can specify whether it will be visible regardless
of whether it is yet visible.
Signed-off-by: Nicholas Rishel <rishel.nick@gmail.com>
Button events now include tabletdata, so move is unnecessary.
Generate mouse button events when the system has an event but Wintab did not find a correlated event.
Only filter mouse button events, not Win32 Pointer events.
Signed-off-by: Nicholas Rishel <rishel.nick@gmail.com>
Maniphest Tasks: T75566
Differential Revision: https://developer.blender.org/D7404
This introduces a new operator to edit the detail size of constant
detail mode in dyntopo. The way this operator works and the
functionality it provides is similar to the "Voxel size edit" operator
for the voxel remesher.
It also includes a sample mode. When pressing Ctrl, the detail size
will be sampled from the surface under the cursor, updating the
preview in real time. This allows quick resolution changes without
using the operator multiple times.
The operator is set to Shift + D, replacing the old way to change
the constant detail size of dyntopo. Shift + R will remain available to
be enabled when the voxel remesher works with dyntopo. Deciding
if both detail sizes can be unified needs a separate discussion as the
new dyntopo can work with detail sizes in parts of the mesh that can
easily crash the remesher.
The structure of these operators is similar, but the data they control,
ranges, drawing and setup functions are completely different, making it
hard to merge them into one.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D9355
Global automasking options are usually turned on/off for quick
adjustments both in brushes and filters, so it is convenient to have them
in a pie menu.
This uses the Alt + A shortcut.
Reviewed By: dbystedt, HooglyBoogly
Differential Revision: https://developer.blender.org/D9282
This adds an option to orientate the trimming shape using the surface
normal instead of the view when lasso trim is used.
Reviewed By: dbystedt, sergey
Differential Revision: https://developer.blender.org/D9231
Theoretically the modifier's properties could be changed or removed by
Python, so it may be useful to have errors printed and added to the
modifier in that case.
I kept the check as a separate step from the `compute_geometry`
pass because the object is needed to set the modifier error message.
But it could easily be moved there in the future.
If we add the "name" label manually with uiItemR, we can use the socket
identifier instead of the name for the IDProperty name. This will also
allow us more flexibility in how to draw the settings in the future, and
removes the empty _RNA_UI proprety that was drawn before.
Caused Victor.blend from the cloud to crash with ASan. Four floats are
fetched from the stack, but the buffer was only three wide.
Caused by 042143440d. Issue was probably harmless since the fourth
element wasn't actually touched.
Ensure parenting an object to itself is seen as error, by returning
`false`.
This error was introduced as part of a supposed-to-be-non-functional
cleanup rBb8d4a2aff8069dd7d6fb91ad0d9427eed489b68f.
Non-memfile undo steps never properly initialized the flag allowing to
re-use old Main data during undo/redo. This lead to doing a complete
full re-reading of data when undoing/redoing mode switches e.g.
Note that current undo system is supposed to support any kind of mode
switch across those steps, however this needs to be properly deeply
tested, so only comitting this to master. It would be way too risky for
2.91 release.
This adds support for treating multiple undo steps as a single step
from the user perspective.
This is needed for outliner mode switching and `object.switch_object`
operator which change active object and mode in a single action.
For edit/pose modes clicking in the mode-column would include other
selected objects (besides the active object), causing other objects to
remain in the mode even though Control wasn't held.
Also replace `ED_object_mode_generic_exit(...)` on all objects with a
single call to `ED_object_mode_set(...)` so switching modes will only
exit the current mode, leaving objects in other modes unchanged.
Basic support for velocity updates with the APIC method.
This commit adds APIC to the already existing dropdown menu for the simulation method. The APIC plugin within Mantaflow has been updated to the latest version.
This commit uses an enum to access expansion for specific panels for
each modifier, constraint, etc. Even though these values are quite simple,
this can help make the code more explicit when the ui_expand_flag is
accessed directly. Also update comments about this bitfield to make
them consistent.
Caused by my own cleanup commit rBa308607a5334. Just a simple copy-paste
error. Here the difference between `curve` and `editnurb` makes quite a bit
of difference.
Don't use the current mouse position at the time the event is handled, but
rather the position at the time of the event. This should make e.g. brush
stroke paths more accurate.
In addition, this may solve issues with other software that does mouse
position smoothing. Ref T82143.
Use of the current mouse position was added in 12b642062c as part of a
large commit that also made continuous grab work. But it appears to still
work getting the mouse position from the event.
The system for exposing property settings like min, max, default, subtype,
etc. for ID properties is quite convoluted currently, so I won't give a
full description here, but this commit creates the tree of ID properties
needed to store that information. This means that property subtypes like
"angle" or "XYZ" will affect the display in the modifier.
Limitations:
- The _RNA_UI property is displayed in the modifier. This may require a
modification to uiDefAutoButsRNA to fix.
- IDProperties must have unique names, but node sockets don't have
that limitation. This can be solved by adding a "UI name" field to
IDProperties.
This one was a bit more tricky, because the file loading is
mixed with versioning code and because collections are
embedded into scenes.
All tests that passed before, still pass.
The issue stems from the fact that scene arrays are not cleared when rendering is done. This was not really an issue before the introduction of the ownership system (rB429afe0c626a) as the id_map would recreate scene data arrays based on their new content. However, now that the id_maps do not have access to the scene data anymore the arrays are never created.
Another related issue is that the BlenderSync instance is never freed when the persistent data option is activated.
To fix this, we delete nodes created by the id_maps in their destructors, and delete the BlenderSync instance before creating a new one, so the id_maps destructors are actually called.
Reviewed By: brecht
Maniphest Tasks: T82129
Differential Revision: https://developer.blender.org/D9378
Allow Trackpad Ctrl-Pan to cycle enum values. Also allows cycling in more button types.
Differential Revision: https://developer.blender.org/D8535
Reviewed by Brecht Van Lommel
Changes NanoVDB to be a standalone dependency that is independent of the OpenVDB one.
It works by downloading the "feature/nanovdb" branch of OpenVDB, but using the NanoVDB
CMake in the "nanovdb" subdirectory. Since it is header-only, only the install target is used.
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D9383
When using a trackpad Zoom to Mouse Position would always zoom to center of canvas.
Differential Revision: https://developer.blender.org/D8683
Reviewed by Brecht Van Lommel
Allow project to build without definition of optional UI_PROP_DECORATE.
Differential Revision: https://developer.blender.org/D8370
Reviewed by Julian Eisel
Rather than just printing a message and falling back to the CPU. For render
farms it's better to avoid a potentially slow render on the CPU if the intent
was to render on the GPU.
Ref T82193, D9086
We can't use that for now, because the original mesh might be freed
before the new copied mesh. When the original mesh is freed, the shared
layers will be freed as well.
This can probably be improved in the future, so that we can actually
share attribute arrays between meshes. An approach similar to how
geometries and components are shared should work as well.
A geometry now contains zero or more geometry components.
A geometry component is a subclass off GeometryComponent.
Currently, there is a MeshComponent and PointCloudComponent.
A geometry contains at most one component of each type.
Individual components can be shared between multiple geometries
to avoid unnecessary copies. For that, each component has a user
count that determines when the component will be freed and whether
it is mutable.
Code working with geometries can either work with the components
directly, or use the utility functions on the geometry that cover the
most common operations.
In the future, additional component types can be added. For example,
we'll probably need components for curves and volumes.
Furthermore, something like an InstancesComponent can be added,
which contains points with attributes and references a geometry/object/collection
that is instanced on all the points.
Relabel the Playback Synchronisation menu so that it's clear:
- what does happen (instead of describing what does //not// happen), and
- that the synchronisation options don't just affect audio.
The changes are:
- Change label from "Audio" to "Sync"
- Change the labels of the sync enum:
- No Sync → Play Every Frame
- Frame Dropping stays the same
- AV-sync → Sync to Audio
The "Audio" label has moved one option down, as that option does
actually relate to audio.
Reviewed By: looch, Severin, HooglyBoogly, campbellbarton
Differential Revision: https://developer.blender.org/D9269
Now the behaviors are consistent for blend, image and text files:
- If the file is not writable, will report it.
- If the file is new (without a path), save as will be used.
- If the file was deleted, will try to recreate it.
Ref D6755
Some subtypes never made it to all neccessary places when they were
introduced. This was throwing warnings when accessing such a properties
subtype from python.
The sub_type enums were also defined in 4 different places:
- RNA_types.h 'PropertySubType'
- rna_rna.c 'rna_enum_property_subtype_items'
- rna_rna.c 'subtype_items'
- bpy_props.c as multiple enums
This patch syncs the definitions across all places so that they are the
same everywhere. It also looks redundant to define these twice in
rna_rna.c, now just use 'rna_enum_property_subtype_items' there (and get
rid off 'subtype_items').
Also moved 'POWER' & 'TEMPERATURE' to number enum in bpy_props (these
were defined in the array enum).
Fixes T82167.
Maniphest Tasks: T82167
Differential Revision: https://developer.blender.org/D9371
Wasnt checking the symmetrized 'grab_delta_symmetry' so result could
flip around.
Maniphest Tasks: T82188
Differential Revision: https://developer.blender.org/D9373
The alpha of the first layer was always used
here since introduction in rBee4453f08369 and was not updated when
customdata support for alpha was added.
Now also use the interpolated alpha.
thx @brecht noticing!
ref T81914
Reviewers: brecht, mont29
Maniphest Tasks: T81914
Differential Revision: https://developer.blender.org/D9358
When the combined alpha [the 'tmp' variable having the mixfactor applied
already] - reached zero it was handled like a no-op (for the alpha as
well) and just copied the first color.
So e.g mixing 255/255/255/255 with 0/0/0/0 with a factor of 1.0 gave
alpha of 255, which looks wrong.
cases where tmp gets zero:
src1 alpha:0 src2 alpha:whatever mixfactor 0.0
src1 alpha:whatever src2 alpha:0 mixfactor 1.0
src1 alpha:0 src2 alpha:0 mixfactor whatever
Now set alpha to zero in that case.
ref T81914
Maniphest Tasks: T81914
Differential Revision: https://developer.blender.org/D9357
`sys.executable` is documented to be a Python interpreter or None.
This was set to Blender's executable which caused the multiprocessing
module to spawn new instances of Blender instead of Python on WIN32.
See issue described in D7815.
Deprecate 'bpy.app.binary_path_python' & warn when using.
Blender's executable remains accessible via `bpy.app.binary_path`.
Modified 04c5471cee, setting `sys.executable` instead of using
Py_SetProgramName, which is needed for a bundled Python installation.
Ensure that Zoom does not crash on Mac Trackpad by checking for existence of Continuous Zoom timer.
Differential Revision: https://developer.blender.org/D8682
Reviewed by Julian Eisel
Add "Selection to" as prefix for those menu items that move the selected
keyframes to something, for both the Key → Snap menu and the Shift+S pie
menu.
No functional changes.
This was missed in rB477d983c2e8ab298cbf638d5aadd77fad9c2f677
Differential Revision: https://developer.blender.org/D9304
When outliner datablocks are selected, switch to the corresponding tab
for that datablock in properties editors. Only properties editors
that share an edge with the outliner will change tabs.
Additionally, when modifiers, constraints, and shader effects are
selected from the outliner, the panel will be expanded in all properties
editors.
Part of T77408
Manifest Task: https://developer.blender.org/T63991
Differential Revision: https://developer.blender.org/D8638
Improved calculation of checkbox width to include line width preference.
Differential Revision: https://developer.blender.org/D9025
Reviewed by Hans Goudey
Various changes to some labels and descriptions to be more accurate, clear, or less confusing.
Differential Revision: https://developer.blender.org/D8394
Reviewed by Hans Goudey
Change the sequence of characters shown on the 'Color Grid' generated image.
Differential Revision: https://developer.blender.org/D8603
Reviewed by Hans Goudey
Move popover down a very slight amount to fix clipping of its arrow tip.
Differential Revision: https://developer.blender.org/D9030
Reviewed by Julian Eisel
This reverts commit 2ddecfffc3.
The correct fix is to compile with -fno-strict-aliasing on
release/relwithdebinfo builds also.
See the last commit/ {D9372}.
Maniphest Task T81077
The old code was added in {rBbaa4a9c7d4dd}
While the hotfix committed in {rB2ddecfffc3d3a3a1db4ae45e8665ca} fixed
the id_management test, the correct reason was found at
https://bugs.llvm.org/show_bug.cgi?id=47984
This matches the behavior on Linux.
Now it's possible to trace a sequence of images and not just a single one
When the trace is for more than one image, a bacth job is started to process all frames.
Note: All trace data is generated by Potrace library.
Differential revision: https://developer.blender.org/D9316
The execute callback of a geometry node gets more domain specific
types as parameters now: GeoNodeInputs and GeoNodeOutputs.
Those types are also aware of what node is being executed and can
provide better error messages when they are used incorrectly.
The existing code for this was incomplete. Each instance can now have a set
of attributes stored separately from geometry attributes. Geometry attributes
take precedence over instance attributes.
Ref D2057
This avoids OpenCL inlining heavy volume interpolation code once for every
data type, which could cause a performance regression when we add a float4
data type in the next commit.
Ref D2057
We already had the ability to bake fcurves but no way to convert the
baked result back without using python. This patch adds and operator
that is available now next to the bake operator in the drop down menu,
Reviewed By: Sybren
Differential Revision: http://developer.blender.org/D6379
Not sure why Collection was the only ID for which depsgraph building did
not check for potential recursion?
Reviewed By: sergey
Maniphest Tasks: T82149
Differential Revision: https://developer.blender.org/D9365
Previously, only predefined and limited set of intrinsics combinations
could have been refined. This was caused by a bundle adjustment library
used in the early days of the solver.
Now it is possible to fully customize which intrinsics are to be refined
during camera solving. Internally solver supports per-parameter settings
but in the interface they are grouped as following:
* Focal length
* Optical center
* Radial distortion coefficients (which includes k1, k2, k3, k4)
* Tangential distortion coefficients (which includes p1, p2)
Differential Revision: https://developer.blender.org/D9294
Historically the refine options had a hardcoded list of possibilities.
This was caused by an old bundle adjustment code which did not support
all possible combinations.
Now the bundle adjuster is based on Ceres solver, allowing to refine
anything in any combination.
Introduced recently in 09139e41ed.
While this worked in the cases it was used, '--threads' for example
was failing to parse the number when it's pass was set to 0.
Increase the enum values to start at 1 &
add asserts so this wont happen again.
Avoid passing the pass argument to BLI_argsAdd, instead set this
once for each group of passes.
This means we can add new passes without having to bump the arguments
to BLI_argsAdd.
This function from 2017 came with a comment: "TODO Nuke this once its
only user has been correctly converted to use generic IDmanagement"
Since it is unused after rB91462fbb31ba, now is time to remove it.
Differential Revision: https://developer.blender.org/D9368
`sys.executable` is documented to be a Python interpreter or None.
This was set to Blender's executable which caused the multiprocessing
module to spawn new instances of Blender instead of Python on WIN32.
See issue described in D7815.
Deprecate 'bpy.app.binary_path_python' & warn when using.
Blender's executable remains accessible via `bpy.app.binary_path`.
Moves the Text Editor 'Resolve Conflict' button closer to data-block selector and with 'Question' icon.
Differential Revision: https://developer.blender.org/D9266
Reviewed by Hans Goudey
ASAN reported a use after free in after rB15d78ea85b602c. This commit
removed a reassignment of the `data` variable here that was actually
required because it may have been freed earlier in the function. This
sort of error would be avoided if the same variable wasn't reused for
different purposes.
The new option to filter the layers used by the Fill tool was not using the boundary strokes. The problem was the layers were skipped and any boundary stroke was not used.
Now, the layer is not skipped, but the strokes that are not boundary are skipped.
wiki.blender.org/wiki/Tools/Git#Tips has been updated.
A follow up to
lists.blender.org/pipermail/bf-committers/2020-October/050698.html
will be sent after commit.
Reviewed By: brecht, campbellbarton
Differential Revision: https://developer.blender.org/D9234
The logic of `BKE_sculpt_update_object_for_edit` was not correct. such
low-level functions should typically never preform depsgraph evaluation
themselves, they should be able to rely on getting a fully evaluated
depsgraph and just get needed data from there.
Supporting that required fixing other broken code higher in the
callstack, namely:
* `ED_object_sculptmode_enter_ex` was freeing evaluated data, for no
valid reason it would seem.
* `sculpt_undosys_step_decode` was ensuring an evaluated depsgraph
**before** calling `ED_object_mode_generic_exit`, which would
invalidate a lot of evaluated data.
Note that it is fairly difficult to track down all code paths leading to
`BKE_sculpt_update_object_for_edit`, so there may be still cases where
this gets called with improperly evaluated depsgraph.
Reviewed By: sergey
Maniphest Tasks: T81854
Differential Revision: https://developer.blender.org/D9270
When selecting multiple layers, the redo operator might not correctly
update the pose data. To make sure it is in a good state we have to
ensure that the pose data is good.
Reviewed By: Bastien
Differential Revision: http://developer.blender.org/D9354
The simple subdivision as a type only causes issues like no-continuous
normals across edges, inability to reliably switch the type and things
like this.
The new subdivision operators supports wider variety of how to add
details to the model, which are more powerful than a single one-time
decision on the subdivision type.
The versioning code is adjusting topology converter to specify all
edges as infinitely sharp. The reason for this (instead of using
settings.is_simple) is because in a longer term the simple subdivision
will be removed from Subsurf modifier as well, and will be replaced
with more efficient bmesh-based modifier.
This is finished up version of D8436.
Differential Revision: https://developer.blender.org/D9350
Sub-systems that use directories from BKE_appdir needed to be
initialized after parsing '--env-system-datafiles'.
This meant the animation player needed to call IMB_init it's self.
Avoid this complication by having a pass that can be used to setup
the environment before Blender's resources are accessed.
This reverts the workaround from 9ea345d1cf
The first is a "Point Distribute" node, which takes a mesh and outputs a
pointcloud, scattering points randomly on the surface based on the input
density. The distribution algorithm is extremely basic at this point, and
doesn't take into account an attribute weight at each vertex.
The second node is a "Point Instance" node, which is mainly a placeholder
until the engineering design for how to work with instancing in geometry
nodes is decided. For now it just runs the same code that the convert
pointcloud to mesh operator uses.
This adds another set of the same mesh functions but for pointclouds.
There are probably better ways to generalize this functionality, but that
may have to be rethought in the future anyway if we want to store
multiple of each type. And anyway it's handy to have a specific set of
"pointcloud" functions available in the node implementations.
Improved user experience by using viewport focal length
to calculate the new camera distance.
Also resizing the border to the same aspect ratio as
the window will help not zooming in more than expected.
Ref D9341
This encapsulates Node socket members behind a set of specific methods;
as such it is no longer possible to directly access Node class members
from exporters and parts of Cycles.
The methods are defined via the NODE_SOCKET_API macros in `graph/
node.h`, and are for getting or setting a specific socket's value, as
well as querying or modifying the state of its update flag.
The setters will check whether the value has changed and tag the socket
as modified appropriately. This will let us know how a Node has changed
and what to update, which is the first concrete step toward a more
granular scene update system.
Since the setters will tag the Node sockets as modified when passed
different data, this patch also removes the various `modified` methods
on Nodes in favor of `Node::is_modified` which checks the sockets'
update flags status.
Reviewed By: brecht
Maniphest Tasks: T79174
Differential Revision: https://developer.blender.org/D8544
This studio light preset is designed for color painting tasks. As color
are multiplied on top of the current studio light/matcap, this should be
as white as possible and with very soft speculars to avoid color
distorsion while showing the volume of the mesh.
Reviewed By: jbakker, JulienKaspar
Differential Revision: https://developer.blender.org/D8209
This uses the code from the rewritten boolean modifier from 2.91 as a node.
The implementation is about as minimal as it can get, since for now the use
case of this node is mostly to test multiple geometry outputs and inputs.
The boolean code requires a BMesh, so the node converts the input meshes
to BMesh, so the bmesh is added as a dependency to the nodes module.
The settings can only be set via Python currently. The matching between properties
and group inputs is based on the socket identifier (which is e.g. `Input_5`).
Maybe we'll have to use a different matching strategy in the future, will see.
The properties are not used yet, but can already be accessed with Python
using `modifier.settings['setting name']`.
The plan is to use id properties to store the parameters that the modifier
passes into the geometry node group.
For Customdata layer copying, interpolation with the mixfactor is only
done for certain mix modes, now set the UI inactive if the mixfactor is
not in use.
Namely, the modes are the "Above / Below Threshold" which are only used
for flags, colors and normals and mixing is not supported in these cases.
Spotted while looking into T81914.
Differential Revision: https://developer.blender.org/D9327
Group inputs are not yet exposed in the modifier. For now I just added
a simple float setting that will be passed to every float input of the group.
Inputs of other types have some default value.
Draw the handles for the active keyframe only when the interpolation type
is set to Bézier. This now matches the behaviour of handles of regular
(non-active) keyframes.
Activate an FCurve only on selecting, and not on deselecting a keyframe
or a handle.
Reviewed By: HooglyBoogly, Severin, looch, #animation_rigging
Differential Revision: https://developer.blender.org/D9328
Code was assuming frustrum planes are symmetrical which is not the case
for shifting. This lead to a shrinking region if shift was negative (and
a growing region if shift was positive)
So instead of only keeping track of plane on one side (and mirroring
over in code) get the actual planes after shifting and use these
instead.
This code corrects this for ortho and perspective cameras, it does not
touch panoramic cameras.
Reviewed By: brecht
Maniphest Tasks: T69911
Differential Revision: https://developer.blender.org/D9342
This patch improves the single core performance of the lattice deform.
1. Prefetching deform vert during initialization. This data is constant for
each innerloop. This reduces the complexity of the inner loop what makes
more CPU resources free for other optimizations.
2. Prefetching the Lattice instance. It was constant. Although performance
wise this isn't noticeable it is always good to free some space in the
branch prediction tables.
3. Remove branching in all loops by not exiting when the effect of the loop
isn't there. The checks in the inner loops detected if this loop didn't
have any effect on the final result and then continue to the next loop.
This made the branch prediction unpredictable and a lot of mis
predictions were done. For smaller inner loops it is always better
to remove unpredictable if statements by using branchless code patterns.
4. Use SSE2 instruction when available.
This gives 50% performance increase measured on a
Intel(R) Core(TM) i7-6700 CPU @ 3.40GHz with GCC 9.3.
Also check other compilers.
Before:
```
performance_no_dvert_10000 (4 ms)
performance_no_dvert_100000 (30 ms)
performance_no_dvert_1000000 (268 ms)
performance_no_dvert_10000000 (2637 ms)
```
After:
```
performance_no_dvert_10000 (3 ms)
performance_no_dvert_100000 (21 ms)
performance_no_dvert_1000000 (180 ms)
performance_no_dvert_10000000 (1756 ms)
```
Reviewed By: Campbell Barton
Differential Revision: https://developer.blender.org/D9087
This looks like a optimizer bug where it makes wrong assumptions.
The code inside lib_id_delete:264 on rBafd13710b897cc1c11b
`for (id = last_remapped_id->next; id; id = id->next) {..}`
is not executed in release/relwithdebinfo builds.
This can be "fixed" by several ways:
- Adding a line that prints the `last_remapped_id->name` right before
the said for-loop starts.
- Turning off optimization for the whole function `id_delete`:
`#pragma clang optimize off/on` Ray Molenkamp
- Marking `last_remapped_id` volatile. Julian Eisel
- Marking `tagged_deleted_ids` volatile. But it adds a warning when
calling `BLI_addtail`: discards volatile qualifier. Discovered by
accident.
Fix T81077
Reviewed By: mont29
Maniphest Tasks: T81077
Differential Revision: https://developer.blender.org/D9315
Set property split in the higher level panel functions so that it carries
over to buttons added after. Also discard the redundant "Symmetry"
to make sure there is enough space for the checkbox label.
The now redundant "X" checkbox is removed since it's also present in the
Symmetry panel above. The Topology Mirror is moved into the Symmetry
panel also.
This was needed because `Mesh.use_x_mirror` has recently been turned into
different functionality, and its old functionality now lives under
`Mesh.use_mirror_vertex_group_x`. Something went wrong in the UI in This
transition.
Differential Revision: https://developer.blender.org/D9287
Intern definitions are moved to sequencer/intern/sequencer.h
BKE_sequencer.h was also cleaned up a bit to make sure that functions
and structs are in correct category.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D9259
Recursion happens in case when scene strip point to it's own scene
indirectly by using SEQ_SCENE_STRIPS option.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D9264
It is possible to create scene strips pointing to each other. This is
sanitized when rendering, but in dependency graph such setup will cause
infinite loop.
This patch fixes loop in dependency graph, but same problem exists in
audaspace
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D9262
Main DB and it's structs can point to different address after undoing.
In this case problem was that bmain was not updated. Same fix was
done for scene as well.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D9240
When using local area, all nodes need to build their constraints first
before activating them for simulation. THis ensures that nodes get their
structural constraints from the initial location of each symmetry pass.
Reviewed By: sergey
Maniphest Tasks: T81904
Differential Revision: https://developer.blender.org/D9303
The plane deformation falloff was introduced in the first version of the
cloth brush, but due to the lack of all the new features and fixes in the
solver it was causing a lot of artifacts for deformation brushes. In
order to avoid that, the cloth brush was always using radial falloff for
the grab brush.
Now the plane falloff is properly implemented using the deformation
constraints.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D9320
This adds a tool property for sculpt line gesture tools (line and
project) to limits its effect to the segment of the gesture instead of
using the infinite line to bisect the mesh in two parts.
To achieve that, the line gesture now has two extra side planes that can
be enabled/disabled for getting the nodes from the PBVH and to test the
vertices.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D9307
This adds an operation mode to the Face Set Edit tool which deletes the
geometry of a Face Set by clicking on it.
The operator also checks for the mesh having a single Face Set to avoid
deleting the entire object by accident.
This is also disabled for Multires to avoid modifying the limit surface
without control (it is not an important limitation as base meshes for
multires are usually final, but maybe it can be supported in the future).
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8938
This commit contains some improvements to this function to make this
function more purposeful and readable.
- Split updating information of the old button to a new function.
- Remove some 7 year old code disabled with `#if 0`.
- Add comments explaining some of the less obvious aspects.
Differential Revision: https://developer.blender.org/D9117
The previous fix forgot the case where there is an intermediate
edge and everything isn't in one plane.
Differential Revision: https://developer.blender.org/D9336
Approximately 141 changes of capitalization to conform to MLA title style.
Differential Revision: https://developer.blender.org/D8392
Reviewed by Julian Eisel
The versioning code to default to old booleans for old files was
faulty because really old files had a 'solver' field (later removed,
but then added back for new booleans).
The versioning code to default to old booleans for old files was
faulty because really old files had a 'solver' field (later removed,
but then added back for new booleans).
This commit uses continue in loops and returning early to reduce
indentation in long functions, only where this results in a significant
improvement. Also includes a few LISTBASE_FOREACH macros.
Changing language could sometimes leave File Browser System list showing incorrect text until restart.
Differential Revision: https://developer.blender.org/D9323
Reviewed by Brecht Van Lommel
The operator was using a secondary python operator to ask parameters before running, but this can be done in invoke.
Differential Revision: https://developer.blender.org/D9330
Don't allocate a new buffer for refitting meshes, but update the existing one.
It's not clear from the API docs if this is required, but it appears to solve
the issue and should be more efficient.
Activate an FCurve only on selecting, and not on deselecting a keyframe
or a handle.
Reviewed By: HooglyBoogly, Severin, looch, #animation_rigging
Differential Revision: https://developer.blender.org/D9328
Avoid accessing mesh emitter and hair at the same time. This is not ideal for
performance, but once we have a dedicated hair object this will resolve itself.
Differential Revision: https://developer.blender.org/D9322
Draw the handles for the active keyframe only when the interpolation type
is set to Bézier. This now matches the behaviour of handles of regular
(non-active) keyframes.
We might not want to have all those nodes in a final version.
Some of them have been added with particle nodes in mind.
However, to test the evaluation system it is useful to have a
couple of nodes available.
Those nodes should "just" work, because their implementation
is reused from the particle nodes project.
This adds support for nodes that have a multi-function implementation.
That includes various function nodes like Math, Combine Vector, ...
Furthermore, there is support for implicit conversions now. So it should
work when one connects e.g. a float to an integer and vice versa.
Caused by rB7878adf49cff.
When getting the stroke location via raycast in ortho view, the above
commit flipped the condition of the check to perform adjustments on the
rays start/end. This would thus happen (even though it shouldnt),
resulting in wrong depth and stroke location.
Now just flip the condition back, so adjustments only happen when
clipping is OFF.
Maniphest Tasks: T81934
Differential Revision: https://developer.blender.org/D9318
These two functions "snode_notify" and "ED_node_tag_update_id" appear to
be mostly duplicates. However, there is already a case for each type of
built-in node tree, so it makes sense to add one for the geometry node
tree as well. This doesn't solve the update issues for changing number
in buttons, that must be handled somewhere else.
GPencil: Change Interpolate shortcut to Ctrl+E
Before the shortcut was Ctrl+Alt+E, but it's more logic remove the Alt.
This was missed in rBee49ce482a797a5937829de497abd69bcd1edb48
GPencil: Change Interpolate shortcut to Ctrl+E
Before the shortcut was Ctrl+Alt+E, but it's more logic remove the Alt.
This was missed in rBee49ce482a797a5937829de497abd69bcd1edb48
This regression was caused by rB57de5686048f which disabled srgb
transform after the python callback.
The right thing to do is to only rebind the framebuffer once to
reset the no-srgb override.
For objects with shared data, it makes sense to show the mode icon for
every object sharing the same data if one of them is in edit mode.
This also disables the "extend" functionality in this case, because
being in edit mode for multiple objects with the same data isn't
supported.
Differential Revision: https://developer.blender.org/D9273
There was a weird bug in the API where a value of 0 gave a mask value of
1. I am not sure why this is but the current code works as desirable.
This was missed in rB6faa765af8954948de3cec75a2261a5aa139b4e5
There was a weird bug in the API where a value of 0 gave a mask value of
1. I am not sure why this is but the current code works as desirable.
This was missed in rB6faa765af8954948de3cec75a2261a5aa139b4e5
Move the checks for whether to draw the button to the beginning of the
function and return early. Also use a shorthand variable for ob_active.
Committing to 2.91 as a patch for an upcoming bug fix depends on these
changes.
Differential Revision: https://developer.blender.org/D9272
The solo mode was skipping the layer creation data and the loop of masks expect to have all layers in the array or the loop crash.
The solution is just create the layer array data for the layer, but don't draw any stroke.
This adds a property to the grab that masks vertices based on its
original normal and grab delta. When used on thin meshes, it allows to
grab the silhouette from one side of the object without affecting the
shape of the other side.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D9205
Paint and smear tools are only implemented for regular mesh PBVH,
meaning they are not supported by the dynamic topology and multires
sculpting.
These tools are to be ignored for an unsupported sculpt modes, regardless
of state of user preferences.
Reviewed By: sergey
Maniphest Tasks: T81932
Differential Revision: https://developer.blender.org/D9308
This adds a layer of abstraction between the code calling callbacks
on sockets and the implementation of those callbacks.
This abstraction layer allows some sockets to not implement all
callbacks when those can be derived from some other callback.
The new parameter allows to define if after cutting the stroke the cap of the cut side will be set as flat.
Before, the cap shape of the cut side always was equal to the original stroke, and in some situations, the rounded cap was visible.
Note: If the angle of the join is very extreme, it's still possible to view some sections of the cut stroke.,
All tools planned for 2.91 now have icons, so this option can be
removed.
Reviewed By: dfelinto, Severin
Differential Revision: https://developer.blender.org/D9299
From the backtrace it looks like in some cases file save (which triggers
a general override updates) is done before other code has a chance to
re-generate pose data, leading to rna accessing freed memory.
I was never able to reproduce that here, so this is a tentative fix in
master, if it proves to be working for the studio it will be
cherry-picked into 2.91 release branch later.
This is still very basic and does quite a few unnecessary computations.
Also the error handling is quite weak currently, so when invalid things are
connected, it will probably just crash.
Also the interface that individual nodes have to implement will have to change,
but the current solution is a good starting point.
Only the triangulate node is implemented for now.
It was rather a huge chunk of code, which started to become
more harder to maintain with the transition to OpenSubdiv based
implementation. Because of this transition, the compatibility was
also rather on a poor side.
Remove compatibility support for pre-2.50.9 multires.
Ref T77107
Reviewed By: brecht, mont29
Differential Revision: https://developer.blender.org/D9238
Historically the result of the keying node was violating alpha
pre-multiplication rules in Blender: it was simply overriding
the alpha channel of input.
This change makes it so keying node mixes alpha into the input,
which solves the following issues:
- The result is properly pre-multiplied, no need in separate
alpha-convert node anymore.
- Allows to more easily stack keying nodes.
This usecase was never really investigated, but since previously
alpha is always overwritten it was never possible to easily stack
nodes. Now it is at something to be tried.
Unfortunately, this breaks compatibility with existing files, where
alpha-convert node is to be manually removed.
From implementation side this is done as a dedicated operation since
there was no ready-to-use operation. Maybe in the future it might
be replaced with some sort of vector math node.
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D9211
`imb_gpu_get_data` could reuse `data_rect` when it was already in used (double alloc).
making the first use leak. This was detected after enabling OpenGL Texture
Limit.
Reviewed By: Clément Foucault
Differential Revision: https://developer.blender.org/D9280
This avoids accidents using user-preferences in the main BLF API,
which could cause preferences to be used unintentionally
(such as stamping into renders or creating generated images).
As well as uses of BLF when preferences aren't loaded
such as animation playback.
Selecting an F-Curve handle caused an assertion as well as treating
the key-frame as inactive.
Allow active the keyframe to be active when it's handle is selected,
as is done with bezier curves.
These changes should result in more readable and undestandable code,
especially where while loops were use instead of for loops. They are
not comprehensive, and I skipped wherever the change was not obvious.
In practice, there are only a limited number of operations we need to
use repeat such as navigation, stepping operations that cycle states
and text input.
Now we don't need to disable repeat explicitly when a modal operator
uses checks for a key being held as was needed for 17cb2a6da0.
Repeat is now included in exported keymaps.
Use versioning so existing exported keymaps are loaded properly.
This was caused by unprotected drawing callbacks.
As of 2.91, we require that all python callbacks used for
drawing needs to be safeguarded by `GPU_bgl_end()` to end the
state tracking override.
This was caused by unprotected drawing callbacks.
From 2.91, we now require that all python callbacks used for
drawing needs to be safeguarded by `GPU_bgl_end()` to end the
state tracking override.
This avoid strange discrepency between the general purpose variant and
the specialized glass variant which did not have a way to turn
multi-scatter off.
This patch helps the case of intricate reflections where the
ray does not travel far before intersecting the geometry.
In these cases there could be false negative exclusion of the ray
caused by the backface rejection threshold.
The artifact manifested as lines of different values caused by faillure to
trace the depth buffer correctly.
Adding a ad-hoc value to the step size to mitigate the issue.
Outliners listener (outliner_main_region_listener) needs ND_PARENT
notifier to redraw, the parenting operator only spawned ND_TRANSFORM
(which doesnt do a redraw).
Maniphest Tasks: T81896
Differential Revision: https://developer.blender.org/D9295
nothing
Caused by rB6faa765af895.
Since above commit, we have to use paint.mask_box_gesture instead now.
Maniphest Tasks: T81926
Differential Revision: https://developer.blender.org/D9300
shapekeys/modifiers
This was failing for all mask filters (sharpen, grow, invert, clear,
shrink, contrast, smooth) and mask gestures (box, lasso).
Also have to recalc shading, use SCULPT_tag_update_overlays for this.
ref D8956
Maniphest Tasks: T81929
Differential Revision: https://developer.blender.org/D9302
This improves performance in scene synchronization when there are many
mesh, hair and volume objects. Sync time speedups in benchmarks:
barbershop 5.2x
bmw 1.3x
fishycat 1.5x
koro 1.0x
sponza 3.0x
victor 1.4x
wdas_cloud 0.9x
Implementation by Nicolas Lelong, and Jagannadhan Ravi (AMD).
Differential Revision: https://developer.blender.org/D9258
Line gesture use always the right side of the line as active (the area
of the mesh that is going to be modified) by default.
This adds the ability to change the active side when the line gesture is
active by pressing the F key.
This allows more freedom to position the line after starting the
gestures, as it won't be required to cancel the operation or undo if the
line was used in the wrong direction.
Reviewed By: Severin
Differential Revision: https://developer.blender.org/D9301
Line gesture use always the right side of the line as active (the area
of the mesh that is going to be modified) by default.
This adds the ability to change the active side when the line gesture is
active by pressing the F key.
This allows more freedom to position the line after starting the
gestures, as it won't be required to cancel the operation or undo if the
line was used in the wrong direction.
Reviewed By: Severin
Differential Revision: https://developer.blender.org/D9301
The new preset I made for 2.91 is way more controllable with lower
strength values and does not have the accumulate bug, but until the
brush management is in place to ship multiple versions of the brush,
probably most people expect something closer to the old version to be
the default.
Reviewed By: sergey
Maniphest Tasks: T81901
Differential Revision: https://developer.blender.org/D9289
Swaps the order of the '+' and '-' button in the File Browser file name field,
so that '-' comes first.
For increasing or decreasing a value it makes more sense to have decreasing
first, increasing last. Consistent to how you press on the left side of a
number button for decrease, and right to increase.
However this is inconsistent in another way: Usually we have a '+' button
before a '-' button, but that refers to adding and removing items, not
increasing or decreasing. The icons are also placed in their own buttons then,
making them look more separate.
So the UI Team agreed on accepting that trade-off, see today's meeting notes:
https://devtalk.blender.org/t/2020-10-21-ui-team-upcoming/15849
For some reason the layout code doesn't center the search button properly. Add
a blank icon button to add some padding, dynamically resized as the region size
changes. This is quite finicky and not at all perfect. But it makes the search
button look far less off-place.
The pin button should be next to the data-path, which is what it belongs to.
Note that this makes the placement of the search button in the header look
quite off. That is because it's centered to the absolute header width, not the
width of the main region (which is smaller because of the tab region on the
left).
Technically it's correct that way, visually it looks wrong. This will be
addressed in a followup commit.
When the stroke was not flat, the bounding box projected could not be right and the strokes could not be painted or sculpted.
Now, the 2D bounding box is calculated using the extremes of the 2D bounding box and not the original 3D min and max values.
In order to prevent the panel code from using the type after it is freed,
the field needs to be set to NULL. This needs to be done recursively
for subpanels as well as top-level panels.
Those two features are not directly related and one might be activated
in master earlier than the other.
WITH_PARTICLE_NODES was removed, because we continue the project
under the name "Geometry Nodes".
Until it is decided whether to work on, or ignore these
warning, disable them. See T78535
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D9281
Currently render passes in the draw manager (eevee) must be predefined
in the render result. This patch would ask the render engine for the
render passes it needs, and create these as a preparation step during
rendering. This allows any draw engine to define more complex render
passes setup.
Render passes can only be added before the call to `RE_engine_begin_result`.
`RE_engine_begin_result` makes a full copy of the render passes. During
rendering the render engines renders to the duplicated passes.
`RE_engine_end_result` syncs the data back to the original render passes,
but only if the passes existed in the original render result.
Currently we work-around this issue by registering the passes in
`render_result_new`. This is legacy blender internal structure and should
be avoided.
With upcoming projects (AOV/Cryptomatte) it becomes a bit of a mess as we
are extending legacy code to support new features. This patch allows us to
let each draw engine register their own render passes at render time
(similar to cycles and other render engines). In the future we could get
rid of legacy render passes registration in render_result_new.
Reviewed By: Clément Foucault
Differential Revision: https://developer.blender.org/D9088
Replace ED_transform_snap_object_context_create with
ED_view3d_give_base_under_cursor & ED_view3d_autodist_simple so object
visibility is respected and non-geometry objects can be supported.
Ref D9255
Fixes error in determining 3DView Overlay Line Height. Do not base on current default font height.
Differential Revision: https://developer.blender.org/D9288
Reviewed by Campbell Barton
17cb2a6da0 missed a break statement after a nested switch,
while it didn't cause a bug nesting switch doesn't read well
and is prone to errors like this.
Split modal-keymap checks into their own branch to avoid this happening,
also use matching event checks for all gesture operators.
Being able to adjust the distance between fluid and obstacles comes in handy when trying to achieve a fluid motion over inclined obstacles.
Depending on the slope of such obstacles, already small adjustments of this value can help when particles stick to obstacle surfaces (i.e. make particles not stick to obstacles).
This adds support for snapping for line gesture tool. It is implemented
in the modal keymap as Snap, which is a toggle (similar to how snapping
in the transform operator works).
Right now it snaps the angle of the line to 15 degree increments, which
is defined in code. This should be easy to expose in the UI in the future
if we need to.
Reviewed By: Severin
Differential Revision: https://developer.blender.org/D9115
This patch adds a modal key to move the selection box/lasso/line while
drawing it. It also sets "repeat": False on the animation playback key
to prevent accidental playback if the spacebar is released after the
mouse button.
Reviewed By: #user_interface, pablovazquez, Severin
Differential Revision: https://developer.blender.org/D9227
This patch fix most self intersection comming from reflection rays.
We regenerate the ray if it goes below the shading normal (should be the
geometric normal but we have no access to it here).
Also add the same precision based bias we use for contact shadows.
This fix T81105 Eevee SSR quality regression in 2.91 alpha
SCULPT_tilt_apply_to_normal expects a normal, and offset was already
scaled by radius. The funcion returns a normalized vector, so the
strength of the brush was changed.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D9275
This patch adds icons for the multires displacement
eraser tool in sculpt mode
Reviewed By: HooglyBoogly
Differential Revision: https://developer.blender.org/D9286
This patch adds icons for line gesture tools
Line mask and line project
Reviewed By: HooglyBoogly
Differential Revision: https://developer.blender.org/D9285
In comments made by a tester on rB17b89f6dacba007bf, it seems that baking
of the spray maps would be useful. This commit adds that capability. Both
the spray map and its inverse are baked out in this change, for maximum
convenience and to avoid assuming what the user wants.
Differential Revision: https://developer.blender.org/D8470
Shorten name of 'Inverse Cursor Color' to 'Inverse Color' so that it does not overflow its popover.
Differential Revision: https://developer.blender.org/D9274
Reviewed by Brecht Van Lommel
With D8234 a new drawing method for UV/Image editor was introduced. For debugging
reasons we left the old drawing method in the code base. This patch will remove
the old drawing method.
Reviewed By: Clément Foucault
Differential Revision: https://developer.blender.org/D9011
Caused by rBaafd71a8a160.
In the process of CustomData Correction, we need to make sure we also
have matching layer names [as was done before above commit], otherwise
this will create layers with default names, applying
(mesh_customdatacorrect_apply and friends) will fail then.
Maniphest Tasks: T81865
Differential Revision: https://developer.blender.org/D9278
The alpha out socket output the average transmittance, not the alpha.
This patch will convert the transmittance to alpha.
Found during research of T80919; Issue introduced when `Closure.opacity` was migrated to `Closure.transmittance`.
Reviewed By: Clément Foucault
Differential Revision: https://developer.blender.org/D9010
Use the newly introduced packed intrinsics, which allows to remove
code which was initializing parameters block based on distortion
model type.
Now such initialization is done by a particular implementation of
a distortion model.
Differential Revision: https://developer.blender.org/D9192
This is a common class which can be used in all sort of minimization
problems which needs camera intrinsics as a parameter block.
Currently unused, but will replace a lot of hard-coded logic in the
bundle adjustment code.
This fixes the crash and the assert, but undo is done in multiple steps
Reviewed By: campbellbarton
Maniphest Tasks: T81076
Differential Revision: https://developer.blender.org/D9256
Version reported by package manager is not always satisfying (on Debian
testing currently e.g. `llvm-dev` is reported as `9.0`, when exact one
is actually `9.0.1`, this break CMake build of Blender then).
Just use version reported by `llvm-config` instead, when using
distribution packages.
The design for how we approach the "Everything Nodes" project
has changed. We will focus on a different part of the project initially.
While future me will likely refer back to some of the code I remove here,
there is no point in keeping this code around in master currently.
It would just confuse other developers working on the project.
This does not remove the simulation modifier and data block. Those are
just cleaned up, so that the boilerplate code can be reused in the future.
This changes how the simplify volumes setting works. Before, it only
affeted viewport rendering. This was an issue, because all internal
computations would still have to happen on the high resolution volumes.
With this patch, the simplify setting already affects file loading and
procedural generation of volumes.
Rendering does not have to care about the simplify option anymore,
it just gets the correct simplified version from the depsgraph.
Reviewers: brecht
Differential Revision: https://developer.blender.org/D9176
When creating the boundary edit data, the loop can stop because a new
vertex was found further from the boundary than the brush radius or
because all vertices of the mesh were already processed. In this second
case, the max_propagation_step was not set, so the code that laters
calculates the falloff was not working, preventing the mesh from
deforming.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D9215
This adds a smooth deformation type to the boundary brush, which smooths
the boundary and has a regular falloff towards the inside of the mesh.
For smoothing, only vertices parallel to the boundary are taken into
account, creating this effect.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D9204
Now that sculpt mask and face sets can also be drawn without using the
PBVH, these operators need these extra updates when the data changes.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8956
Previously the base displacement for solving the constraints was always
using 0.5, which may introduce artifacts when multiple constraints of
different types are computed for the same vertex. This introduces a
factor that reduces the base displacement of the solver, reducing the
artifacts.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D9202
Menus with categories gain a dividing line and omit the title.
Differential Revision: https://developer.blender.org/D5135
Reviewed by Brecht Van Lommel
Slight adjustments to widths, and adds gaps below titles, of 'Viewport Gizmos' and 'Object Types Visibility' popovers.
Differential Revision: https://developer.blender.org/D9257
Reviewed by Pablo Vazquez
Corrects incorrect usages of the fragment 'apart of' when 'a part of' was required.
Differential Revision: https://developer.blender.org/D9245
Reviewed by Campbell Barton
Corrects incorrect usages of the word 'loose' when 'lose' was required.
Differential Revision: https://developer.blender.org/D9243
Reviewed by Campbell Barton
Only build avx/avx2 unit tests if supported by the compiler and
WITH_CYCLES_NATIVE_ONLY is off, otherwise the appropriate compiler flags
are not available.
Corrects incorrect usage of contraction for 'it is', when possessive 'its' was required.
Differential Revision: https://developer.blender.org/D9250
Reviewed by Campbell Barton
Although the `ELEM` macro wraps logic into parentheses, it's not intended to be
used that way. Unexpanded macros should still follow regular coding style for
readability and for tools parsing the code (it confused clang-format for
example).
This is a stricter version of the change made in
{rBbb872b25f219d1a9bc2446228b6dc}
Cmake must never look into Frameworks when the system
library guards (`without_system_libs_begin`/`without_system_libs_end`)
are present.
OpenAL didn't follow this and OpenAL.framework in Xcode would be used.
The Cmake's `FindOpenAL.cmake` looks for both library (in this case,
the .framework file), and include dir.
Precompiled libraries don't contain the former. So `find_package`
cannot be used, or it becomes the hack that {rBb2c707747da9} removed.
So hardcode the include dir path, and other variables.
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D9267
Issue introduced in rB4f616c93f7cb.
Issue here is that the the `StrokeCache` `mirror_symmetry_pass` is still
in its previous state when entering
`wpaint_do_symmetrical_brush_actions`.
For the initial stroke this means that the (wrong) cache
`mirror_symmetry_pass` ends up in SculptBrushTest `mirror_symmetry_pass`
as well and thus the clipping test in `sculpt_brush_test_clipping` will
fail.
[ This one flips the coords to test against clipping according to (now
wrong) `mirror_symmetry_pass` ]
Solution seems simple: just ensure we start of with a
`mirror_symmetry_pass` of zero in `wpaint_do_symmetrical_brush_actions`
for the initial stroke.
Same thing is done for vertex paint as well.
Maniphest Tasks: T81484
Differential Revision: https://developer.blender.org/D9268
Add "Selection to" as prefix for those menu items that move the selected
keyframes to something, for both the Key → Snap menu and the Shift+S pie
menu.
No functional changes.
There would be eigen alignment issues with the custom libmv vector
class when compiling with AVX optimizations. This would lead to
segfaults.
Simply use the std::vector base class as suggested by the old TODO in
the class header.
Reviewed By: Sergey
Differential Revision: http://developer.blender.org/D8968
This modifier is the opposite of the recently added Mesh to Volume modifier.
It converts the "surface" of a volume into a mesh. The "surface" is defined
by a threshold value. All voxels with a density higher than the threshold
are considered to be inside the volume, while all others will be outside.
By default, the resolution of the generated mesh depends on the voxel
size of the volume grid. The resolution can be customized. It should be
noted that a lower resolution might not make this modifier faster. This
is because we have to downsample the openvdb grid, which isn't a cheap
operation.
Converting a mesh to a volume and then back to a mesh is possible,
but it does require two separate mesh objects for now.
Reviewers: brecht
Differential Revision: https://developer.blender.org/D9141
Related to T81199. When saving a rendered image with transparency (RGBA)
to a 3 channel PNG image the emissive colors were not exported. This
change adds the emissive colors to the written file.
NOTE: this does not fix the limitation of writing emissive colors to a 4
channel PNG file as the file format does not support this.
Fixed a couple of typos in comments in CMakeLists.txt and GNUmakefile
Reviewed By: #platforms_builds_tests, mont29
Differential Revision: https://developer.blender.org/D9261
Issue caused by {9582797d4b50} in b2.90. The surface per material used
an index buffer owned by the batch. These index buffers are created at
the same time the surface tris index buffer was created. When a material
per batch buffer was invalidated it used the surface tris index buffer
rendering all materials on all surfaces making the last draw command
render succeed.
This patch stores the surface tris per material in the cache so they can
be reused. There is also no need to use the `saved_elem_ranges` anymore as they are
now part of the cache.
The ugly bit of the implementation is that in `extract_tris_finish` the
MeshBufferCache is retrieved. But as this part was already documented as
a hack and it is something that is only used for final meshes. Other
solutions would impact performance or made the fix not condensed
(passing parameters that shouldn't be used).
Reviewed By: Clément Foucault
Differential Revision: https://developer.blender.org/D9136
Field time_base of video stream must be set for some containers,
otherwise avformat_write_header() will set it to default values.
Rendered file in such case won't be played at desired frame rate.
See init_muxer() in mux.c in ffpmeg sources.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D9213
rB78a5895c96 introduced a "use after scope" warning, where a buffer
from a lower scope was used later. The solution is to only use one
variable and store whether to use it more explicitely with a bool.
The code for Bevel's percent (and absolute) modes were pretty bogus.
It assumed, like the rest of the modes, that the offset lines are
parallel to the beveled edge. Which is not true for these modes,
though it accidentally works sometimes if the legs are equilength.
Also the clamping code for those modes was completey wrong.
It is too hard to really fix the clamping code for absolute mode,
but it is a little better now. Percent mode clamping is fixed.
A general refactor / fix commit that should clear out the issues that have been reported on external forces and moving effectors (e.g. T79537, T81660, T80088).
Setting this type is required to prevent fluid particles from being treated like physics particles. The actual fix for this was made in rB11a8a6d0e6b5.
Scale widths of popovers and dialogs with Text Style font point changes.
Differential Revision: https://developer.blender.org/D9132
Reviewed by Hans Goudey
The "active_pose_bone" context variable isn't accessed from
`buttons_context`, it uses `screen_context` instead. This means
it can't account for pinning in the property editor.
Using "pose_bone" context variable first means the property
editor context will be used and the pinning will work.
Differential Revision: https://developer.blender.org/D9242
Code could call CustomData_get_layer_index_n with a negative index (if
no active and/or render UV layers are found). This would assert since
rBe86785c51445.
Spotted while looking into T81398.
Differential Revision: https://developer.blender.org/D9212
Caused by rBe65c78cd43aa.
Since above commit, only geometry and lights received the update,
previous to this check an instancer would receive that is well (in case
of 'is_updated_shading').
Now check for an instancer (checking OB_DUPLI via ob.is_instancer()) and
do an update then as well.
Reviewers: brecht
Maniphest Tasks: T81729
Differential Revision: https://developer.blender.org/D9222
This was caused by canceling operator if strip has more than 0 inputs.
Logic should be reversed - cancel only if strip has 0 inputs.
BKE_sequencer_render_loop_check() arguments had to be sanitized because
seq_effect_find_selected() can set seq1,2,3 to NULL
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D9197
- Expand the "Type" toggle at the top. This is consistent with other
modifiers where there is a "Type" option at the top. It conveys the
property's importance and makes it faster to switch it.
- Expand the "Delimit" option vertically so the text isn't squashed.
There isn't enough space on one line for this, and is has to be
expanded because more than one option can be selected. This is also
consistent with how "multi-select" enums are often displayed, like
the 3D view snapping settings.
| Before | After |
| {F9000996} | {F9000985} |
This patch implements panel expansion saving and resetting for property
search. While search is active, the panel expansion is based on whether
or not it has a search result. When the search finishes, the panel
expansion returns to its state before the search started. However, any
panels interacted with during the search won't reset their expansion.
This requires adding a new runtime flag for panels to store whether to
use search result status as expansion. It also requires better handling
for animation when panel expansion changes with another new runtime flag.
`UI_panel_is_closed` gets the search-dependent expansion, but it is
intentionally not used to access expansion in every case-- sometimes it's
necessary to use `PNL_CLOSED` directly.
Differential Revision: https://developer.blender.org/D8984
Add operator to snap the 2D Cursor value to selected keyframes. This is
doing almost the same as the "Cursor to Selected" operator, except that
it doesn't affect the current frame, just the Y-coordinate (the value)
of the 2D cursor.
The "snap cursor" operators are added to the Key → Snap menu and to the
Snap pie menu. This means that these menus are now extended in meaning,
to not only mean "snap the selected keyframes to the cursor", but also
for some options "snap the cursor to selected keyframes".
This fixes T76596.
Split up `GRAPH_OT_frame_jump` exec function and added some local variables
to give names to the cryptic `ked.f1`, `ked.f2`, and `ked.i1`.
No functional changes.
Remove sculpt/paint checks in getTransformOrientation_ex
This code goes back a long time (early 2.5x). I couldn't find any
reason why sculpt/paint checks were being made.
This makes the following changes:
- When in object mode, the object must be selected.
Since this function typically operates on the selected items.
- When in paint/particle modes, the objects matrix is always
used regardless of selection, since object selection can't be
controlled in these modes.
- When there is no active object, the first selected object is no
longer used as it's quite an arbitrary decision & not something
done elsewhere with objects in Blender.
Basically first letter of (most) words is to be capitalized.
These settings (Noodle curving and Grid levels) had this wrong since
their first commit (2011 and 2020 respectively).
Remove the code that synchronises selection state of shader node animation
channels. This code is only used in a few cases where selection of these
animation channels is changed, and then potentially does the wrong thing
and disallows selection of animation channels altogether.
This removal is meant to be a temporary situation, to unblock animation
channel selection. See T74159 for the overall effort to improve selection
sync.
This lets add-on authors avoid false positive leaks when exiting.
In particular GPUShaders's although it applies to any PyObject that
stores memory allocated by guarded-alloc.
While this does add overhead on exit, on my system it's
under 1/100th of a second with all addons enabled.
See: T71362
TBB includes Windows.h which defines a min/max macro
leading to issues when you want to use std::min and
std::max.
This change prevents Windows.h from defining them
sidestepping the issue.
Needed after d1b3439b80. Think the error only happened with fresh builds,
where dna_type_offsets.h didn't already exist.
We have to do the same in other places too, see 8594cdb456.
The `outliner.item_rename` operator needs to decide if it should rename the
active or the hovered item. Previously it checked if the event is a press
event, which is a hacky way of doing this and limit how the operator can be
used in the keymap.
Now use a operator option to let this be controlled on the keymap level.
Doesn't change any default behavior.
* Use existing and optimized lookup function, rather than own duplicated logic.
* Move low-level coordinate check into general function, alongside similar
ones.
* `do_outliner_item_rename()` enables the rename mode for the item under the
cursor. Issue is, collapsed children end up having stored the same coordinate
as their parent, so they too would get the rename mode enabled (there is no
early-exit that would hide this).
* The items displayed as inline icons do get the proper coordinates of the
icons, so they are not mistaken as being under the cursor.
After rBb077de086e14, the Outliner tree is rebuilt less often, so the
coordinates are cleared less often too.
As far as I can see we can always clear coordinates of invisible items now. No
code seems to depend on keeping the old coordinates anymore.
The new parameter made so that previously keyed Alpha values were lost
and instead the new "Emission Strength" was keyed.
Issue introduced with the original commit of Emission Strength: b248ec9776
Note: Files created since the issue (September 17) that keyframed the
Emission Strength will have to fix their files manually.
Differential Revision: https://developer.blender.org/D9221
The cloth brush grab mode was creating constraints at 1.0 strength in
the area of the brush where the fade was evaluated to 1. This was causing
stability issues in the simulation and not producing ideal results.
Now the constraint strength is scaled with an empirically found factor.
The values in this patch may require further tweaking after experimenting
a little bit more with them.
Reviewed By: sergey, zeddb
Differential Revision: https://developer.blender.org/D9201
Restore the old `correct_bezpart()` (pre-rBda95d1d851b4) function as
`BKE_curve_correct_bezpart()`, and use that where the old behaviour was
desired (that is, curve maps like used by the RGB Curves shader node).
The new (post-rBda95d1d851b4) function is also renamed to
`BKE_fcurve_correct_bezpart()` to avoid confusion.
When using anchored stroke, the stroke operator was modifying the
coordinates on the "mouse" rna property by setting them to the original
position. Because of this, all the sculpt delta calculation was failing
and the delta for these brushes was set randomly (with a 0 vector) at
the beginning of the stroke.
There is now an extra property that uses the unmodified coordinates of
the mouse to calculate the delta. Now delta orientation works as expected
in all brushes and features that require brush tip orientation.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D9183
Dynamic area should use the radius instead of the initial radius to get
the nodes as the radius can now change during the stroke. In case of
anchored strokes, the current radius can be bigger than the initial
radius, simulating vertices outside the falloff area and breaking the
mesh.
Reviewed By: sergey
Maniphest Tasks: T81649
Differential Revision: https://developer.blender.org/D9181
Previously, all Face Set visibility logic was using mvert flags directly
to store the visibility state on the vertices while sculpting. As Face
Sets are a poly attribute, it is much simpler to use mpoly flags and let
BKE_mesh_flush_hidden_from_polys handle the vertex visibility, even for
Multires.
Now all operators that update the Face Set visibility state will always
copy the visibility to the mesh (using poly flags) and the grids, all
using the same code.
This should fix a lot of visibility glitches and bugs like the following:
- Sculpt visibility reset when changing multires levels.
- Multires visibility not updating in edit mode.
- Single face visibible when surrounded by visibile face set, even when
the face set was hidden.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D9175
ME_POLY_LOOP_NEXT and ME_POLY_LOOP_PREV expect the offset of
the loop in the poly as an argument, in other words, corner index of the poly.
This was violated in some places. It didn't cause issues when base mesh consists
of only quads due to the way how modulus worked inside of the macro. However,
if mesh had non-quad faces adjacency information was returning wrong vertex
indices. This was causing multiple brushes to work erratically, including brushes
like Face Set, Boundary automasking, mesh relax, and others.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D9173
When using the sculpt filters, global automasking settings that affect
all brushes were ignored because the automasking system was not
implemented for filters, making filters and brushes react differently
to the global sculpt settings which creates confusion.
This makes all filter tools (mesh, cloth, color) use the same general
automasking settings and features as the brush tools. Filters will now
use the settings in the options panel to limit their effect.
This also removes the "use Face Sets" option from the Mesh filter code,
as it was duplicated from the automasking code just to have that
funcitonality. This is now handled by the regular automasking system.
The "Use Face Sets" option is still available in the cloth filter as that
option limits the action of the forces, not the displacement.
After this, it is possible to initialize the automasking system
independently from the StrokeCache and Brush settings, so it can also be
added to more tools and features in the future.
Fixes T81619
Reviewed By: dbystedt, sergey
Maniphest Tasks: T81619
Differential Revision: https://developer.blender.org/D9171
This adds an operator property to use the paint cursor radius and
position for the depth of the trimming shape created by the trimming
tools.
When enabled, the shape is located in the surface point when the gesture
started and it will have the depth of the cursor radius. When the cursor
is not over the mesh, the shape will be positioned at the center of the
depth of the whole object from the viewport camera.
Reviewed By: dbystedt, sergey
Differential Revision: https://developer.blender.org/D9129
This patch enables the paint brush cursor with all tools in sculpt mode,
even with the ones that are not brushes. The motivations for this change
are:
- The filters are using the position of the active vertex for certain
features without any visualization of what the active vertex is.
- You can call operators like mask expand (which depends on the brush
cursor position and active vertex) with a non brush tool enabled.
- Having the cursor in the rest of the tools allows to have a scene
scale representation of the radius and a direct control of radius and
strength (using the current brush shortcuts), which will allow to make
features more intuitive without relying on modifying values in the
topbar. For example, the brush radius can be used to control the depth
of the cut in the trimming tools or the size of the sphere in the
sphere mesh filter
Reviewed By: #user_interface, dbystedt, pablovazquez
Differential Revision: https://developer.blender.org/D9071
This forces the mask expand modal operator to use the maximum
iteration when the cursor is not over the mesh, masking the entire
connected component.
This fixes the issue for both expanding masks and face sets.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D9105
This adds support for pen tilt in sculpt mode. For now, pen tilt is used
by tweaking the tilt strength property, which controls how much the pen
angle affects the sculpt normal. This is available in Draw, Draw Sharp,
Flatten, Fill, Scrape and Clay Strips brushes, but it can be enabled in
more tools later.
The purpose of this patch is to have a usable implementation of pen tilt
in a painting mode, so users can test and see in which hardware and
platforms this feature is supported and how well it works. If it works
ok, more tools and features that rely on pen tilt can be implemented,
like brushes that blend between two deformations depending on the angle.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8893
The code to handle object transforms was wrong. Now the trimming mesh
and depts is calculated in world space, using the real view origin and
normal and then stored in object space in the mesh and in the original
coordinates array. As now both meshes for the boolean operation are in
the same object space, the space conversion code can also be removed
from the boolean function.
Reviewed By: sergey
Maniphest Tasks: T80625
Differential Revision: https://developer.blender.org/D8852
The voxel remesher was using the voxel size to limit the shrink-wrap
projection distance. Now that distance is increased to help preserving
more detail on hard surface edges.
Reviewed By: pablodp606
Differential Revision: https://developer.blender.org/D6204
Previously I used `CTX_copy` to create a mutable copy of the context in
order to set its area and region fields to temporary variables. This was
a tradeoff to avoid casting away `const` for `bContext`.
However, `bpy.context` is set to this new temporary value, which is fine
for a single `wm_draw_update` pass, but in the next main loop,
`bpy.context` is still set to this value, which was freed at the
end of `property_search_all_tabs`. It would be possible to reset the
`bpy.context` variable ath the end of the function, but this patch
contains an alternate solution: just don't copy the context. It looks
like this was the only use of `CTX_copy` anyway, maybe for good reason.
Differential Revision: https://developer.blender.org/D9216
With all the work on DNA defaults for 2.91, it's nice to expose this
convenient operator. This was already hardcoded in the UI code to the
backspace key, adding it to the keymap instead will make the shortcut
automatically show in the button right click menu.
Differential Revision: https://developer.blender.org/D9219
Split up `ANIM_deselect_anim_channels()` into
`ANIM_anim_channels_select_set()` and
`ANIM_anim_channels_select_toggle()`.
`ANIM_anim_channels_select_set()` is equivalent to the old
`ANIM_deselect_anim_channels(..., false, ACHANNEL_SETFLAG_xxx)`.
`ANIM_anim_channels_select_toggle()` is equivalent to the old
`ANIM_deselect_anim_channels(..., true, ACHANNEL_SETFLAG_ADD)`.
`ANIM_deselect_anim_channels(..., true, ACHANNEL_SETFLAG_CLEAR)` was
also called once. The `true` parameter suggested the "toggle" behaviour
was intended, but the `ACHANNEL_SETFLAG_CLEAR` argument prevented any
toggling. This is now replaced with `ANIM_anim_channels_select_set(ac,
ACHANNEL_SETFLAG_CLEAR)` to make this explicit.
No functional changes, just a cleanup in order to better understand how
the selection works.
Rename `animchannels_deselectall_exec` → `animchannels_selectall_exec` so
that it matches the operator `ANIM_OT_channels_select_all`.
No functional changes.
We need a separate time stamp for each axis, and we have to add a few
milliseconds of padding to those, to ensure wrapping on both axes get
properly performed when it happens almost simultaneously, and avoid
extra wrapping caused by very close events sometimes.
This only addresses the Linux (X11 backend) case.
Differential Revision: https://developer.blender.org/D9209
Being in render 'context'was not taken into account in code evaluating
modifiers for meshes in Edit mode.
Reviewed By: #modeling, mont29
Differential Revision: https://developer.blender.org/D9217
When creating a particle system to display simulated particles, the phystype needs to be set to 'no physics' so that particle positions are just copied and not integrated.
This shows the file's full path, its modification date, and its size in
the tooltips for the open recent fiels menu. When no file path is set,
the original operator description is used.
Differential Revision: https://developer.blender.org/D9028
This was caused by a wrong flag equality check when in xray mode because
the xray mode was masking the effect option flags that are not supported
in this mode. This means the never passed and the TAA was reset before
every redraw, leading to infinite rendering.
This was caused by the sequencer using a sRGB buffer without using
the sRGB transform.
This patch make it so that the framebuffer is rebound using the
sRGB transform before the python draw callbacks.
When there is no active texture, a NULL pointer was dereferenced.
It didn't crash because the dereference was for the first item at the
pointer, the ID. To fix this, return with no data when `texture is NULL.
A lot of animator request an option to invert the filter of the dopesheet channels. This patch adds that invert filter option. This is not for Grease Pencil only, affect to all modes.
{F8983328}
Note: I have seen the new button has a rounded borders on the left. It would be better get rectangle shape, but not sure how to do it.
Reviewed By: campbellbarton, pepeland
Maniphest Tasks: T81676
Differential Revision: https://developer.blender.org/D9182
c68a2a
This makes the icon equals to the invert icon used in all modifiers.
{F8986444}
Reviewed By: #user_interface, Severin, Blendify
Differential Revision: https://developer.blender.org/D9194
3bb3b2
This will make the "Reset to Default Value" operator in button right
click menus work for the fluid modifier. Before they always reset
the values to 0.
Differential Revision: https://developer.blender.org/D9206
Don't override other `LSAN_OPTIONS` like suppression file set in
the environment variable.
Old code added in {rB38ff5064b33ccb8} and {rB5f4e99b7a2b8376}
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D9207
Calling PyEval_ReleaseLock() was crashing with Python 3.9
because it accessed the NULL pointer set by PyThreadState_Swap().
This happened when calling ViewLayer.update() for example.
While the existing logic could be fixed by swapping the thread-state
back before calling PyEval_ReleaseLock(), this depends on functions
which are tagged to be removed by v4.0.
Replace use of deprecated functions by calling PyEval_SaveThread(),
instead of inlining the logic, using _PyThreadState_UncheckedGet()
to prevent Python aborting.
The call to PyEval_ThreadsInitialized has been removed
as threads are now initialized with Python.
This could be replaced with Py_IsInitialized() however it doesn't look
like this is necessary.
This is compatible with Python 3.7 & 3.9.
With constructive + deform modifiers, loop-cut visualization
wasn't following the displayed mesh.
This now gets the coordinates from the cage when available.
This was just not implemented. Now make this consistent with
vertexpainting / texturepainting.
Maniphest Tasks: T81670
Differential Revision: https://developer.blender.org/D9195
Previously the softbody strength property was controlling the strength
of the constraints that pin all vertices to the original location. This
was causing problems when the forces were trying to deform the vertices
too much, like when using gravity or grab brushes.
Now softbody is implemented with plasticity, which creates constraints to
a separate coordinates array. These coordinates are deformed with the
simulation, and the plasticity parameter controls how much the
simulation moves the coordinates (plasticity 0), or the coordinates move
the simulation back to its previous position (plasticity 1).
This creates much better and predictable results and adding softbody
plasticity to the brushes can increase its control and the stability of
the simulation.
Reviewed By: sergey, zeddb
Differential Revision: https://developer.blender.org/D9187
This patch enables property search for all tabs in the property editor.
To make interaction faster, if the editor's current tab doesn't have a
result, the current tab changes to the next tab that has a match.
This patch implements basic code that only searches panels.
While we could run the existing "single tab" property search for every
tab, that would also do everything else related to the layout pass,
which would be less efficient, and maybe more complicated to maintain.
The search match status for every current tab of the property editor is
stored in a runtime bitfield and them displayed later by dimming icons
in the tab selector panel to the left. Using `BLI_bitmap` properly in
the runtime struct required moving it to `buttons_intern.h` and
adding a small API to access the search filter instead.
To make sure the editor isn't influenced by anything that happens while
building the layout for other tabs, most of the context is duplicated
and the new search is run in the duplicated editor.
Note that the tool settings tab works slightly different than the other
tabs, so I've disabled searching it for this commit. That would be a
relatively simple improvement, but would just require a bit of
refactoring of existing code.
Differential Revision: https://developer.blender.org/D8859
This adds arguments to `uiLayout.prop_tabs_enum` and the C equivalent
(`uiItemTabsEnumR_prop`) to gray out tabs based on a boolean array.
For property search in multiple tabs, we need a way to show which tabs
have a search result, but we still need to show which tab is active.
Differential Revision: https://developer.blender.org/D8858
Changing the color of monochrome alert icons would not change until the theme was reloaded.
Differential Revision: https://developer.blender.org/D9062
Reviewed by Harley Acheson
This diff adds support for respecting the `BLENDER_USER_SCRIPTS`
environment variable when setting up the IDE environment.
Previously the scripts from the users profile folder were always
used even when this variable was set, leading to the possibility
of the wrong scripts being visible in the IDE.
Differential Revision: https://developer.blender.org/D9189
Reviewed By: Ray Molenkamp
With this patch the build system checks whether the "CUDA10_NVCC_EXECUTABLE" CMake
variable is set and if so will use that to build sm_30 kernels. Similarily for sm_8x kernels it
checks "CUDA11_NVCC_EXECUTABLE". All other kernels are built using the default CUDA
toolkit. This makes it possible to use either the CUDA 10 or CUDA 11 toolkit by default and
only selectively use the other for the kernels where its a hard requirement.
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D9179
Seems to be an oversight/typo in rB83980506957c.
Now set back to 0.01m (as it was before).
Maniphest Tasks: T81672
Differential Revision: https://developer.blender.org/D9193
A .blend file cannot use itself as a library.
This is prevented when linking manually, but was still possible when
relocating, which is now prevented.
Maniphest Tasks: T81656
Differential Revision: https://developer.blender.org/D9191
This was reported for sculpting, the same is true for weightpaint or
vertexpaint though.
When viewport clipping and radial symmetry are enabled, the
'sculpt_brush_test_clipping()' function was not considering radial
symmetry at all, so if the coordinate was outside the clipping planes,
no action would take place. Now the coordinte is brought back to where
the stroke actually happens and that is checked against clipping.
Since other mirroring options while painting/sculpting (as well as
editmode operations with mirroring) usually still take place even if the
mirrored coord is outside the clipping planes, this should also be the
case for radial symmetry.
This grows the 'SculptBrushTest' struct a bit, but should be acceptable?
Fixes T81466
Maniphest Tasks: T81466
Differential Revision: https://developer.blender.org/D9120
When we override a whole collection, we want to add non-instantiated
objects to a hidden sub-collection at the end of the process.
However, this makes no sense when instantiating an object, if other
dependencies objects get also overridden on the process, we should just
add them to the same collection owning the root object.
- BKE_bezt_subdivide_handles -> BKE_fcurve_bezt_subdivide_handles
- binarysearch_bezt_index -> BKE_fcurve_bezt_binarysearch_index
These functions are specific to F-Curves and don't make sense for other
uses of BezTriple (curve-object data for e.g.)
Also:
- Move detailed doxygen comment above code, following code-style.
- Mark bezt_add_to_cfra_elem unused.
The problem is related to the `use_index_order` option of `BLI_kdtree_3d_calc_duplicates_fast`.
With this option, the higher index is expected to be less than `tree->nodes_len`.
Blender crashed when dragging and dropping color into the outliner.
This issue was cause by a missing check for the correct drag type
in `datastack_drop_poll`. The check is added in this commit.
Additionally, a new drag type is introduced for the "data stack"
drag option, that was introduced in commit 1572da858d, to
differentiate it from the existing WM_DRAG_ID type.
Reviewed By: Severin
Differential Revision: https://developer.blender.org/D9169
Change to monochrome version of the large alert icons and use 'Question' for the the Quit Confirm dialog box.
Differential Revision: https://developer.blender.org/D9062
Reviewed by Pablo Vazquez
Default text output routines (which do not specify a size) will now use Text Style point size.
Differential Revision: https://developer.blender.org/D9107
Reviewed by Brecht Van Lommel
When changing the selected, active or visible object(s), the Outliner
has to be rebuilt while using the corresponding object state filters.
The object hiding operators also have to send the proper notifiers (they
changed visibility without notifying about that).
Check selection state in `BKE_fcurve_active_keyframe_index()`, and only
return the active keyframe index when that keyframe is actually selected.
This is now also asserted in the `BKE_fcurve_active_keyframe_set()` function,
which is now also used when inserting a keyframe.
While there are other bundles which show up as directories,
`.app` are the most common ones.
*Users should not be saving anything inside .app bundles, nor using
Blender to edit any of the files.
*This declutters the File Browser for say ~/Applications folder
or recursive search on a path with apps.
*Matches Finder's behavior of showing apps as files. (We don't have a
"right click > show package contents" button like Finder though)
This change shows `.app` files like incompatible files, or `.exe`s
on Windows.
{F8970986}
Reviewed By: #platform_macos, brecht, mont29
Differential Revision: https://developer.blender.org/D9162
That was a suboptimal decision from back in the days, which ended up
being problematic. It is no longer used, so remove it from API making
it so new code does not depend on this weak concept.
The order got broken when Brown distortion model has been added.
Made it so the indexing of parameters is strictly defined in the
parameter block, matching how parameters are used in the cost
function.
There is some duplication going on accessing parameters. This can
be refactored in the future, by either moving common parts packing
and cost function to an utility function in bundle.cc.
Alternatively, can introduce a public PackedIntrinsics class which
will contain a continuous block of parameters, and each of the
camera models will have API to be initialized from packed form and
to create this packed form.
The benefit of this approach over alternative solutions previously
made in the master branch or suggested in D9116 is that the specific
implementation of BA does not dictate the way how public classes need
to be organized. It is API which needs to define how implementation
goes, not the other way around.
Thanks Bastien and Ivan for the investigation!
This reverts commit 7e836bde11.
Reverting the incomplete workaround, due to the following reasoning:
- There should be no dependency between CameraIntrinsics and the bundler
code. This violates intended abstraction of the intrinsics class.
- The fix was not complete, or wrong. Even in the world where there is
a requirement to the parameters storage size this should have applied
to all distortion models, including Divisions, Nuke.
Proper fix will be committed next.
The leak was happening when problem did not have any parameters blocks
defined. This happens, for example, if there are no 3D points at all,
or when all markers are set to 0 weight.
Was noticeable in libmv_modal_solver_test when building with LSAN
enabled.
Fix regression introduced in rBe12767a0352a9e113892b4a07c6c8446d3ff361f
The volumes are not a line type and should not be render into the line
framebuffer nor it should change the framebuffer.
Before I was double confused about how openvdb stores its
transformation matrices. Now, I know they have the same layout
physically in memory, but are logically transposed (i.e. the translation
is in the last row instead of in the last column).
Previously Blender would only match the existing curve slope when the
to-be-inserted key value was already very close to the curve. This check
is now removed, allowing for sliders in the graph editor to subtly
change the curve, and for keyframes added with ctrl+click to follow the
curve better.
`CustomData_bmesh_interp` use the same CustomData decryptor (in this case, `bm->ldata`) in both blocks.
So make sure that all CustomData layers match.
This commit also removes redundant `BM_elem_attrs_copy_ex` calls.
Maniphest Tasks: T81060
Differential Revision: https://developer.blender.org/D9159
This resolves a long list of linker warnings that is currently only showing up on macOS arm builds.
The warnings themselves are of this shape (one example):
```
ld: warning: direct access in function 'Manta::MeshDataImpl<Manta::Vector3D<float> >::_W_39(_object, object, object*)' from file '../../lib/libextern_mantaflow.a(mesh.h.reg.cpp.o)' to global weak symbol 'typeinfo for Manta::MeshDataImpl<Manta::Vector3D<float> >' from file '../../lib/libextern_mantaflow.a(mesh.cpp.o)' means the weak symbol cannot be overridden at runtime. This was likely caused by different translation units being compiled with different visibility settings.
```
Just would like to get your opinion to make sure this is an acceptable way to handle this on all platforms.
Reviewed By: LazyDodo
Differential Revision: https://developer.blender.org/D9002
The strength has a high impact on performance. With the previous
step and default one could easily and accidentally make Blender unusably
slow, because of a too high displacement strength.
Clean up the code that handles mouse clicks on animation channels in the
non-NLA animation editors, by splitting into separate functions.
No functional changes.
Using the diagonal has the benefit, that the adaptive voxel size changes
when the bounding box changes its shape in any way. Having a changing
voxel size looks bad when rendering an animation, therefore one should
usually use a fixed voxel size when rendering an animated volume.
This becomes apparent earlier, when the adaptive voxel size changes
when the bounding box changes in any way.
Otherwise this can easily go unnoticed when rendering only a few frames
of an animated volume.
For multi-config generators, tests are in `bin/tests/<config>` folder.
and OpenMP cannot be found at `@executable_path/../Resources/lib`.
So create that folder and put the library there to be used by tests.
It is not ideal to copy the library around. When minimum CMake version
is changed to 3.12, FindOpenMP by CMake can be used for this whole
block of code. Then a nice rpath based solution can be used.
The "Apple" part is guaranteed by the platform file.
"Clang" is enough to differentiate it from other compilers.
Came across this due to the custom built LLVM toolchain I'm using.
The "Geometry start & End" panel was grayed when using an object
to create the curve bevel. It should only be grayed out in this case
when no object is chosen in the "Bevel" subpanel.
The code that restored collection flags after they are rebuilt when
moving a collection didn't take into account collection children. The
flag for the active collection was properly restored, but all of its
children would take on the exclude flag of the collection the active
collection was dragged into.
This commit builds a temporary tree structure to store the flags for
the moving collection and its children. Then it reapplies these flags
after `BKE_main_collection_sync`.
Differential Revision: https://developer.blender.org/D9158
`targetIsDirectory` slipped through the code review of
{D6679}/{rBafb1a64ccb81}.
`BLI_is_dir` exists to check for directory status of a file.
Remove some `else-after-return`s.
Use `r_` prefix for return value arguments, and move it to the end
in the list of arguments.
The routine to find dissolvable vertices had a check to ensure
that the vertex was exactly in line with the two neighbors.
I have convinced myself that this check is unneccesary (it was
failing with only a 1e-9 difference from 0 on a cross check),
so have removed it.
After tests were bundled in a single executable and cycles and libmv
created their own tests, the warnings on macOS have gone over 800.
The reason is setting `*_LIBRARIES` to names of the libraries
and later using `link_directories` to link them properly.
https://cmake.org/cmake/help/latest/command/link_directories.html
> Note This command is rarely necessary and should be avoided where
> there are other choices. Prefer to pass full absolute paths to
> libraries where possible, since this ensures the correct library
> will always be linked. The find_library() command provides the
> full path, which can generally be used directly in calls to
> target_link_libraries().
Warnings like the following popup for every target/executable,
for every library it links to.
```
ld: warning: directory not found for option
'-L/Users/me/blender-build/blender/../lib/darwin/jpeg/lib/Debug'
```
The patch completes a step towards removing `link_directories` as
mentioned in TODO at several places.
The patch uses absolute paths to link libraries and removes
all `*_LIBPATH`s except `PYTHON_LIBPATH` from
`platform_apple.cmake` file. (The corner case where it's used seems
like dead code. Python is no longer shipped with that file structure.)
Also, unused code for LLVM-3.4 has been removed.
Also, guards to avoid searching libraries in system directories have
been added.
`APPLE` platform now no longer needs `setup_libdirs`,
`cycles_link_directories`, and `link_directories`.
The number of warnings now is less than 100, most of them being
deprecation ones in dependencies.
This patch depended on {rBb746179d0add}, {rB2fdbe4d05011},
{rB402a4cadba49} and {rBd7f482f88ecb}.
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D8855
The real maximum was `GPU_SAMPLER_ICON`, not `GPU_SAMPLER_REPEAT`, my
bad. {rBa31a87f8943aa40}
Move `GPU_SAMPLER_MAX` out of the enum since it's used as an `int`
at many places.
Also, the macro `ENUM_OPERATORS` needs a maximum, and this enumerator
cannot be used as the argument of that macro. It creates wrong values
in the `~` NOT operator.
Thanks @deadpin for catching this.
Reviewed By: fclem
Differential Revision: https://developer.blender.org/D9157
The 0.1 limit was too large. The issue with making it smaller is that
one can easily crash blender by making it to small. To counteract this,
the step has been reduced as well.
A voxel size/amount of 0 disables the modifier.
This is a follow up commit for rB309c919ee9.
Clearing hash tables is now parallelized as well. Surprisingly, most of
the time is actually spent in `free` (a couple of milliseconds per call
in my test).
Benchmark of individual functions:
reserve_hash_maps: 17%
add_polygon_edges_to_hash_maps: 49%
serialize_and_initialize_deduplicated_edges: 12%
update_edge_indices_in_poly_loops: 14%
clear_hash_tables: 5%
`BKE_mesh_calc_edges` was the main performance bottleneck in D9141.
While openvdb only needed ~115ms, calculating the edges afterwards
took ~960ms. Now with some parallelization this is reduced to ~210ms.
Parallelizing `BKE_mesh_calc_edges` is not entirely trivial, because it
has to perform deduplication and some other things that have to happen
in a certain order. Even though the multithreading improves performance
with more threads, there are diminishing returns when too many threads
are used in this function.
The speedup is mainly achieved by having multiple hash tables that are
filled in parallel. The distribution of the edges to hash tables is based on
a hash (that is different from the hash used in the actual hash tables).
I moved the function to C++, because that made it easier for me to
optimize it. Furthermore, I added `BLI_task.hh` which contains some
light tbb wrappers for parallelization.
Reviewers: campbellbarton
Differential Revision: https://developer.blender.org/D9151
While not full width this gives a nice trade off between improved screen
use space and readibility.
The old value was 80 so this gives 25% more screen space.
See T73223
Was failing because of a mismatch between RNA and DNA defaults on
`courant_target` property, removed RNA one (DNA came from initial value
defined in BKE's particle creation code, think it's best to keep that
one).
Due to the old AA method, it was necessary to add transparency to the
outline of the popover arrow to match the box outline opacity. This is
no longer required after rB15dda0115c78.
See the differential for before and after screenshots.
Differential Revision: https://developer.blender.org/D9026
The underlying type of the enum cannot be fixed here due to its usage
in C code.
All the values possible in the width of the underlying type are not
valid for an enum.
Only 0 to (2*max - 1) if all enumerators are unsigned.
So the macro asks for the biggest value among the //listed// ones.
If any enumerator C is set to say `A|B`, then C would be the maximum.
(2*max-1) is used as the mask.
The warnings (for each enum modified in this commit):
GPU_vertex_buffer.h:43:1: runtime error: load of value 4294967291
which is not a valid value for type 'GPUVertBufStatus'
https://github.com/llvm/llvm-project/commit/1c2c9867
Ref T81340
Reviewed By: fclem
Differential Revision: https://developer.blender.org/D9067
- Move some security checks outside of `interp` callbacks.
Namely, that we do get interpolation weights, and have something to
interpolate.
Some callbacks where not checking on those anyway, safer to move that
up into calling code.
- Cleanup usage of sub-weights, lots of interpolation callbacks wher
actually using those completely wrong.
- Change default behavior when no weights are given to higher-level API
functions: prevriously, each callback was responsible to handle that
case (and one did not even do it!), they were switching to purely
additive behavior then.
Instead, we now default to expected simple average of source values.
Note that the only real important change here is defaulting to actual
average of source value when no inertpolation weights are given (afaik,
this only happens in Weld modifier code).
Differential Revision: https://developer.blender.org/D9114
This patch "modernizes" `DNA_struct_switch_endian` similar to
how I updated `DNA_struct_reconstruct` recently. Furthermore,
some special case handling have been moved to another place.
Reviewers: campbellbarton
Differential Revision: https://developer.blender.org/D9089
This patch adds missing icons for the new trim tools in sculpt mode.
Although these tools recently got "add geometry" modes, it's more
essential to highlight the most important feature of the tool,
trimming, than to try to portray everything they can do.
Differential Revision: https://developer.blender.org/D8963
This patch adds missing icons for the new sculpt face set tools, and
updates the mask icons to be more consistent with the changes.
Currently draw face sets and draw mask icons are inconsistent and don't
relate to drawing/painting masks or face sets. This commit makes the
icons consistent and reusable for future tools like sculpt vertex
colors.
Differential Revision: https://developer.blender.org/D8875
This directly adress the issues caused by rB536c2e0ec916.
Since the state tracking is done at a lower level, using the bgl
functions needs to be safegarded by the state manager.
The current workaround is to bypass `apply_state` when inside a
callback that used a `bgl` function.
Related to T80730.
This fix T81003.
Also this fix the default blend equation for callbacks.
Fixes T80169 T81289.
`APPLE` platform handles ASan compiler and linker flags using
`add_compile_options` and `add_link_options`. {rB74bcb32c9f02}
Arguments in `CMAKE_{LANG}_FLAGS{_CONFIG}` are also passed to
`try_compile` which will fail due to linker errors, since link flags
are not set. `try_compile` is used by `find_package(Boost)` for
`thread` library.
See CMP0066 [1] also.
[1] https://cmake.org/cmake/help/latest/policy/CMP0066.html
Ref D8855
Both the default theme and Blender Light have been updated.
Blender Light also includes automatically generated colors based on the
default theme that were not set previously.
Better use higher-level code from common ID management when possible.
Helps to de-duplicate logic, and reduces outside usages of more
'dangerous' functions.
Note that we could get rid of many of those `BKE_<id_type>_add`
functions now, but on the other hand several of those take extra
parameters and perform additional actions, so think we can keep them all
for now as 'non-standard ID specific creation functions'.
Generic ID management code can now do those local temp copy handling, so
no need for duplicated own code for that.
No behavioral changes expected here.
Python 3.7x added support for module `__getattr__` & `__dir__`.
Make use of this for operator access,
previously these were module like classes.
No functional changes, this is mostly to use a type which is expected,
rather than faking the type with a class instance.
After rBeb911610d05 there are some differences in how the layout code
decides whether to add decorators (aka animation buttons). Whether or
not that logic is correct, it's nice to specify that we don't want them
turned on in this popover.
Those were only shallow wrappers around `BKE_id_copy`, barely used (even
fully unused in some cases), and we want to get rid of those ID-specific
helpers for the common ID management tasks. Also prevents weird custom
behaviors (like `BKE_object_copy`, who was the only basic ID copy
function to reset user count of the new copy to zero).
Part of 71219.
Note that possibility to pass the new ID pointer as parameter was kept,
as this is needed for some rather specific cases (like in depsgraph/COW,
when copying into already allocated memory).
Part of T71219.
This modifier uses a 3D texture to displace a volume.
For now, this can only use the previously existing texture system,
because we do not have a better alternative yet. Still, the results
can be quite good and interesting. See D9075 for some examples.
Reviewers: brecht, simonthommes
Differential Revision: https://developer.blender.org/D9075
Changes Workspace tabs so that the text size and tab width are both set by style.widget.points.
Differential Revision: https://developer.blender.org/D9108
Reviewed by Julian Eisel
Move position of Alerts and Progress bar back to the center of the Statusbar.
Differential Revision: https://developer.blender.org/D9118
Reviewed by Brecht Van Lommel
Regular rendering uses a custom blend mode, but render passes renders to
2 separate textures. This wasn't configured correctly inside the
fragment shaders. This patch adds a switch to configure the fragment
shader with the correct attachments.
Backport to Blender 2.83.
Reviewed By: Clément Foucault
Differential Revision: https://developer.blender.org/D9038
The overlay options in the image/uv editor is hidden in side panels and menus. Sometimes this panel is even hidden, while still useful.
The goal of this task is to introduce an overlay pop-over just like the overlay-popover of the 3d viewport.
Popover has
* UV Stretching (only available in the UV mode, when active object mode is a mesh and in OB_EDIT mode)
* Display As (only available in the UV mode, when active object mode is a mesh and in OB_EDIT mode)
* Show Modified (only available in the UV mode, when active object mode is a mesh and in OB_EDIT mode)
* Show UV Edges (including opacity slider; available UV, View, Paint, when active object mode is a mesh and in OB_EDIT mode)
* Udim tiles when no image is available.
Like the 3d viewport, there will be a editor toggle to enable/disable the overlays
For compatibility reasons the RNA properties are added to both the `SpaceImage.uv_editor` amd `SpaceImage.overlay`. On DNA level they are still stored in the SpaceImage. only new properties are added to the SpaceImageOverlay struct. During the next major release we could remove these options from `SpaceImage.uv_editor`. This should be noted in the Python section of release notes.
Reviewed By: Julian Eisel, Pablo Vazquez
Differential Revision: https://developer.blender.org/D8890
Particularly useful now that we can dynamically generate volume using
modifiers. Not exposed in the user interface currently, but it can be used
by add-ons that need to export an entire scene including volumes.
There was a wrong duplication of all strokes and this accumulated all previous data for each frame.
Also a cleanup has been done in order to clarify what means each variable.
The alloc was done in the init, but the free was done in the discard.
As discarding can happen more often there were some unneeded if
statements to work around use after free.
This patch moves the free to the `mesh_batch_cache_clear` and removes
the if statements to check for use after free.
The panel category tabs (In the 3D view N-panel) did not look great
when zoomed in or with high DPI screens, with multiple overlapping
outlines, and roundness that didn't match elsewhere in Blender. In
addition, there was some unecessarily low-level drawing code in the
panel code.
This commit uses an existing function `UI_draw_roundbox_4fv`to draw
the tabs instead. There are some slight visual differences, though
these are easily adjusted with theme colors.
Differential Revision: https://developer.blender.org/D9045
In the graph editor there is a panel that says "Active Keyframe" for
numerically editing a keyframe's values, but in the code there is no
concept of the "active keyframe." Since this is a useful concept to
have for some other features anyway, this commit adds an active
keyframe index value to FCurves. It also displays it with a theme
color for the active vertex (which didn't exist before) if the
FCurve is active.
The active keyframe in the graph editor is treated similarly to the
active vertex in the 3D view. It is the keyframe most recently selected
with a single click, and it is always selected.
For now, the only real functional change is that the active keyframe
appears in white and it should be more predictable which keyframe is
being edited in the sidebar panel.
Differential Revision: https://developer.blender.org/D7737
Move the information about the type of the source objects to the item
descriptions instead of the names. The extra information in the names
made the labels cluttered, and it's less important information.
Differential Revision: https://developer.blender.org/D8352
Follow the MLA style, agreed upon in T79589. This means "from" within UI
labels should be lowercase.
Differential Revision: https://developer.blender.org/D8345
The settings in these popups are not animateable, so also turn off
property decorate. And also use better widths in order to fit the
full text.
Differential Revision: https://developer.blender.org/D9060
Aligning the columns removes unecessary white space.
The same as in the D8480 / rBb2e0c8f902fa.
Differential Revision: https://developer.blender.org/D9111
This path makes use of more framebuffers when using the gpu workarounds.
Increase the MAX FBO attachments per texture. Maybe we could use a growing
vector in the future.
At the end of `GPU_framebuffer_blit` when `prev_fb->bind(true);` is called,
the `context_->active_fb` was not in sync and lead to the wrong framebuffer
being modified by bind function.
This fix T81055 SSS light is missing from the diffuse light render pass
Endian switching when loading 64bit blend files on a 64bit system was
crashing as the endian switching is only applicable when loading
on 32 bit systems.
This crash goes back to 2.7x, it looks like this never worked
all the way back to the first commit.
Avoid overwriting user-chosen `SACTION_SLIDERS` flag when switching the
Dope Sheet editor to Shape Key mode. The Shape Key mode now ignores the
flag, and always shows the sliders.
The obvious limitation is that you can't hide the sliders in the Shape
Key editor anymore
Reviewed By: looch, sybren
Differential Revision: https://developer.blender.org/D9121
Besides the NodeTree case (which remains unchanged), the localize code
is only used in one place (to generate previews of shading data-blocks).
This commit introduces a new `LIB_ID_CREATE_LOCAL` option for ID
creation/copying, which essentially implements the behavior of the
removed `BKE_XXX_localize()` functions into regular mainstream ID copy
code. When this option is set:
- new ID is tagged with `LIB_TAG_LOCALIZED`;
- Some ID copying callbacks have specific behaviors, mainly the root
nodetree of shading IDs gets duplicated with specialized
`ntreeLocalize()` function.
Note that I would not consider getting rid of `ntreeLocalize` for now,
this function is recursive, which should ideally never happen within ID
management copying code (this introduces all kind of complications).
No behavioral change expected from this commit.
Add a specific flag for nodetree deep-copy special localization code.
And add a new `LIB_ID_CREATE_LOCALIZE` flag, similar to
`LIB_ID_COPY_LOCALIZE`, for creation purposes.
No behavioral changes expected here.
colors
If the mesh was constant, no check was done if there were animated
vertex colors and thus creation of a MeshSequenceCache modifier was
skipped.
Thx @sybren for feedback!
Maniphest Tasks: T81330
Differential Revision: https://developer.blender.org/D9057
If this doesnt work in a particular mode, it would be better to indicate
that in the Overlay popup.
Differential Revision: https://developer.blender.org/D8981
In that case, weightpainting would still be done threaded, but current
mirroring code cannot be run in parallel.
Caused by rB5502517c3c12 [which was checking mesh 'flag', needs to be
'editflag' -- thus test was always true].
Reviewers: pablodp606
Maniphest Tasks: T81461
Differential Revision: https://developer.blender.org/D9126
For other areas in Blender that have a toggle and related settings,
we put a popover with the settings right next to the toggle. This
combination is nice because it organizes the settings without making
interaction slower. It also makes the settings more discoverable since
they're right next to the toggle.
Differential Revision: https://developer.blender.org/D8537
The edit_constraint_property_get function was using bone constraints
in pose mode even for object constraints. This change mirrors the code
in `uiTemplateConstraints` as well.
Differential Revision: https://developer.blender.org/D9128
The flood-fill helper serves as a sanity check for scenes with complex obstacles. Similarly to the holes filler function it ensures that there are no holes in obstacles.
In addition, it ensures that there will be no holes resulting from multiple intersecting obstacles.
In general, this commit should improve stability.
This enables a union boolean mode for the trimming gestures tools which
adds geometry to the mesh instead of cutting it.
It also adds a Join mode, which adds the geometry directly without
using a boolean operation.
Depending if you plan to use dyntopo or not, it is useful to have both
options available.
This is using the full depth of the object from the camera view for the
depth of the geometry, but options for controlling the trimming depth in
all modes are going to be added in later patches
Reviewed By: mont29
Differential Revision: https://developer.blender.org/D9066
The example file in T81218 has a driver that maps a bone's X-location to
the number of BBone segments. This caused a dependency cycle, which
resulted in bad thread serialisation, which caused the crash.
This patch breaks the dependency cycle `BONE_LOCAL` →
`DRIVER(bones["Bone"].bbone_segments)` → `BONE_LOCAL`. The 'Driver Data'
relation now points to `BONE_SEGMENTS` when the driven property starts
with `bbone_`.
Differential Revision: https://developer.blender.org/D9122
When sequencer was moved from kernel, the DEFINES based on audaspace
were not moved along.
That means users had no way of creating new audio strips since
BKE_sequencer_add_sound_strip was ifdef'out
The buttons hide when the search finishes based on whether they are in
the "panel header" group. These buttons were not protected with a new
group. This adds a new group for operator button calls, and also makes
it so a new group is always created after the header buttons.
- Density, interpolation and slicing options in Volume Object's
Viewport Display are now not aligned to each other as they are not
closely related.
- Changed the enum property for slicing ('Method') to a boolean
property 'Slice' and added a sub-panel for slicing options under
the 'Viewport Display' panel in Volume Object and Fluid for better
clarity.
- Renamed `axis_slice_method` to `use_slice` in the python API for
Volume Object and Fluid.
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D9097
When extracting a mask, the new object is going to have the same mask on
the same vertices, so it can't be sculpted without clearing the mask
first. It makes more sense to delete the mask datalayer from the new
object.
Reviewed By: mont29
Differential Revision: https://developer.blender.org/D9070
This helps blending artifacts with dynamic simulation areas as the
damping increases when the vertex.
Reviewed By: mont29
Differential Revision: https://developer.blender.org/D9084
This adds a small gradient to the right side of the line to preview
which side of the mesh is going to be affected by the gesture operation.
Reviewed By: Severin
Differential Revision: https://developer.blender.org/D9106
rendering range
The Allow Negative Frames option, introduced in rB21a2350248fd, allows
for negative frames, but this should only apply for playback and
animations, the rendering range should still be clamped to MINFRAME /
MAXFRAME, because rendering does not allow for negative frames. Preview
range should not be affected afaict (I am not aware of operators that
allow for rendering this range).
Maniphest Tasks: T81380
Differential Revision: https://developer.blender.org/D9112
Move the collection exclude checkbox to be with the other restrict
buttons. This makes the button consistent with the other restrict
buttons, and it makes the visual hierarchy more clear with the contents
of a collection properly indented under the collection name, rather than
the collection icon.
Differential Revision: https://developer.blender.org/D8928
When the limit surface is disabled OpenSubdiv generates a set of linear
patches which are only C0 continuous, not C1. This makes it impossible to
evaluate derivatives at vertices which, in this mode, are by definition put
at boundaries of patches. Normals are calculated from those derivatives.
Solution is to disable normal calculation and let it be done downstream, as
for other modifiers. This limitation is also the reason that non feature
adaptive subdivision is badly suited for GPU evaluation.
Differential Revision: https://developer.blender.org/D9103
Because the active context is stored as a short in DNA it's better
to use short in the code surrounding it. Also adjusted a comment
that reflected an incorrect assumption.
During the refactor, this feature was removed by error.
When this option is enabled, only the frames with a keyframe are displayed in Draw mode. This is used when fill an animation to paint only in the frames with real drawing.
Fix T81429.
This was an intentional change in rBc75a665c442e as it maintains the
same behavior as the constraint with or without modifier.
But from the user's PoV, it is better to keep the old behavior.
This makes drawing and behavior more intuitive.
NanoVDB is a platform-independent sparse volume data structure that makes it possible to
use OpenVDB volumes on the GPU. This patch uses it for volume rendering in Cycles,
replacing the previous usage of dense 3D textures.
Since it has a big impact on memory usage and performance and changes the OpenVDB
branch used for the rest of Blender as well, this is not enabled by default yet, which will
happen only after 2.82 was branched off. To enable it, build both dependencies and Blender
itself with the "WITH_NANOVDB" CMake option.
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D8794
This has serious performance implications as the additional check makes
it O(N^2) when there are a lot of merged vertices.
The goal is to move the merging to the weld modifier where it would
be much easier to optimize.
Ref D8946
This commit to complex solidify swaps the side of the geometry which
is affected by the material offset when the normal flip option is
enabled. This matches simple solidify and is considered correct/useful.
Ref D8948
Apply the De Casteljau algorithm to split the Bèzier curve at the X
coordinate where the new key is inserted, and uses the result to update
both the newly inserted and surrounding handles.
For curves that use Auto keyframes this has been largely addressed by
the new algorithm from D2884. This commit extends this to non-auto
handles.
This code is heavily based on D3172 by Alexander Gavrilov (@angavrilov).
Manifest Task: https://developer.blender.org/T81353
Initialize CLOG in the blendfile-loading unit test superclass. Since
rB8683d4e88f2e CLOG is used by more areas in Blender, and without
initialisation it crashes.
Current libmv_modal_solver_test fails since rB3a7d62cd1f5e.
It appears that the issue is caused by the insertion of the new OFFSET_K4 parameter,
as, if camera intrinsics are not required to implement/use all of those deform
parameters, they absolutely have to keep order (values) matching those defined
in bundle.cc, otherwise `PackIntrinisicsIntoArray` and `UnpackIntrinsicsFromArray`
will mangle them around.
This reverts commit 66cd82a8d. This was based on a misunderstanding. We do want
to take into account transforms similar to other modifiers like booleans, to
support taking into account animated transforms, combining multiple meshes into
one volume, etc.
Declaring variables where they are initialized and used makes their
scope much more explicit and it aids readility since types are visible.
This commit doesn't touch the ID template code since there is an
active patch in that area.
This is first step of refactoring task T77580.
Next step will be breaking up files into smaller ones.
Reviewed By: sergey, brecht
Differential Revision: https://developer.blender.org/D8492
Remove word proxy from eSpaceSeq_Proxy_RenderSize members if proxy is
not used with that item. Otherwise this can be a bit confusing.
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D8861
If current frame is set to before start frame, with AV sync there is
delay when playback is started.
Tag scene with ID_RECALC_AUDIO_SEEK, so audio playback starts from
correct position.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8809
Functionality was hard to follow as the meaning of fullname
depended on basename being NULL or not.
- where_is_temp is now only responsible for finding the temp directory.
- Split out tempdir_session_create into it's own function
that wraps mkdtemp.
- Rename basename to tempdir, fullname to tempdir_session.
This choice was made when preferences were not read in background mode,
so the temp dir needed to be initialized with something.
See bc9848f7e6.
Now preferences are always read there is no reason to ignore this
particular preference in background mode.
Anyone wanting to do background renders that ignore their preferences
can run with factory startup.
These were printf's, disabled by a define that could be uncommented.
Use logging since this can be useful when investigating any issue
with paths not being found at the expected location.
Without this any problem finding app-templates, Python's installation,
icons ... etc are quite difficult to troubleshoot especially on systems
without access to system-call tracing.
To use this run:
blender --log "bke.appdir" --log-level 3
Initialize ImBuf (and color-management) after passing arguments
that set environment variables such as `--env-system-datapath`
This also fixes a bug where BKE_appdir logging failed since it was
called before the `--log` argument was passed.
Add asserts so this doesn't happen again.
Internally appdir functions would test if a path exists,
returning false if it doesn't, leaving the string set instead
of clearing it.
This is error prone as invalid paths could be used accidentally.
Since BKE_appdir_folder_id_user_notest & BKE_appdir_folder_id_version
depend on this, add a 'check_is_dir' argument so the path can be used
even when the directory can't be found.
Actually, fixes compile error and an incorrect poll.
Caused by 21fc4ae206 and 3a1cf838ca.
The poll has to perfrom the `view_zoom_poll()` logic to ensure correct context.
Uncovered by a4aa94c41c. Before that, it would actually be a possible `NULL`
pointer dereference.
It would happen if a `UILayout.prop()` was placed with the `full_event` option
set to `True` (or the C equivalent) for a non keymap-item property.
Now this the `full_event` option is ignored then with a warning print.
Multiple related issues:
* `view_zoomdrag_exec()` required `view_zoomdrag_init()` to be called first to
set up data. Can now easily be done in the `_exec()` function as well though.
* Because of that the operators could not be called in an `EXEC_` context (e.g.
`EXEC_REGION_WIN`).
* With "Zoom to Mouse Position" enabled, zooming from a menu or script would
zoom to the mouse position. That should only happen if the operators are
called directly in the editor.
`view_zoomdrag_init()` would perform a number of sanity checks that
`view_zoom_poll()` already executed. The design of operators forsees that
(non-expensive) context sanity checks are done by the `poll()` callbacks, and
the execution callbacks can then just assume that happened. No reason to be
overly pedantic, we don't do that elsewhere either.
Note that this code is from the earliest days of an operator design. So it's
not surprising that it wasn't using the design "properly".
This commit makes the panel expansion set based on the search results
when the active tab in the properties editor changes. The multi-tab
search patch (D8859) actually doesn't handle this because it uses a
different code path.
This feature uncovered a subtle but fairly significant issue with the
implementation of property search (More details in T81113). Basically,
the search needed multiple redraws to properly display the expansion of
panels based on the search results. Because there is no animation of
panel expansion when switching tabs, the problem was exposed only now.
With this commit, hiding of "search only" buttons and panel size
calculation happens in a single final step of the panel layout pass.
The "search only" layout root flag is removed. Instead every button
inside a panel header is in a single "uiButtonGroup" marked with a
specific "in header" flag, an idea which could be generalized in the
future.
Differential Revision: https://developer.blender.org/D9006
Change `test_path` function so:
- Joining paths calls a function instead of being performed inline.
- Optional NULL paths must always be ordered last (easier to follow).
Other minor changes:
- Remove FIXME comment is it's handled properly by BLI_path_join.
- The wrong size was being used copying into `targetpath`.
We have our own assert implementation, `BLI_assert()` that is prefered over the
C standard library one. Its output is more consistent across compilers and
makes termination on assert failure optional (through `WITH_ASSERT_ABORT`).
In many places we'd include the C library header without ever accessing it.
Revert 76b1a27f96 since there is no
reason windows should behave differently to other platforms.
This was added so Windows users wouldn't see "/tmp/" in the UI.
Since then the default temporary directory is a blank string,
leave blank on all systems as Python script authors may accidentally
use this instead of `bpy.app.tempdir`.
This last worked in v2.27 (2003) where all paths were initialized to "/"
which was still checked to initialize the temp directory.
This hasn't been the case since 932e9e8316
where it changed to "/tmp/", then an empty string (current default).
This makes it unecessary to create a button group when the block
is created, giving more flexibility when creating the first group-- for
example, creating the first button group with special parameters.
For a future patch (D9006) we need these groups for longer than just the
the layout process, in order to differentiate buttons in panel headers.
It may also be helpful in the future to have a way to access related
buttons added in the same uiLayout.prop call. With this commit, the
groups are stored in and destructed with the uiBlock.
When blend files were loaded with app-templates,
setup_app_data was running before defaults were updated.
This is likely to cause problems with order of initialization
so always update the startup file beforehand.
BLO_version_defaults_userpref_blend -> blo_do_versions_userdef
The name was misleading as it was declared along with
BLO_update_defaults_startup_blend making it seem these functions were
related.
In fact preference defaults don't need to be updated as is done for
startup.blend since an in-memory blend file isn't used.
Rename the function to match other versioning functions
called from readfile.c. Also add/update comments on these differences.
Now versioning UserDef is run in readfile.c,
as is done for other Blender data.
Previously versioning was mixed with other run-time initialization,
so it needed to be called later by the window manager.
readfile.c's versioning function was only used for 2 variables.
Move versioning into versioning_userdef.c so everything
is done in one function.
Note: DNA_struct_elem_find checks have been replaced with checks for
the next released version.
This is harmless, as only old preferences saved between releases can
have their values overwritten.
Note: userdef versioning should be called from `do_versions_userdef`,
this will be done separately.
This replaces the blue theme color for subpanel titles with the the same
fade as for parent panels.
The search color doesn't work well for subpanel title colors. And actually,
because there are often buttons with checkboxes in the panel headers, we
don't have to treat this indicator any differently than regular buttons.
Differential Revision: https://developer.blender.org/D8976
Currently there's no way to know if a button is inactive when it doesn't
match the search results, because they use the same 50% gray level.
This isn't a huge problem, but it could lead to confusion. This commit
uses a subtle solution, a 25% opacity when the button is inactive and
also filtered by search.
This requires flipping the meaning of the UI_SEARCH_FILTER_MATCHES
flag in the code, and also adding a widget_alpha_factor utility in
the widget code.
Differential Revision: https://developer.blender.org/D8975
3D Viewport -> Sidebar -> View -> Local Camera is disabled. When you
click on it, box select starts. This behavior isn't intended, and fixing
it is quite simple. The only case where this would continuing would
still be desired is if a disabled button overlapped with a non-disabled
button, but this shouldn't be the case anywhere.
Differential Revision: https://developer.blender.org/D8832
The current way of setting the compute device makes sense for local
use, but for headless rendering it it a massive pain to get Cycles
to use the correct device, usually involving entire Python scripts.
Therefore, this patch adds a simple command-line option to Blender
for specifying the type of device that should be used. If the option
is present, the settings in the user preferences and the scene are
ignored, and instead all devices matching the specified type are used.
Differential Revision: https://developer.blender.org/D9086
Define enum `eContextResult` and use its values for returns, instead of
just returning 1, 0, or -1 (and always having some comment that explains
what -1 means).
This also cleans up the mixup between returning `0` and `false`, and `1`
and `true`. An inconsistency was discovered during this cleanup, and
marked with `TODO(sybren)`. It's not fixed here, as it would consititute
a functional change.
The enum isn't used everywhere, as enums in C and C++ can have different
storage sizes. To prevent issues, callback functions are still declared
as returning`int`. To at least make things easier to understand for
humans, I marked those with `int /*eContextResult*/`.
This is a followup of D9090, and is intended to unify how context
callbacks return values. This will make it easier to extend the approach
in D9090 to those functions.
No functional changes.
Differential Revision: https://developer.blender.org/D9095
Specific armature editing undo code would duplicate and store a list of
editbones. However, those are not linked to main data storage and should
therefore never affect ID usercounting.
Armature freeing would not correctly free its editbone IDProperties.
Add a utils to free the whole list of edit bones, and properly handle
their potential IDProperties.
Freeing of bones' IDproerties from Armature `free_data` callback would always
attempt to do user refcounting, which should never be done from that code.
This would generate crashes in depsgraph/COW context e.g.
So far data management code would simply fully ignore potential ID
pointers in custom properties of edit bones (which are a copy of those
from regular `Bone`). This would lead to all kind of issues, among which
refcounting inconsistencies, missing clearing of data uppon deletion,
etc.
Picking is disabled
Affected were (at least)
- Select Linked pick
- Select Shortest Path
Was originally caused by rB8b347fc2cdc6 [which got improved in
rBfc8a7a44b200 - but only partially].
After removal of BONESEL_NOSEL we have to filter out -1 from the hits for
any bone picking it seems (since armature drawing uses this).
In rBfc8a7a44b200, this was done for 'ED_armature_edit_select_pick' (or
'get_nearest_editbonepoint' precisely), now do this for
'ed_armature_pick_bone_impl' as well.
Reviewers: campbellbarton
Maniphest Tasks: T81200
Differential Revision: https://developer.blender.org/D9080
The sticky_select_mode property is an enum where each item has a
different behavior, but currently the description is for a single one
of the items. It should be more general, especially because one of the
items is "Disabled."
A couple more problems:
- "Automatically" is not so helpful. Many things any program does are
"automatic", it doesn't really help to say it.
- "Also" is bad grammar.
- Overly complex wording
This also adjusts the SHARED_VERTEX tooltip, because "irrespective"
should have "of" after, but it's also unecessarily complex.
Differential Revision: https://developer.blender.org/D9049
Refactor `ed_screen_context()` to use `GHash` lookups instead of a
sequence of string comparisons. This should provide a nice speedup,
given that the hash for `member` only has to be computed once instead of
matching it to each possible string.
Reviwed by: brecht
Differential Revision: https://developer.blender.org/D9090
Refactor `ed_screen_context()` to call separate functions, instead of
having the entire functionality in one function. Each function now only
retrieves the data it needs from the context. Furthermore, some string
comparisons are removed.
No functional changes.
Reviwed by: brecht
Differential Revision: https://developer.blender.org/D9090
There was an oversight when adding new experimental user preferences.
I can try to overengineer this later to make it more fail-proof. But for now
it should be clear what to update when adding a new variable.
Version patching userpref.blend wasn't using the correct version,
causing settings not to be properly updated.
This seems the likely cause of T70196 and similar bugs.
The license was renamed GPL-license.txt => license/GPL-2.0txt in 0434efa09d.
Although the renaming made all the license files consistent, we are better off
not updating the copyright file in Blender. Instead we can live with some mild
inconsistency in the license names.
Trying to open the 2.74 Fishy Cat file would generate endless asserts in
customdata code. This commit refactors and cleans up the code of
`psys_get_dupli_texture` to avoid useless calls and data access, and
validate data before trying to further access it.
No behavioral changes expected here (besides getting rid of the assert).
`workspace_hook` of wmWindows store pointers for runtime data and to
data belonging to other IDs (workspace's layouts). That kind of pointers
should always be cleaned up on read, it allows for cleaner segfault
crash in case of mistakes in code updating/re-setting them, and avoids
potential security issue of accessing random memory address.
No behavioral change expected here.
Relying on pointer addresses across different data-blocks is extremely
not recommended (and should be strictly forbidden ideally), in
particular in direct_link step of blend file reading.
- It assumes a specific order in reading of data, which is not ensured
in future, and is in any case a very bad, non explicit, hidden
dependency on behaviors of other parts of the codebase.
- It is intrinsically unsafe (as in, it makes writing bad code and making
mistakes easy, see e.g. fix in rB84b3f6e049b35f9).
- It makes advanced handling of data-blocks harder (thinking about
partial undo code e.g., even though in this specific case it was not
an issue as we do not re-read neither windowmanagers nor worspaces
during undo).
New code uses windows' `winid` instead as 'anchor' to find again proper
workspace hook in windows at read time.
As a bonus, it will also cleanup the list of relations from any invalid
ones (afaict it was never done previously).
Differential Revision: https://developer.blender.org/D9073
Gathers information for time spent in the various managers or object (Film, Camera, etc.) being updated in Scene::device_update.
The stats include the total time spent in the device_update methods as well as time spent in subroutines (e.g. bvh build, displacement, etc.).
This does not qualify as a full blown profiler, but is useful to identify potential bottleneck areas.
The stats can be enabled and printed by passing `--cycles-print-stats` on the command line to Cycles, or `-- --cycles-print-stats` to Blender.
Reviewed By: brecht
Maniphest Tasks: T79174
Differential Revision: https://developer.blender.org/D8596
Followup for rB8398050695
This requires moving a few enum definitions to DNA instead of BKE, and
adding default definitions for the two structs the cloth modifier where
the cloth modifier stores its settings. The defaults are also reordered
to be consistent with the order of each item in the struct.
When a Shape Key is active, use the PBVH deformed coordinates for the
preview.
Reviewed By: sergey
Maniphest Tasks: T80873
Differential Revision: https://developer.blender.org/D8921
This was introduced in 6c9ec1c893. The overlays can now be drawn when
PBVH drawing is not enabled, but the PBVH should still exist in the
SculptSession in order to draw them. Before, it was avoiding the crash
by checking use_pbvh as BKE_sculptsession_use_pbvh_draw also checks if
the PBVH exists.
Reviewed By: sergey
Maniphest Tasks: T81268
Differential Revision: https://developer.blender.org/D9044
When using anchored strokes, inbetween events are not needed since there
is no stroke path to be reconstructed. Handling the inbewteen events are
only producing unnecessary updates, creating performance issues.
Reviewed By: sergey, brecht
Differential Revision: https://developer.blender.org/D9053
This simulation area mode moves the active area with the brush. When
enabled, the cloth brush has no restrictions on stroke length, area or
mesh vertex count.
In order to work, this enables PBVH nodes dynamically for simulation as
the stroke location moves and builds the constraints for new nodes
during the stroke. When a node is not inside the simulated area, all the
constraints that were created for it and vertex collisions are not
computed. The simulation limits falloff areas and constraints tweaking
control how the simulated and no simulated nodes blend.
Reviewed By: sergey, zeddb
Differential Revision: https://developer.blender.org/D8726
1ul << n will still be a 32 bit integer regardless
of the value of n, given the target here is 64 bits
the upper 32 bits will always be zero. Using 1ull
will yield the expected result.
This removes `foreachObjectLink` from `ModifierTypeInfo`, `GpencilModifierTypeInfo`
and `ShaderFxTypeInfo`. There is no need to have both, `foreachObjectLink` and `foreachIDLink`.
There is not code that actually depends on `foreachObjectLink`.
Reviewers: brecht
Differential Revision: https://developer.blender.org/D9078
The adds a new option to simplify volumes in the viewport.
The setting can be found in the Simplify panel in the render properties.
Volume objects use OpenVDB grids, which are sparse. For rendering,
we have to convert sparse grids to dense grids (for now). Those require
significantly more memory. Therefore, it's often a good idea to reduce
the resolution of volumes in the viewport.
Reviewers: brecht
Differential Revision: https://developer.blender.org/D9040
Ref T73201.
Remove the workaround for T59395 that increases the minimal refinement to 2
to work around an OpenSubdiv bug. This bug appears to be fixed in the latest
OpenSubdiv version we are using.
Problem found by Piotr Ostrowski.
Ref D9076
As noted in T80164, there are quite a few area of Blender where the
"Reset to Default Value" operator in button context menus doesn't work.
Modifiers are one of them, because the DNA defaults system was never
set up for them.
Additionally, this should make modifier versioning easier. Whenever a
new field is added it should be automatically initialized to the
default value.
I had to make some ordering changes in the following modifiers to work
around an error with `-Wsign-conversion` in the macros:
- Solidify Modifier
- Corrective Smooth Modifier
- Screw Modifier
Some modifiers are special cases and are skipped in this commit:
- Data Transfer Modifier
- Cloth Modifier
- Fluid Modifier
- Softbody Modifier
Differential Revision: https://developer.blender.org/D8747
This was already changed for Material nodetrees on duplication in
rBa75ac18638f4.
Since it is not obvious from the UI how change World actions - and to be
consistent with Material actions, it is best to copy the action as the
default behavior.
So use generic BKE_id_copy functions with LIB_ID_COPY_ACTIONS flag [which
also enables us to get rid of `BKE_world_copy`]
Note: taking the User Preference `USER_DUP_ACT` into account here (for
both material and world actions) could be a followup step.
Maniphest Tasks: T81248
Differential Revision: https://developer.blender.org/D9046
Most collection operations should not be allowed when either the scene
or the parent collection are linked or liboverrides.
This commit adds more stricter checks in both polling and execution
aspects of several related operators.
Fix T81282: Crash on instance to scene of a collection of a linked scene.
Now Wintab is not initialized when starting Blender with the tablet API
preference set to native, since that disables Windows Ink.
Note that changing the tablet API requires restarting Blender for changes
to take effect. This serves as a stopgap to allow use of Windows Ink until
runtime API switching is merged.
Differential Revision: https://developer.blender.org/D9051
- Rename `find_elem` to `elem_offset` (matching `elem_exists`).
- Remove unused `SDNA_StructMember` return argument.
- Return an offset instead of a pointer which was being converted
back into an offset by one caller,
in this case there was no way to tell the difference between
and element that doesn't exist and a struct member
at the start of the array.
Resolves UBSan warning raised in T81340.
It looks like this code was left over from tabbed panels in the
properties editor. It wasn't used anywhere except for in one line of
the horizontally-aligned panel code that was recently removed.
Differential Revision: https://developer.blender.org/D8651
Running Blender on Ampere cards was already possible with ptx, this fix is
needed to support building CUDA binaries.
Note the CUDA version used for official Blender builds is still 10, this is
merely the change to make it possible for those using CUDA 11 and specifying
the sm_8x kernels to be compiled.
Found by Milan Jaros.
This adds the option to either smooth the entire boundary, or to keep
corners sharp, for the Subdivision Surface and Multiresolution modifiers.
This mainly helps with compatibility with other software. The default
behavior remains to smooth the entire boundary.
Differential Revision: https://developer.blender.org/D8485
This makes subdivision surfaces compatible with the old subdivision
surface modifier and other applications that do not use the limit surface.
This option is available on the Subdivision Surface modifier.
Differential Revision: https://developer.blender.org/D8413
On user level this fixes dead-lock of OpenCL render on Intel Iris GPUs.
Note that this patch does not include change in the logic which allows
or disallows OpenCL platforms to be used, that will happen after the
kernel fix is known to be fine for the currently officially supported
platforms.
The dead-lock was caused by wrong usage of memory barriers: as per the
OpenCL specification the barrier is to be executed by the entire work
group. This means, that the following code is invalid:
void foo() {
if (some_condition) {
return;
}
barrier(CLK_LOCAL_MEM_FENCE);
}
void bar() {
foo();
}
The Cycles code was mentioning this as an invalid code on CPU, while in
fact this is invalid as per specification. From the implementation side
this change removes the ifdefs around the CPU-only barrier logic, and
is implementing similar logic in the shader setup kernel.
Tested on NUC8i7HVK NUC.
The root cause of the dead-lock was identified by Max Dmitrichenko.
There is no measurable difference in performance of currently supported
OpenCL platforms.
Differential Revision: https://developer.blender.org/D9039
Implemented Brown-Conrady lens distortion model with 4 radial and
2 tangential coefficients to improve compatibility with other software,
such as Agisoft Photoscan/Metashapes, 3DF Zephir, RealityCapture,
Bentley ContextCapture, Alisevision Meshroom(opensource).
Also older programs: Bundler, CPMVS.
In general terms, most photogrammetric software.
The new model is available under the distortion model menu in Lens
settings.
For tests and demos check the original patch.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D9037
The "UI_panel_set_expand_from_list_data" doesn't need to be in the
public API since it's just called every time an instanced panel is added.
This commit just sets the expansion automatically and adjusts some
naming to account for the moved function.
Before this change, when users switch from edit mode to sculpt mode, the
entire mesh would be visible. Even if in the edit mesh mode part of it was
set to invisible.
With this change the visibility is preserved, by creating a separate face set
for the visible and invisible parts of the mesh and setting their initial visibility.
Implementation details: This adds a function to initialize a new Face Set
datalayer taking the current mesh visibility into account which is stored
in the ME_HIDE flag of the vertices.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8901
This tool projects all vertices to the right of the plane defined by the
line gesture towards the plane. By doing this, this tool can create cuts
and plane surfaces in the mesh without modifying the geometry or using
boolean operations, so it is much faster than bisecting the mesh for
cases where the geometry was going to be remeshed afterwards.
Added as experimental as it does not have icon.
Reviewed By: sergey, Severin
Differential Revision: https://developer.blender.org/D9021
`SOCKET_OFFSETOF` was added in the initial commit {rBec51175f1fd6c91d5}
when `offsetof` [1] was not supported well enough. GCC and LLVM
support it since C++17.
Other two changes: type and size check can be done without creating
an invalid address too.
[1] https://cppreference.com/w/cpp/types/offsetof
Reviewed By: campbellbarton, brecht
Maniphest Tasks: T81100
Differential Revision: https://developer.blender.org/D9042
This took more than a day to fully investigate and understand, one of
the reasons being that the probability of the issue to show up was
extremely low, and subjected to very specific random factors.
Root of the issue is that, in some very rare cases, a newly read ID
might get the exact same memory address as the one it had when the blend
file was saved.
In that case, `BKE_workspace_active_set` would return immediately, since
the pointer to the active workspace would remain unchanged. But that
lead to having an unset NULL active layout pointer, which would crash
when attempting to get e.g. the active screen.
For the record, I ran into this when running a specific build (master
with one flag added to the `LIB_ID_CREATE` ones, with value `1 << 3`),
using a specific set of options (`--background --factory-startup -noaudio`),
and passing the .blend file from T80090 as argument.
Remove the attempt to update the active layout pointers of each window
from whithin `direct_link_workspace`.
This piece of code was a nonsense for at least to critical reasons:
* Do not, never, ever, access data from another datablock within the
direct_link_... functions. Just don't. Don't try to be smart.
* Since it was trying (and failing) to update the active layout of every
window for every workspace, it was effectively setting those
`act_layout` pointers to NULL (remapping can only ever happen once,
trying to remap and already remapped new pointer is bound to fail in
any case).
Luckily (and funnily), this piece of code was actually harmless, since
setting the active layout would be overridden/redone later, in
`lib_link_windowmanager`, when updating their `workspace_hook` in
`lib_link_workspace_instance_hook`.
Note that the similar horror with `WorkSpaceDataRelation->parent` (which
points at a window) is kept for now, because that one is needed. Hope to
refactor it soon though.
Although I haven't seen this cause any visible errors, there is some
incorrect handling for setting panel expansion during search:
- Properly check if child panel is active first
- Don't stop traversal at headerless panels that could theoretically
have children
Add an option to disable Alembic vertex interpolation.
Bump subversion from 5 to 6.
Alembic stores mesh samples at specific time keys; when a frame in
Blender maps to a timecode between two samples, Blender will interpolate
the mesh vertex positions. This interpolation only happens when the mesh
has a constant topology, but sometimes this was not detected properly
when the vertices change order, but the number of mesh elements remains
the same. This would result in a mesh with jumbled up vertices (T71981).
With this patch, users have the ability to disable vertex interpolation.
An alternative would be to have better detection of topology changes,
but that that'll cause a considerable slowdown.
Maniphest Tasks: T71981
Differential Revision: https://developer.blender.org/D9041
This modifier can only be added to Volume objects. It takes a mesh
as input and generates a "density" grid near the surface or in
the enclosed volume.
Ref T73201.
Reviewers: brecht
Differential Revision: https://developer.blender.org/D9032
Previously, one could only select a volume object in the outliner
or by clicking on the object origin. This patch allows you to click
on the actual volume.
Furthermore, the generated (invisible) mesh that is used for
selection is also used to draw an outline for the volume object now.
Reviewers: brecht
Differential Revision: https://developer.blender.org/D9022
When loading large files that are more than a couple weeks old
(such that DNA has changed in that time), a significant amount of
time is spent in `DNA_struct_reconstruct`. This function takes a struct
in the old layout and creates a struct in the new layout from it.
This was slow because it was computing the diff between the struct
layouts every time a struct is updated. Now the steps for the struct
reconstruction is computed only once per struct. This information is
then used to actually reconstruct all structs that changed.
I measured about 10-20% speedup when loading Spring files.
E.g. `10.6s -> 8.7s` for `06_055_A.anim.blend` in BKE_blendfile_read`.
This percentage varies a lot based on the number of blocks that have
to be reconstructed and how much DNA has changed since they have
been written. In none of my tests was the new code slower than the
old code.
Reviewers: campbellbarton
Differential Revision: https://developer.blender.org/D8959
The data layout remains exactly the same..
This change just gives all the elements in `SDNA->structs` names,
making it more comfortable to work with the data.
Reviewers: campbellbarton
Differential Revision: https://developer.blender.org/D8926
The behaviour is a bit different from other find modules:
`FFMPEG_FIND_COMPONENTS` should contain the
list of required libraries before calling `find_package(FFmpeg)`.
If not set, a default set of most common libraries is searched for:
`avcodec avdevice avfilter avformat avutil`
http://ffmpeg.org/download.html
Ref D8855
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D8918
Add a dependency graph relation Force Object Animation → Scene Rigid
Body World Rebuild. This ensures that the rigid body world is rebuilt
when a force object is re-tagged for animation updates.
The extra relation doesn't add any new calculations when the animation
is running, as the Time Source node already had a relation to the
scene's `RIGIDBODY_REBUILD` node.
The relation is created directly to the `RIGIDBODY_REBUILD` Operation. I
would have liked to target the containing Component instead. However,
that has the `RIGIDBODY_SIM` operation as entry node, which isn't enough
to actually fix T80121.
Reviewers: Sergey
Differential Revision: https://developer.blender.org/T80121
Draw dashed lines for object child hierarchies in the outliner. This
makes the difference between collection and object hierarchies more
clear. Dashes are also consistent with viewport relationship lines.
Differential Revision: https://developer.blender.org/D8997
Compare mesh loop count with number of loop normals before reading the
loop normals.
Houdini doesn't always write the correct loop normals to Alembic. When a
mesh is animated and then replaced by a fluid simulation, Houdini will
still write the original mesh's loop normals, but the mesh
verts/loops/polys are from the simulation. In such cases the normals
cannot be mapped to the mesh, so it's better to ignore them.
Simplify `anim_flush_channel_setting_up()` by flipping conditions and
returning early. This makes it easier to understand what is actually
happening in the code.
No functional changes.
Split up `ANIM_flush_setting_anim_channels()` into smaller functions. This
makes it easier to understand how it works, and makes future improvements
also easier.
No functional changes.
In the 3D view sidebar, the active tool settings panel can be pinned to other
categories, and in those other categories it doesn't redraw when the active
tool changes. This commit checks for pinned panels from the "Tool" category
when checking whether to redraw.
Note that the relatively expensive string comparison is only done for
currently visible pinned panels.
Differential Revision: https://developer.blender.org/D9012
This adds support for line gesture to SculptGestureContext and
implements a Mask Line tool, which affects everything to the right of a plane
defined by the straightline gesture.
For this to work, a new WM_gesture_straightline_oneshot_modal is needed
which only runs exec when the gesture is over.
Added as experimental as it does not have icon.
Reviewed By: Severin
Differential Revision: https://developer.blender.org/D8722
face_set_color_default was not copied when copying meshes, so when
extracting the colors in the draw cache it was using a wrong default
color ID.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8957
This reverts commit 5eb5978043.
The change was problematic:
*Advanced users may prefer managing schemes manually and adding only
a few that they need, instead of one for every library/executable
Blender builds.
*If CMake creates schema files, it overwrites the changes a developer
made every time CMake is run. If Xcode creates scheme files, it keeps
settings/ environment variables intact.
If someone runs CMake very frequently, pass
`-DCMAKE_XCODE_GENERATE_SCHEME=ON` to `cmake` while configuring.
I didn't realize there were more duplicates while committing
{rBf1fee433be92}.
Original was added in {rB83f8223543f58c3b0881a03b6e9ddffff91}.
Duplicate was added in the merge {rB9e09b5c418c0a436e3c84ccf}.
Ref D8822
Setting `PLATFORM_LINKLIBS` is not required for clang, compiler and
linker flags are enough. Note that the change made in
{rBa4c5811e2127}) to `platform_apple.cmake` (appending to
`PLATFORM_CFLAGS`) has not been reverted. platform file shouldn't be
overwriting the flags.
`PLATFORM_LINKFLAGS` is overwritten by `platform_apple.cmake`, so no
point in setting it. Fixing that like `PLATFORM_CFLAGS` is out of the
scope of this change.
`PLATFORM_LINKFLAGS_DEBUG` has been replaced with generator expression
to include RelWithDebInfo and MinSizeRel build types also.
Object in the 2.7x file is on no layer at all, while this is not really
expected nor considered valid, better to cope with the case gracefully
than crash.
To be backported to 2.90 and 2.83.
Where the if statement was just a check for an error or an unhandled
condition and there are no else statements, the rest of the function
doesn't need to be indented.
Before, it might have been possible in some cases (when there were no
items at all anymore in the collection) to miss that cleanup step,
leaving ghost useless and invalid operations around.
These breadcrumbs used to be a panel in 2.79 and 2.80 development.
If a file was saved in those versions, and the panels rearranged in a
version that didn't have this panel, its sortorder would be changed
even though it wasn't visible or even used in that version.
A simple fix is to use a new ID name for the new version of this
panel in 2.91. "Buttons" is an old name anyway.
These tooltips mostly just repeated the property names, which is not
helpful at all. Changing that leaves space to be much more specific
about what the properties actually do.
Differential Revision: https://developer.blender.org/D8994
This was the behavior in old versions of blender.
During a transformation operation, when pressing a contrain key, the chosen
orientation is that of the scene.
If you press the same key, the orientation changes to Global or Local.
However, if you choose another contrain axis with the orientation changed,
the orientation does not return to the set for the scene.
It remains Global or Local.
Now when changing a contrain axis, no matter what the current orientation is,
it always returns to the scene orientation.
Actions are either locked to a specific ID type, or "floating". Actions
in the floating state are now locked when they are assigned to an ID block.
Previously (rB94b99b5d4a7c20cf2) this was done at evaluation time, which
caused various problems:
- The ID type was set on the evaluated copy, and inconsistently flushed
back to the original.
- A newly created Action could not be assigned to an Action constraint,
unless a depsgraph evaluation was be forced first.
This is now resolved by calling `BKE_animdata_set_action()` to set the
action (instead of direct assignment) where possible, and calling
`BKE_animdata_action_ensure_idroot()` otherwise.
Manifest Task: https://developer.blender.org/T80986
Reduce complexity of `rna_SpaceDopeSheetEditor_action_update()` by flipping
conditions and returning early.
The depsgraph tagging has slightly changed, in that the depsgraph is not
tagged at all when there is no actual animation data added. I still see
this as a non-functional change, though, as in that case nothing changed
and tagging and re-evaluating wouldn't make any actual difference.
No functional changes.
Simplify `BKE_animdata_set_action()` by flipping some conditions and using
early returns, and document which parameters can be `NULL`.
No functional changes.
This reduces the number of places that have to be modified
when a new modifier is added.
Reviewers: campbellbarton
Differential Revision: https://developer.blender.org/D9000
With this change `outliner_draw.c` does not have to be
edited anymore when a new modifier is added.
Reviewers: campbellbarton
Differential Revision: https://developer.blender.org/D8998
Handling of pointer replacements in collection items lead to validating
all operations for a given property, when some might actually be no more
valid and would need to be cleaned up.
Related to T81059, found while investigating it.
Print out messages into console instead when se cannot find expected
data (this is actually normal when source library gets edited).
Note that there seems to be some issue with the 'auto-cleanup' mechanism
removing no more valid override rules on file save, will check on that
next.
Related to T81059, found while investigating it.
Fully revert D7913 "Fix T77276: Generating Python API docs raises many warnings"
{D7913} broke xrefs/links to other types and permalinks in the docs. This makes the python api docs for 2.90 and 2.91 completely unusable.
It got partially reverted in commit e893430a63. That didn't fix those two issues though, so it should be fully reverted.
As you can see here, i'm not able to click `bpy_struct.id_data`, because no <a> tag got generated in the html.
{F8889934}
Here you can see a working, but wrong permalink, generated by clicking the little chain. (It should be `bpy.types.MeshVertices`, not `bpy.types.MeshVertices.MeshVertices`)
{F8889938}
Reviewed By: Blendify
Differential Revision: https://developer.blender.org/D8913
The operator has some problems with Undo, so better put it in
experimental for now.
Reviewed By: campbellbarton
Differential Revision: https://developer.blender.org/D9007
We can only re-apply overrides fron the old local overrides to the newly
generated ones after all IDs have been properly remapped and renamed.
Otherwise override operations based on ID names may fail.
Related to T81059, found while investigating it.
Part of the code handling deletion of old, not needed anymore local
override IDs, was not working properly, effectively only deleting one
ID ever.
New code should also be a bit faster, though this should not be really
visible from user perspective.
Related to T81059, found while investigating it.
This commit replaces the BVH Tree currently used by the Weld Modifier
with the KD Tree used by `Merge > By Distance`.
This changes the result of the Weld Modifier to more closely match
`Merge > By Distance`.
There is also a big performance advantage.
Here is an overview (models in D8995):
| 2.90 (Duplicate Limit = 0) | 2.90 (Duplicate Limit = 1) | master (BVH) (Duplicate Limit = 1) | patch (KD) |
| 1.69s| 0.17s | 0.12s | 0.029s |
Differential Revision: https://developer.blender.org/D8995
More agressive optimization made the results differ a bit more than the
current error margin would allow. Bump the error margin to be 1e-6
instead of the previous 0.5e-7.
Since the search is applied all in one phase, there is no need to store
a reference to the search filter in every uiBlock. Instead just pass it
as an argument to UI_block_apply_search_filter.
Root of the issue is how we generate the storage ID for the differential
override operations.
However, since those are disabled anyway currently, simply comment out
creation of this copy for now, we can revisit this if/when we decide to
re-activate differential overrides.
Related to T81059, found while investigating it.
Before we would continue checking normal array values even if we knew
that the normal array would be conidered valid.
Break early to avoid excess iterations and make the code more clear what
it is doing.
No functional changes.
In very rare occations, the returned derivates would be the same. This
would lead to the normal calculation breaking (zero normals).
Solution: Add this edge case to the other corner case checks.
Reviewed By: Sergey
This is was caused by incorrectly set `preview_render_size` in VSE
rendering context. Value was set to `SEQ_PROXY_RENDER_SIZE_FULL`, but
it should be `SEQ_PROXY_RENDER_SIZE_SCENE` as scene render size is
being used.
Alternatively we can check for `context->for_render` in
`input_preprocess()`, but I think fix above is more correct.
Reviewed By: sergey, brecht
Maniphest Tasks: T80424
Differential Revision: https://developer.blender.org/D8838
Tag an Action for recalculating animation when it's pushed down onto the
NLA stack, as its effective boundaries change. Normally an Action spans
all of the timeline, and the effects of Cycle modifiers are visible
everywhere. When the Action is converted to an NLA clip, the modifier's
effect should only be visible for the duration of the NLA clip.
Since porting this to C in rB850234c1b10a, Smart UV Project would not
add a UVMap (if none existed already) anymore.
Not having a UVMap already is a reasonable situation though when e.g.
starting off fresh by deleting an existing UVMap or also when applying
certain generative modifiers. This is also inconsistent with all other
unwrap operators (all of them will create a UVMap if none exists
already)
Differential Revision: https://developer.blender.org/D9001
For the final builds instead of leaving all the license files in the main
folder, we move them to a "license" folder.
Also, adding more licenses here (MIT, Apache, ...).
Differential Revision: https://developer.blender.org/D8999
Original was added in rB83f8223543f58c3b0881a03b6e9ddffff91.
Duplicate was added in the merge rB9e09b5c418c0a436e3c84ccf.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8822
Setting the search match flag every time property search runs can
invalidate the results for panel headers. Instead, clear the flag on
every redraw and or the result of every search in the panel to it.
The "bevel_factor" for curves also applies when there is no bevel and
only extrusion. rB4e667ecef92f addressed this by moving the factor
and mapping properties to their own subpanel, but the property
descriptions still don't reflect that. This commit replaces "Bevel" with
"Geometry" for this situation. The property identifiers are not changed.
Own mistake in rB9dcae4eb17d7b.
We cannot use ED_object_constraint_active_list in uiTemplateConstraints
since it is dependent on mode (pose vs. object). Now get object
constraints directly when use_bone_constraints is false.
Note: unfortunately a derivate of this bug has made its way into 2.90.1
Maniphest Tasks: T81057
Differential Revision: https://developer.blender.org/D8989
Do not escape flow / effector objects if they have a disabled use flow / use effector flag.
Vertex velocities still need to be kept track of in order to have correct object velocities when enabling flows / effectors intermittenly. It is possible though to skip the emission loop if the flags are disabled.
This fixes and reverts commit c7287ffaec
Due to hardcodded keys, the modifier for auto contrain plane did not
work with custom keymaps and was in conflict with other keyitems.
Its usability is also confusing since it cannot be used without
`MOD_PRECISION`
But instead of removing it, it is better to make this modifier compatible
with custom keymaps and keep the conflict.
Caused by still having 'texture_opacity' in the preests and the preset
operator (but this was removed in the 2.83 rewrite).
Maniphest Tasks: T81081
Differential Revision: https://developer.blender.org/D8990
This code is not used by the new fluid simulation system.
Therefore, it does not make sense to keep it around.
We do want to integrate the fluid system with the generic cache
system eventually, but refactoring the system is easier when there
is less dead code that has to be refactored.
Note, I could not remove BKE_ptcache_id_from_smoke entirely yet,
because the depsgraph seems to expect that object that
uses DEG_add_collision_relations also has a pointcache.
That is because it wants to reset the point cache when any of the
other objects changed (this does not work with fluids currently).
Reviewers: sebbas
Differential Revision: https://developer.blender.org/D8987
Script create_msix_package.py will download the ZIP file
from the given URL. It will create the MSIX package
with the version number and publisher ID given.
Strongly recommended are the path to a valid PFX file, and the
password to use that PFX file. These are needed for signing
the resulting MSIX package. The signing step is optional though,
but the resulting MSIX package cannot be installed outside of the
Microsoft Store
Example
set VERSION=2.83.2.0
set URL=https://download.blender.org/release/Blender2.83/blender-2.83.2-windows64.zip
set PUBID=CN=PUBIDHERE
set PFX=X:\path\to\cert.pfx
set PFXPW=pwhere
python create_msix_package.py --version %VERSION% --url %URL% --publisher %PUBID% --pfx %PFX% --password %PFXPW%
Requirements:
* Python default from the Microsoft Store should do (3.8)
* requests can be installed with `pip install requests`
Note that for an LTS release that gets uploaded to its own LTS application release
in the store you need to specify the `--lts` switch on the command-line to the script.
Upon completion there will be a file with the
name blender-2.83.2.0-windows64.msix. In case PFX file and its password were
given on the command line MSIX package will also be signed for the Microsoft Store.
Related Wiki page: https://wiki.blender.org/wiki/Process/Release_On_Windows_Store
Reviewed By: jbakker
Maniphest Tasks: T77348, T79356
Differential Revision: https://developer.blender.org/D8310
In the end the process is surpringly simple, we only need to manually
convert the proxy itself into an override (which is trivial), and then
run common code with the default 'make override' operation.
Fix T81059: Add operator to convert proxies to library overrides.
Some RNA setters require ID data they operate on to be in G_MAIN.
Unfortunately, when we apply overrides as part of a .blend file reading,
new Main is not yet made global one, so we have to do it temporarily
here.
This is a fairly ugly hack, but it should be harmless and safe.
This is a first step towards supporting conversion of proxies, done
separately to make it easy to pinpoint in case it would create problems.
It is not expected to cause any change in behavior currently.
Constructive modifiers were not initializing an original index layer
in the case a previous deform modifier created 'mesh_final'.
This happened in the case of multiple deform modifiers that deform along
normals, as requesting normals caused the final mesh to be created.
Ensure mapping data is created in the case
only non-constructive modifiers have run.
Mainly a UI adjustment, no functional changes
To have the default mode in the advanced panel as separated option is not the best solution.
Now, there is a pin option and when it is enabled, the brush keeps this mode.
Differential Revision: https://developer.blender.org/D8974
The problem is due to lack of precision with small and coplanar triangles.
Also, triangles that have vertices with the same coordinate are at the
threshold of the intersection.
We could add an epsilon to consider the minimum distance for intersection.
But that would add a lot of overhead to the code.
The solution used is to increase precision using doubles.
Some characters: `'`, `"`, and `\`, can cause problems with RNA paths.
Instead of using more complicated handling to deal with those cases,
we can just prevent these characters from being used in custom property
names.
This commit checks for these characters to `idp_try_read_name`, where
other checks like length are already done.
Differential Revision: https://developer.blender.org/D8839
Those flags are meant for detecting which socket has changed, so in the
future we can have more granular updates.
`Node` now stores an `update_flags` member which is modified every time
a socket is changed though `Node::set`. The flags are or-able bits
stored in `SocketType` instances. Each `SocketType` stores a unique bit
out of 64, for the 64 bits of an uint64_t; the bit
corresponds to the index of the socket in the `Node`'s sockets array +
1, so the socket at index 5 will have the 6th bit set as its flag. This
limits us to 64 sockets per Node, which should be plenty for the current
set of `Nodes` that we have.
This does not change the behavior of other parts of Cycles.
This is part of T79131.
Reviewed By: brecht
Maniphest Tasks: T79131
Differential Revision: https://developer.blender.org/D8644
This refactor is in response to an unreported bug in which overlapping, moving colliders produced an unstable simulation.
Things that change apart from cleanup through this refactor:
- Effector objects with no velocities (either non-animated or animated but non-moving object) will explicitly set zero velocities in their flow bounding box.
- When applying object velocities to the global grid, they will now be accumulated per cell (add and not set velocities). Later they will be averaged with the object count at any cell.
The OpenVDB update added a load() function. This function clashes with a helper IO function (only used when exporting and running the simulation externally) that was also named load().
Removed invel MAC grid since it is sufficient to use the cell centered vec3 representation. When setting initial velocities these will be interpolated by the setter functions.
With the new image editor drawing there were was some mutual exclusive
functionality. When rendering the alpha was shown correctly or the pure
emissive colors were shown correctly, but never both. The cause of this
is that the image_gpu did not used the correct alpha mode when generating
gpu textures for non-images (render results, compositors viewer)
The implementation always checked the alpha_mode. Alpha mode is an
attribute for images, but aren't set for non images. This patch adds
a more detailed check to ensure that the gpu texture is premultiplied.
The issue has been tested using several bug report files and production
files.
Reviewed By: Brecht van Lommel
Differential Revision: https://developer.blender.org/D8978
Since {D8234} the image editor is drawn using a depth buffer.
When using `draw_texture_2d` the image is drawn using the 2D_IMAGE
shader. inside the vertex buffer the image was pushed to the background.
This was introduced by {648924333234} what seems to be out dated as we
have done several overhauls in this area. (workbench refactor, overlay
engine refactor, color management pipeline).
This patch removes the pushing of the image to the background.
During viewport render animation the gpu textures weren't tagged as
invalid. As the image uv editor now draws the gpu texture only the first
was shown as it wasn't refreshed with the actual image data.
I created a bugreport T80859 earlier.
The problem is in the file source/blender/editors/sculpt_paint/paint_image_proj.c (line numbers against commit d376aea61)
at the lines 5309 and 5320 the functions `IMB_rect_from_float(...)` and `IMB_float_from_rect(...)` are called from threaded code on a shared ibuf.
However, ps->reproject_ibuf is shared between threads, so this is not thread-safe.
This is a race condition and leads to wrong output when projecting float data onto a uchar texture or vice versa on a multithreaded system.
The checks on line 5308 and 5319 are already a race condition.
I created this patch which fixes the problem , but I'm not sure if I'm overlooking something.
This makes sure reproject_ibuf contains the correct formats *before* the threadpool is started.
I replaced the original location of the conversions with BLI_asserts(). That may be overkill?
Reviewed By: #sculpt_paint_texture, mont29
Maniphest Tasks: T80859
Differential Revision: https://developer.blender.org/D8936
When introduced in rB1ca1744c29e2, the Weight Paint Overlay XRay's
corresponding depth pass was not considering clipping planes.
Maniphest Tasks: T81013
Differential Revision: https://developer.blender.org/D8970
target in certain situations
Regression from rBdeaff945d0b96.
mesh_ensure_looptri_data would overflow.
Crash would only happen if a Data Transfer modifier (transferring
UVs) follows, so exact reason for this is not yet entirely clear. Also
there are edit-mode versions of the following BVH lookup functions so it
could be avoided (since this is a expensive operation), marking as TODO.
Similar fix as
- rB0945a79ed1eafae444d3021a5912cb39801a7209
- rB56d7e39b92997768b3db8ce2dbc262f869994145
Reviewers: mont29, campbellbarton
Maniphest Tasks: T80996
Differential Revision: https://developer.blender.org/D8973
The placement of the start and end factor and mapping settings for
curves has been quite misleading for a long time. They were in the
"Bevel" subpanel, but they aren't related to bevel because they affect
curves with only extrusion and no bevel.
This commit moves these properties to their own subpanel, labeled
"Start & End Mapping".
Differential Revision: https://developer.blender.org/D8910
This commit contains the Performance improvement, that was originally
proposed in D8966.
It improves the performance of the Weld Modifier by a lot.
It had a loop with execution time O(N^2) which is now O(N*log(N)) at a
bare maximum.
This patch adds a new operator to convert a black and white image into
grease pencil strokes.
If the image is not B/W, an internal conversion is done.
This is the first operator using Potrace, but we expect to add more features in next Blender versions.
Reviewed By: HooglyBoogly
Maniphest Tasks: T79877
Differential Revision: https://developer.blender.org/D8951
The issue was two fold.
First something sets the loop element tag and doesn't clear it before
the UV code in question tries to use the tags. Added a sanity clear to
make sure that it operates on a clean tag state.
The next one was that the UV maps in question had quite a few points
that had zero length UV loop edges. This would lead to division by
zero.
Reviewed By: Jeroen Bakker, Brecht
Differential Revision: http://developer.blender.org/D8967
It's generally considered a better codestyle to check conditions early
and exit early when they are not met, over having most logic of a
function within a big `if`-block. Otherwise people have to go over the
entire block to see if there's possibly an `else` somewhere, or any
followup logic.
There's the old and ugly hack where the `uiBut.poin` points to the
button itself. When reallocating the button we have to update that
pointer of course.
Those two functions had `BKE_` prefix, were defined in BKE headers, but
implemented in ED code, yuck.
Moved everything to ED area for now, since those do not look fondamental
enough to belong to BKE, and none of their usages requires it currently.
As noted by @lichtwerk (thanks), one can have a local object and/or
material using a linked image data-block, this case needs some different
handling to prevent painting on such linked image.
For now, tweak `BKE_paint_proj_mesh_data_check` (eeeek, that name
prefix!) to consider paintslots with linked image as 'non-existing'.
The function this was in already checks the conditions under which it
can operate (as it should). No need to force the caller to do these
(quite specific) checks, the function can just do nothing if the button
doesn't need these operations.
Otherwise we'd have to always execute these checks before calling, which
makes calling it a hassle, makes the code harder to follow and generally
harder to maintain (what if the conditions change?).
Allows scripters to store additional information in the marker itself instead
of using work-around approach based on marker names and such.
Differential Revision: https://developer.blender.org/D8944
Originally was noticed when transforming mesh created by
object.to_mesh() from an object without modifier, in which case the
result references CustomData layers used by the object itself.
The issue goes a bit deeper: mesh.transform() should never modify
referenced layers, hence it should duplicate referenced layers.
This fix changes one specific aspect of the reported behavior. The
case where vertices coordinates are modified manually will still have
affect on the source mesh (as no referenced CustomData layers are being
duplicated). Proper fix for this case is not yet clear to me.
Differential Revision: https://developer.blender.org/D8939
The issue was that the pointcache was not storing dead particles,
even though they are displayed. This lead to the rendering issue,
because only alive particles can be read from the point cache in
the frame that is rendered.
This also fixes an issue unrelated to rendering: when dead particles
are displayed, their position is incorrect when some frames are
skipped during playback.
Reviewers: brecht
Differential Revision: https://developer.blender.org/D8907
Keep braces balanced where possible, even with ifdef's as it avoids
confusions with editors calculating correct indentation level &
finding matching brackets.
This patch tries to sanitize the types of our size parameters across our read
and write code, which is currently fairly inconsistent (using `int`, `uint`,
`size_t`...), by using `size_t` everywhere. Since in Blender file themselves
we can only store chunk of size `MAX_INT`, added some asserts to ensure that
as well.
See {T79561} for details.
Differential Revision: https://developer.blender.org/D8672
Collections of liboverride data only support insertion of new items (and
deleting or moving those around). No operation on existing items from linked
data is allowed.
Reflect that in RNA functions used to edit py-defined RNA collections.
Note that this is not ideal currently, it does the job but feedback to
user is rather poor. Ideally add-ons should implement higher-level
checks for override cases in their operators.
Supporting IDProperty-based data storage here in a generic way is fairly
simple, unlike actual RNA data (DNA-based), for which getting a proper
generic way to duplicate them from within RNA is currently not really
possible.
Part of T79562.
The bevel type is a more useful check for graying out this property
because its effects will be apparent before choosing an object. Before,
the fill type property would only gray out with a bevel object selected.
Warp & weight vertex-group editing modifiers miscalculated vertex weight
inversion, the weights were multiplied before being subtracted from 1.
Ref D8241
Build modifiers linked or copied to objects from the outliner would not
animate until the file was saved and reopened due to a missing depsgraph
relations tag.
Light rotation is supported both in world and view light orientation in
lookdev, so no reason to disable it.
Reviewed By: Blendify
Differential Revision: https://developer.blender.org/D8958
Expanded objects with children immediately after a colored collection
would have the color applied to their hierarchy lines as well. The fix
is to reset the collection color to `COLLECTION_COLOR_NONE` for each
open subtree.
Based on http://jcgt.org/published/0008/01/03/
This is a simple trick that does *not* have a huge performance impact but
does work pretty well. It just modifies the Fresnel term to account for
the multibounce energy loss (coloration).
However this makes the shader variations count double. To avoid this we
use a uniform and pass the multiscatter use flag inside the sign of f90.
This is a bit hacky but avoids many code duplication.
This uses the simplification proposed by McAuley in
A Journey Through Implementing Multiscattering BRDFs and Area Lights
This does not handle area light differently than the IBL case but that's
already an issue in current implementation.
This is related to T68460.
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D8912
The search should check if a panel is active before changing its
expansion, otherwise it sets the expansion for all of the region's
panels, even invisible ones in other tabs.
This was causing flashing colors in the node editor grid.
This is because in some cases the flat color is only set on the provoking
vertex which is the last of the primitive by default.
This makes wide line supported on MacOS and other implementation that
does not support wide line by default.
This workaround works for all Line types but only if using one of the 5
default shaders.
The workaround is completely isolated and invisible to the outside. It has
no side effect.
Note: This does not affect the GPUBatch drawing.
This is not an exhaustive change, just for the 2D cursor and
UDIM grid properties. Also vertically align the "UV Vertex"
buttons like in the 3D view panels.
Checking for the layout alignment is not a reliable way to filter out
which items need the additional icon/text offset. Instead check if the
buttons are icon-only (which we rarely do, e.g. for collection colors in
the Outliner context menu).
For example you can now start entering text in the Properties or
Outliner search, and press the 'x' while the button is still in
text-edit mode.
This way you don't have to exit text editing first, before being able to
quickly clear the string with a mouse click. So this is a small
improvement for convenience.
It also works for the eyedropper (change to picking an object while text
editing) or the '+' and '-' icons in the file saving dialog.
E.g. the 'x' icons or eyedropper icons in text buttons. They didn't use
to have any mouse over feedback, now we dim the icon until hovered.
This kind of feedback helps users see that the icons are interactive,
and if they are within their interaction hotspot.
I think this wasn't allowed before because the section of a curve was
built in multiple parts. But since rBe34d3e32dda7, the whole slice
of a curve is built in one piece, so we can easily support curve
caps for all geometry types, including the new custom profile option.
Note that this also allows "caps" when the fill type is not full.
They could easily be disabled by checking for "Full" fill type
if that was preferred in the future.
See the patch for images.
Differential Revision: https://developer.blender.org/D8911
This adds XYZ symmetry as a property of meshes and updates all modes to
use the mesh symmetry by default to have a consistent tool behavior
between all modes and when switching objects.
Reviewed By: brecht, mano-wii, campbellbarton
Maniphest Tasks: T79785
Differential Revision: https://developer.blender.org/D8587
This object operator exits and frees the edit data of the
current object and enters the same mode in another one in a single step,
without going through object mode or keeping multiple edit object data
active. It is assigned to the D key.
This solves all conflicts that the right/click select keymap and the
emulate 3 button mouse produces for this operation and it is independent
of the state of Lock object modes.
Also, as the SculptSession is freed, when using Multires objects go
back to their preview resolution level, so it is possible to work on
high vertex count scenes without slowing down the viewport and other
performance problems.
Reviewed By: #user_interface, pablovazquez
Differential Revision: https://developer.blender.org/D7510
This removes the limitation of the sculpt overlays not being visible
with modifiers active.
Reviewed By: fclem
Maniphest Tasks: T68900
Differential Revision: https://developer.blender.org/D8673
This implements a new overlay that blends the bakground color over the
objects that are not in the same mode as the active object, making
them fade with the background.
This is especially needed for sculpt mode as there is no other overlay
or indication in the viewport to display which object is active.
This is intended to be used with D7510 in order to have a faster
workflow when sculpting models with multiple objects.
Reviewed By: fclem
Differential Revision: https://developer.blender.org/D8679
This adds an option for the HDRI rotation to follow the view rotation.
When this option is enabled, this allows EEVEE materials to be used as
matcaps for sculpting and painting.
This has an extra performance cost when orbiting around the model as the
lookdev cache needs to be recalculated, but in my test it is barely
noticeable.
Reviewed By: fclem
Differential Revision: https://developer.blender.org/D8566
This filter scales the mesh as it was a softbody, producing folds in the
surface. The orientation of the folds can be controlled using the filter
axis and orientation.
This is an example of a cloth filter that uses deform coordinates instead
of forces, but probably it does not make much sense to expose it to the
user in a different way and with different parameters. I'll remove
FilterCache->enabled_force_axis in a later commit and use always
enabled_axis in SCULPT_filter_zero_disabled_axis_components
for both forces and deformation filters, so this function can also be used
in the mesh filter.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8661
The deformation constraints strength were too strong and they were
preventing the cloth effects of the brushes with cloth deformation target
to create folds properly. This lowers the default, making the simulation
follow the deformation in a more relaxed way.
I'll make a separate patch to expose this as a property for certain brushes
and cloth deformers that may need higher values (like boundary with loop
falloff on a low poly mesh), but I think this default will work better for
most use cases.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8884
This new feature allows to use the strokes as an eraser of any stroke below. This is very handy to open holes in filled areas.
After running some tests, we have decided to keep the additive effect of the holdout color. To get clean holdout areas, just move the color to black to remove any additive effect. To have additive effect can be used in situations like tint slightly a transparent window with blue to simulate the glass.
See T79878 for more details
Differential Revision: https://developer.blender.org/D8932
This was a bug in OpenImageDenoise, upgrade to latest 1.2.3 which has the fix.
This in turn required a newer ispc version.
Differential Revision: https://developer.blender.org/D8892
Steps to reproduce were:
* Create a marker in the timeline
* With the mouse over the marker region, press Shift+D
The function for moving markers was reused and had handling specific to
tweak events. This is not relevant in case of duplicating markers.
When repeat image was active the image was drawn in the foreground
making most uv overlays not visible. This change creates a shared
defined value for repeating and not repeating images.
Ref {D8855}
Unix and Apple platform files use find_package(OpenSubdiv) which sets
`OPENSUBDIV_INCLUDE_DIR` as an advanced variable, as well as
`OPENSUBDIV_INCLUDE_DIRS` which should be used usually.
Windows sets `OPENSUBDIV_INCLUDE_DIR` which is used by the rest
of the code.
This patch renames it to `_DIRS` everywhere, for it to be like other
similar variables.
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D8917
situations
In the situation that the PTCacheEdit is not available (e.g. when
editing unbaked Cloth or Softbody caches - or, as in the report, being
in Cloth or Softbody Editing Type for Hair particles), accesing the
corresponding ParticleSystem or ParticleSystemModifierData would crash.
Now access these later (after PE_start_edit had the chance to return
early on this non-valid situation).
Reviewers: JacquesLucke
Maniphest Tasks: T80900
Differential Revision: https://developer.blender.org/D8941
Was missing a dedicated entry for LMB select, this is to be consistent
with how it is done for meshes.
Maniphest Tasks: T80905
Differential Revision: https://developer.blender.org/D8935
Caused by rB9a7f5f1bb422.
If using Auto IK (or targetless IK and Auto IK together), two temporary
constraints were added.
- from pose_grab_with_ik_add (even for targetless IK)
- from add_pose_transdata (even for Auto IK)
Since both both do similar things, but cannot work in tandem (with
possibly different chainlengths for example), we have to decide which
type to prefer over the other (as in: do not create a constraint for the
other).
It seems better to ignore the 'Auto IK' option on bones that will
have targetless IK set up for them specificallly [e.g. defining special
chainlength]. This way you can still work with 'Auto IK' ON generally
[with interactive chainlength control], but also have specific bones that
need their own custom chainlength.
For now, the most straightforward fix is to
- only add constraints for Auto IK from pose_grab_with_ik_add()
- only add constraints for targetless IK from add_pose_transdata()
Note: this area has some potential for later refactoring:
- move creation of all temporary constraints to a single place
[preferably pose_grab_with_ik_add]
- use only those temporary constraints in transform code [atm. we still
flip CONSTRAINT_IK_AUTO around on the "original" -- unneccesarily, after
rB9a7f5f1bb422 a dedicated temporary constraint is now always available]
- clarify CONSTRAINT_IK_AUTO vs. CONSTRAINT_IK_TEMP
- obeying standard rotation locks on bones in the chain (not just the
the IK locks) is not consistent between targetless IK and Auto IK
Potential candidate for 2.90.1 as well as 2.83 LTS
Maniphest Tasks: T80437
Differential Revision: https://developer.blender.org/D8930
Allows to more easily access time information about how long Cycles
did spend synchronizing objects from evaluated depsgraph on Blender
side to its own structures.
This timer does not include time spent evaluating render depsgraph.
Regression from {b248ec97769f}. A new parameter was introduced, but the
stub shader macros still had the old number of parametes. This
change adds a new dummy parameter to the stub macros.
mode
The transform convert of type TC_POSE was not being defined if the mode
was OB_MODE_WEIGHT_GPENCIL.
Maniphest Tasks: T80904
Differential Revision: https://developer.blender.org/D8934
Assigning window variables relied on notifiers that always used the
active window, no matter the window being assigned.
Pass the current window to the notifier.
This code had grown overly complicated and hard to understand.
The improvements in this commit:
- Avoid allocating a duplicate of every active panel.
- Instead of complicated logic to calculate each panel's offset,
just keep track of the current offset while iterating through.
- More readable code structure, better comments.
Note that calcuting the X offset here is a relic from pre-2.5 when
panels could be aligned horizontally. I kept this in, but it would
be reasonable to remove it in the future.
This makes these variable names more consistent with current standards,
mostly by not over-abbreviating already short words. The following commit
will refactor this function.
When entering scultp mode the visibility from the Face Sets is copied to
the base mesh. This steps was considering that if a vertex belongs to a
face with a visibible Face Set ID, it should be visible. As loose
geometry may not have any faces, those vertex were set to hidden.
Now this function check if a vertex visibility should be modified by the
face sets (by checking the loops), avoiding modifying the visibility of
loose geometry unintentionally.
Reviewed By: sergey
Maniphest Tasks: T80771
Differential Revision: https://developer.blender.org/D8899
Brushes that target the cloth simulation but are not the cloth brush
affect the entire mesh, so they don't have simulation areas and falloff.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8885
Just converts verts to points and vice versa.
Materials and Attribute layers are preserved (so for example if you set
custom radii on the pointcloud, convert to mesh, then convert back to
pointcloud, this will be preserved).
Also not add a Radius layer by default (it is still added and
filled when adding a pointcloud object from the menu), a global Radius
property that will be used if there is no radius attribute can be added
later. A Radius attribute can also be added in the pointcloud data
properties (and filled via python).
This will also add a new utility function that copies materials between
datablocks: BKE_id_materials_copy
ref T75717
Differential Revision: https://developer.blender.org/D7391
Following rB7a0a60dde8b4, checking for temporary IK constraints in
uiTemplateConstraintHeader is not needed anymore (since the whole panel
will now be skipped earlier for those).
Differential Revision: https://developer.blender.org/D8902
(e.g. cannot be animated)
An RNA path is mandatory for RNA_path_from_ID_to_property, and for
EffectorWeights, this has to be provided using RNA_def_struct_path_func
(since this struct is shared across ParticleSettings, Modifiers and Rigid
Body World). Without this, the animation system cannot work and e.g.
keyframes cannot be inserted.
The case for Rigid Body World in rna_EffectorWeight_path was missing,
now added.
Rigid Body World Field Weights are capable of being animated, tested
with file from T80131 (e.g. gravity, wind).
Maniphest Tasks: T80131
Differential Revision: https://developer.blender.org/D8713
The `find_package()` callers expect the library name as `Potrace`, not
`POTRACE`.
Reviewed By: LazyDodo
Differential Revision: https://developer.blender.org/D8923
rBf2d26409e83b fixed a serious problem with invalid mesh after an
operation with the extrude manifold.
Since BKE_mesh_validate_arrays is a slow operation, the chosen
interval between versions is relatively small and also only the
mentioned invalid mesh problem is tested.
Differential Revision: https://developer.blender.org/D8898
The ffmpeg, guardedalloc and blenlib are quite isolated and putting them in
their own executable separate from blender_test is faster for development than
linking the entire blender_tests executable.
For Cycles, this also bundles all the unit tests into one executable.
Ref T79958
Differential Revision: https://developer.blender.org/D8714
Bump Blender subversion from 4 to 5, and re-execute some versioning code.
Because of a mistake in the versioning code (corrected in rB1a4fc6dcd67b),
some of the versioning code didn't run in certain cases. This subversion
bump ensures that this code is still run.
This solves annoyance that the official RPM repository of nasm requires
newer version of rpm tool compared to what is shipped by default with
centOS 7.
Differential Revision: https://developer.blender.org/D8905
Using context overrides in Python caused problems for any operator that
changed the context and require these changes to be read back.
CTX_wm_area_set() for e.g. would set the struct member but future
calls to CTX_wm_area() would still return the value defined by Python
callers context overrides.
This also resolves a mismatch between polling and calling operators
from Python, where poll would override the Python context where calling
only overrode the context when a new context was passed in.
While this isn't an issue with the default configuration it's possible
to register extra header types for a single region.
In this case the first header-type to successfully poll is drawn
without drawing other header types.
This issue was raised by T60195.
This adds support for the same custom bevel profile widget used in
the bevel tool and modifier to the geometry generation for curves.
This is expecially useful for text and 2D curves with extrusion, as
it works much better than a weld & bevel modifier combination.
It can also be useful for adding quick detail to pipe-like objects.
The curve holds the CurveProfile struct and a new "Bevel Mode"
property decides which type of bevel to build, round, object, or
custom profile.
Although curves can already use another curve to make the bevel
geometry, this is a quicker way, and it also defines the profile of
just one corner of the bevel, so it isn't redundant. It's also nice
to have the same custom profile functionality wherever there is bevel.
Differential Revision: https://developer.blender.org/D8402
There is no ideal solution here, but we can assume by default that we should
use the Alpha output from a texture used for transparency data.
This will break people using a dedicated B&W texture for this alpha
pass, but we cannot really handle all cases properly in this wrapper, we
only try to support the most common ones to some extent.
There must not be any return directly in our versioning patches. It
would return from the entire versioning function.
Mistake in da95d1d851.
As a result, when opening old files with animation data, the versioning
for the new collection color tagging wouldn't run, and all collections
would get the first color assigned.
Adds an optional slider to the action constraint so that it can be
driven without a constraint target.
This is very helpful for more complex rigging and mechanical rigs, as it
means the action constraint can be controlled with a driver/custom
property directly, currently if we want to use a driver to control it we
must add a "dummy" bone/object inbetween to act as a control.
Reviewed By: Sebastian Parborg, Sybren A. Stüvel, Demeter Dzadik, Julian Eisel
Differential Revision: http://developer.blender.org/D8022
The output layout was wrong and it's a mistery why it works on most
implementations since it's clearly a wrong usage.
Thanks @sebbas for helping narrowing down the issue.
Since this buffer is used as an array of 12 32bits integers, and C++
string expect a NULL-terminated C-string, we need an extra char to
ensure last one is always NULL.
See D8906. Thanks to @brecht for noting this one too.
Currently, it's difficult to use the NLA system for users who are only
interested in using it as animation layers.
Entering tweak mode hides strips which are not evaluated. If the user
wanted to edit a different strip, they must exit tweak mode, look for
the strip, select it then re-enter tweak mode.
Solution:
All strips are always shown. The user can now see the next strip they
want to start editing.
Reviewed By: Sybren, Sebastian Parborg
Differential Revision: http://developer.blender.org/D7600
Since this buffer is used as an array of 12 32bits integers, and C++
`string` expect a NULL-terminated C-string, we need an extra char to
ensure last one is always NULL.
See D8906. Found while investigating T80657.
Previously, it was kind of a mess. In different places it was using `char`, `short` and `int`.
The changed properties are flags that are operated upon using bit operations. Therefore, the integer type should be unsigned. Since we only use 2 bits of these flags, `uint8_t` is large enough.
Especially note the change I had to make in `RNA_define.h` to make this work.
Reviewers: campbellbarton
Differential Revision: https://developer.blender.org/D8844
When Showing Meta data for an image where the buffer does not exist
(missing file) it crashed. This patch removes the check on the image and
only checks the availability of the buffer.
CustomData_blend_write_prepare might modify the `CustomData`
instance, making it invalid afterwards. It is not necessary to do this
in other cases explicitly, because when writing shallow copies of
ID data blocks are made.
This commit makes grid theming more consistent and capable by adding
some new theme colors related to grid rendering.
- Add grid theme color for node editor. `UI_view2d_multi_grid_draw`
is called with TH_GRID instead of a shaded `TH_BACK`.
Also color-blend `TH_NODE_GROUP`.
- Make the movie clip editor's clip preview grid respect grid theme
color (`ED_region_grid_draw` uses color-blended `TH_GRID`).
- Add versioning code to allow fixing existing themes (the resulting
themes should visually look the same as before)
These changes did cause some inconsistencies in the movie clip editor,
even after adjusting the themes accordingly:
1. The alpha slider of the grid color affects the background and not
the grid lines themselves.
2. The grids used by graph and dopesheet mode could already be themed
in the past. Now that the clip preview's grid can also be themed,
two different modes share the same theme color.
Differential Revision: https://developer.blender.org/D8699
This adds an operator to allow drag and drop of modifiers, constraints,
and shader effects within the outliner. Referred to as "data stack" in
the code for simplicity.
The following operations are allowed:
* Reordering within an object or bone
* Copying a single modifier/constraint/effect to another object or bone
* Copying (linking) all modifiers/constraints/effects to another object
or bone.
This complements the recent work done for panel-based modifier layouts
by allowing reordering in the outliner. It also makes it simple to copy
a single modifier/constraint/effect to another object.
Differential Revision: https://developer.blender.org/D8642
Adds functions to copy a modifier between objects, and an ED_ level
function for linking modifiers between objects. This will be used in
outliner modifier drag and drop.
These functions support both regular and grease pencil modifiers.
Differential Revision: https://developer.blender.org/D8642
Add functions to copy a single constraint between objects or between
bones, and another function to link constraints. This is in preparation
for constraint drag and drop in the outliner.
Differential Revision: https://developer.blender.org/D8642
Move the notifiers and DEG tagging to the ED_* level functions for
modifiers and gpencil shaderfx in preparation for outliner modifier and
shaderfx drag and drop. No functional changes.
Differential Revision: https://developer.blender.org/D8642
No functional changes. Move the constraint reordering logic into an ED_
level function to be used by outliner constraint drag and drop.
Differential Revision: https://developer.blender.org/D8642
Adds two functions: one to copy a shaderfx between two gpencil objects,
and another to link all shaderfx between two gpencil objects. Added in
preparation for outliner shaderfx drag and drop.
Differential Revision: https://developer.blender.org/D8642
No functional changes. Split the grease pencil and object copy logic
into separate functions. This makes the code cleaner and prepares
utility functions for outliner modiifier drag and drop.
Differential Revision: https://developer.blender.org/D8642
Stack panels (for modifiers, etc..) also get their expansion from their
associated list data. This means that property search expansion needs
to properly update the list data, which can be accomplished by calling
`set_panels_list_data_expand_flag`.
This commit also moves this logic to `UI_panels_end`, where it fits
better.
While testing for {rB40dcf686f04f}, compiler flags got mixed up and
non-working ASan configuration was committed.
Platform file, which is `include`d after the `CMAKE_C_FLAGS_DEBUG` etc.,
are set, overwrites those flags instead of appending to them. To fix this,
`PLATFORM_CFLAGS` is used to pass the `-fsanitize=*` flags to the C/C++
compiler.
Tested on fresh build using both Xcode and Ninja, with & without ccache.
Also silence a clang warning for multi-config generators:
the object size sanitizer has no effect at -O0, but is explicitly
enabled: -fsanitize=object-size [-Winvalid-command-line-argument]
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D8879
The change builds GMP with `--with-pic` flag on non-arm architecture.
The warning:
ld: warning: PIE disabled. Absolute addressing
(perhaps -mdynamic-no-pic) not allowed in code signed PIE, but used in
___gmpn_divexact_1 from lib/darwin/gmp/lib/libgmp.a(dive_1.o). To fix
this warning, don't compile with -mdynamic-no-pic or link with
-Wl,-no_pie
Reviewed By: sebbas
Differential Revision: https://developer.blender.org/D8823
For work the GP team plans to land soon (T79877) potrace was taken
on as an additional optional dependency.
This diff adds building the library to the deps builder and takes
care of the integration into the build-system with the `WITH_POTRACE`
cmake switch.
Differential Revision: https://developer.blender.org/D8662
Reviewed by: brecht, sergey
Adds a new hierarchy line that draws to the left of collections. If the
collection is color tagged, then the hierarchy line is drawn in that
color. This is useful to see the color tag of a collection that exceeds
the vertical height of the outliner when expanded.
This also modifies the object hierarchy line drawing to match the new
collection hierarchy line, with no horizontal lines, and moved a unit to
the left. Object lines are not drawn colored.
Manifest Task: https://developer.blender.org/T77777
Differential Revision: https://developer.blender.org/D8622
This adds color tagging to collections. There are 8 color
options which are themable in the user preferences, with an additional
option for no color tag by default.
This adds a new filled collection icon and 8 colored variants of the
icon that can be themed in the user preferences.
In this commit the only interface to setting the color tags is through
Python, and there is nowhere in the interface where the collections are
shown colored. Setting and viewing the color tags from the outliner will
follow.
Manifest Task: https://developer.blender.org/T77777
Differential Revision: https://developer.blender.org/D8622
All the changes made in the branch `soc-2020-fluid-tools` are included in this patch.
**Major changes:**
=== Viewport Display ===
- //Raw voxel display// or //closest (nearest-neighbor)// interpolation for displaying the underlying voxel data of the simulation grids more clearly.
- An option to display //gridlines// when the slicing method is //single//.
==== Grid Display ====
- Visualization for flags, pressure and level-set representation grids with a fixed color coding based on Manta GUI.
==== Vector Display ====
- //**M**arker **A**nd **C**ell// grid visualization options for vector grids like velocity or external forces.
- Made vector display options available for external forces.
==== Coloring options for //gridlines// ====
- Range highlighting and cell filtering options for displaying the simulation grid data more precisely.
- Color gridlines with flags.
- Also, made slicing and interpolation options available for Volume Object.
Reviewed By: JacquesLucke, sebbas
Differential Revision: https://developer.blender.org/D8705
`ctrl-F` to start the search is obviously necessary, but the clear
operator, `alt-F` requires some of explanation. First, it maps nicely
to the paradigm of "key to set, alt-key to clear," which makes it
unobtrusive. Second, it can be a quicker way to clear the search than
moving the mouse to the top. Finally, in the future, it could a reset
the panels to their expansion before the search started.
Differential Revision: https://developer.blender.org/D8857
This adds a search bar to the properties editor. The full search for
every tab isn't included in this patch, but the interaction with
panels, searching behavior, UI, region level, and DNA changes are
included here.
The block-level search works by iterating over the block's button
groups and checking whether they match the search. If they do, they
are tagged with a flag, and the block's panel is tagged too. For
every update (text edit), the panel's expansion is set to whether
the panel has a result or not. The search also checks for matching
strings inside enums and in panel labels.
One complication to this that isn't immediately apparent is that
closed panel's subpanels have to be searched too. This adds some
complexity to the area-level panel layout code.
Possible Future Improvements:
- Use the new fuzzy search in BLI
- Reset panels to their expansion before the search started if
the user escape out of the text box.
- Open all child panels of a panel with expansion.
Differential Revision: https://developer.blender.org/D8856
Those constraints are added when using AutoIK (or targetless IK as
well).
While not strictly incorrect, these kinds of constraints were not
showing in the UI prior to rBeaa44afe703e (and I think they should not).
ref T80437
Maniphest Tasks: T80437
Differential Revision: https://developer.blender.org/D8895
Caused by rBe490dc4346db: UI: Skip unnecessary cursor setting
Above commit returned early if the cursor was already set, but did this
before visibility was regained, now return (still early) after setting
visibility.
Reviewed by @Severin in T80333
We will use a highlight on panel headers to convey that they have
a search match, so this commit initializes the theme color for the
properties editor.
Differential Revision: https://developer.blender.org/D8854
The context path "breadcrumbs" used to be in a panel in 2.79.
Although they look a bit better in the header, there isn't enough
space for them with the property search field in the header as well.
Maybe there will be another solution in the long term to fit both
the search field and this panel in the header, but for now, this
commit moves these labels back to a header-less panel.
Differential Revision: https://developer.blender.org/D8853
This is needed for property search (upcoming patch D8856) where a
buttons labels are often separate buttons, but we need to highlight
the label to show that a property is a search result. This is
especially important in "property split" layouts where the label is
almost always another button in a separate column.
The button groups here are basically a flattened view of the buttons
in the layout tree. Every function that adds a new set of buttons
creates a new button group, and the new buttons are automatically
added to the most recent group. Then, each group is searched
separately in the property search phase. It's important that every
function adding a new button set calls layout_root_new_button_group.
Note that this won't be disabled when property search isn't active.
It may be useful for other things in the future, and trying to pass
that information to layout functions didn't feel worth it to me.
Differential Revision: https://developer.blender.org/D8783
The UV/Image editor was doing interpolation including over the alpha
value what makes will render incorrectly when interpolating between pure
emissive colors and pre multiplied colors. This change disabled the
interpolation.
Add an option to only draw icons for operator_enum menus. This is used
for drawing inline icon buttons in the outliner context menu for
collection color tagging.
Part of T77408
Differential Revision: https://developer.blender.org/D8880
This adds support for drawing icon buttons as a row in menus. This is
needed for drawing collection color tagging icons in the outliner
context menu in T77777.
Part of T77408
Differential Revision: https://developer.blender.org/D8317
Commit 009971ba7a changed it so Cycles creates a separate
Embree device for each Cycles device, but missed the multi-device case. A multi-device with
Embree BVH can occur when CPU rendering is used with OptiX denoising and BVH creation then
failed to get a valid pointer to the Embree device, which crashed. This fixes that by providing the
correct device pointer in the multi-device case as well.
Relax limits of FCurve Bézier handles during evaluation. FCurve handles
can be scaled down to avoid the curve looping backward in time. This
scaling was done correctly but over-carefully, posing unnecessary
limitations on the possible slope of FCurves. This commit changes the
scaling approach such that the FCurve can become near-vertical.
Bump Blender's subversion from 291.0.1 to 291.0.2 to ensure that older
animation files are correctly updated.
Reviewed By: sybren
Differential Revision: https://developer.blender.org/D8752
A bit of versioning code was not part of the "Versioning code until next
subversion bump goes here" block, making the diff of an upcoming bump
harder to read. This commit just moves the code to the appropriate spot.
No functional changes.
Copy correct time to temporary strips used for evaluation of NLA
Transition strips.
When a transition is placed next to a meta strip, the transition would
use the wrong strip time, using the time of the meta strip instead of
its own.
Reviewed by: sybren
Differential Revision: https://developer.blender.org/D8287
There are some areas that don't handle pure emissive colors well. For
example erasing alpha using 2d or 3d painting. Or blurring an image in
the compositor.
This patch makes the rendering of pure emissive colors optional. In the
side panel of the Image editor it can still be enabled when needed.
There currently isn't a better place to store it as it is related on how
the image (or a layer of the image) is created. A future design needs to
make sure that the full workflow is supported.
The ED_region_panels_layout_ex function was quite long, so separating
some pieces of it can make it easier to understand and more modular.
Additionally, the parts that were split off can be reused for future
property search code.
On MacOS + Intel Iris Graphics 6100 (may affect other config too), the
texelFetch operation bypass the base mip setting of the texture object.
Using textureLod with lod = 0.0 ensure the lowest (after clamping) mip will
always be selected.
Also disable the texture filtering for this sampler to avoid unecessary
fetches.
This should fix T78653 Blender 2.83 broken Depth of Field in Viewport
Remove the condition preventing orientation from being obtained out of
the 3d View context.
Also pass the `ob` and `obedit` arguments obtained from the caller.
This commit adds a label on top of the bake operator in modular bake mode. This way users will immediately see if their current settings will allow them to bake noise, meshes or particles after baking the simulation base.
It was possible to drop `nasm` from the list of required macOS dependencies. However, `pkg-config` had to be added - it was required before but probably no one noticed this.
In order to build `external_clang` successfully, `external_xml2` had to be added to the clang dependencies (`c-index-test` was failing).
Note: On Linux `make deps` still seems to require `nasm` installed via apt.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8870
Debug groups makes it easier to view from where an error comes from.
The backend can also implement its own callback to make it easier to
follow the API call structure in frame debuggers.
The instance offset should be applied before scaling.
This way the scaling is done from the "collection origin".
Reviewers: zeddb, brecht
Differential Revision: https://developer.blender.org/D8889
The bounding box is not updated in the original object when the function is called using evaluated object and keeps wrong while the object is not edited or the file saved.
Reviewed By: mont29
Differential Revision: https://developer.blender.org/D8565
Notes: Minor changes done in the patch following review comments.
During undo/redo read code is expected to clear the `OB_MODE_EDIT`
bitflag of `Object.mode`, for some reasons.
This was not done anymore for re-used Objects, we need to add a special
handling case for that too.
Should be backported to 2.90 and 2.83 (will probably not be straight
forward for the latter).
Oversight in 411c5238a2 ignored sticky selection.
Use 'uvedit_edge_select_set_with_sticky' to make sure
sticky options are respected.
Also skip checking the existing selection since that only checks the
current UV, not all connected UV's which is needed for sticky selection.
The extra checks to avoid updating UV's isn't such an advantage as
only meshed in the selected region are tagged for updating.
Write custom properties (aka ID properties) to Alembic, to the
`.userProperties` compound property.
Manifest Task: https://developer.blender.org/T50725
Scalar properties (so single-value/non-array properties) are written as
single-element array properties to Alembic. This is also what's done by
Houdini and Maya exporters, so it seems to be the standard way of doing
things. It also simplifies the implementation.
Two-dimensional arrays are flattened by concatenating all the numbers
into a single array. This is because ID properties have a limited type
system. This means that a 3x3 "matrix" could just as well be a list of
three 3D vectors.
Alembic has two container properties to store custom data:
- `.userProperties`, which is meant for properties that aren't
necessarily understood by other software packages, and
- `.arbGeomParams`, which can contain the same kind of data as
`.userProperties`, but can also specify that these vary per face of a
mesh. This property is mostly intended for renderers.
Most industry packages write their custom data to `.arbGeomParams`.
However, given their goals I feel that `.userProperties` is the more
appropriate one for Blender's ID Properties.
The code is a bit more involved than I would have liked. An
`ABCAbstractWriter` has a `uniqueptr` to its `CustomPropertiesExporter`,
but the `CustomPropertiesExporter` also has a pointer back to its owning
`ABCAbstractWriter`. It's the latter pointer that I'm not too happy
with, but it has a reason. Getting the aforementioned `.userProperties`
from the Alembic library will automatically create it if it doesn't
exist already. If it's not used to actually add custom properties to, it
will crash the Alembic CLI tools (and maybe others too). This is what
the pointer back to the `ABCAbstractWriter` is used for: to get the
`.userProperties` at the last moment, when it's 100% sure at least one
custom property will be written.
Differential Revision: https://developer.blender.org/D8869
Reviewed by: sergey, dbystedt
Especially with new undo/redo it is even less recommended to perform
complex operations in those callbacks, they should remain as fast and
localized as possible.
Also updated the section about undo/redo a bit.
RNA diffing code was not dealing properly valid NULL PointerRNA (like
the empty texture slots of a ParticleSettings e.g., which were cause of
crash in that report).
To be backported to 2.90 and 2.83.
Running `bpy.ops.script.reload()` from Python was crashing
since the operator being called was it's self freed.
Change the reload operator to defer execution - as supporting
re-registration during execution is quite involved for a corner-case.
Tiled texture uses different texture structure than normal textures.
Normally we add dummy textures and use them, but I found it cleaner to
have 2 shaders and use the correct shader.
Hiding a collection should hide all children objects even when we are in Local
view with one of them.
Note from reviewer: We are doing this already for local collections. So
may as well do it when hiding the collections for the entire view layer.
Developer details: In function "BKE_object_is_visible_in_viewport" object flag
BASE_VISIBLE_VIEWLAYER wasn't being checked when we were in Local view,
It's now changed so that it's checked even if we are in Local view.
And this function was called by some viewport draw functions to check if it
should draw an object or not.
Maniphest Tasks: T72584
Differential Revision: https://developer.blender.org/D7894
This was caused by a left over DRWPass->state modification
that made the subsequent samples redraw without Blending enabled.
This led to incorrect blending.
The fix is to use the new API for pass instancing.
Also added code so that exact solver does the whole collection at once.
This patch allows users to use a collection (as an alternative to Object)
for the boolean modifier operand, and therefore get rid of a long modifier stack.
Entering edit-mode after creating shape keys on a blank mesh would crash.
Regression in 9b9f84b317 which prevented initializing empty
shape keys when there is no shape key offset data available.
Printing an evaluated view layer would show:
Evaluated Scene 'Scene'
- Now __repr__ uses the __str__ fallback for evaluated data,
as done in other situations where we can't create a string that
would evaluate to the data.
- __str__ now shows when the data is evaluated.
- __str__ always includes the memory address
(which was previously only shown for structs without a name).
This was caused by a missing state apply.
We force the GPUState to be set after the callbacks to avoid
desync between our state tracker and the real gl state.
This fixes some issues but a better general fix for all BGL would
be better.
This fix T80297 2.91 texture alpha is not transparent
This wraps the functionality used to speedup EEVEE volumetrics.
This touches the rendering code of EEVEE as it should fix a mis-usage of
the GL barrier. The barrier changed type and location, removing an
unused barrier.
Bone.children_recursive would return edit-bones when in edit-mode
irrespective of the type of the bone.
Check the type of self instead of the existence of edit-bones.
This commit removes the custom callback that's currently used to set the
property editor's pinned data, replacing it with an operator. This means
"pin" button doesn't have to be defined in C.
Differential Revision: https://developer.blender.org/D8376
In a recent update to the fluids modifier (rB03c2439d96e8), I introduced a flush call that sets all grids to NULL if the frame is outside of the allowed frame range. This way, the texture creation function must also check if the data grid is NULL before trying to create a texture.
Reviewed By: fclem
Differential Revision: https://developer.blender.org/D8872
Set the inverse matrix when the Child Of constraint is created. This
prevents the bone/object from jumping away when the constraint is added,
improving usability.
Reviewed by: sybren
Differential Revision: https://developer.blender.org/D8851
Running add-objects operator without an active screen would crash
as CTX_data_edit_object would return NULL in this case.
This could happen when adding objects from frame-change handlers for e.g.
In the case of adding new objects there is no need for a context lookup,
pass this directly to ED_object_editmode_exit_ex.
Add a function HairKey.co_object_set() which can be used to modify hair
keys positions in a way that hair strands respects the new coordinates,
without doing extra trickery with edit mode toggle. The naming comes
from existing HairKey.co_object().
The usage is a bit tricky:
- The function is to be called on the hair key from an original object.
- The arguments are to be passed from evaluated object.
Basically:
hair_key_orig.co_object_set(object_eval, modifier_eval, particle_eval)
This is required because coordinate space conversion needs to know
evaluated state of the hair system in order to access hair space
matrix. It is a deliberate choice to not pass dependency graph and not
to do lookup of evaluated ID on every call of co_object_set() to keep
performance on a manageable levels.
One requirement is that the hair strands are to be calculated once, by,
for example, toggling particle edit mode. Without this there will be no
`particles` to access hair keys from in the original object.
Such functionality is something what got lost during 2.80 development
and is something what is needed here at the studio.
Below is the file with more complete example: convert strands from mesh
edges to actual particle system hair strands. The viewport refresh is
sometimes missing, and this does not seem to be related to this specific
change.
{F8858104}
Differential Revision: https://developer.blender.org/D8849
The old image editor has an option to enable the smooth wire drawing.
This option was stored per editor and disabled by default. This patch
connects the smooth wires in the UV/Image editor to `User Prefereces ->
Viewport -> Quality -> Smooth Wire [] Overlay`.
The old option is left in place and will be removed when the old image
editor drawing code will be removed before BCon 3.
This patch reverses use draw manager for image editor the
experimental feature. Now the new drawing is enabled by default.
Inside the experimental tab in the user preferences there is now an
option to revert back to the old drawing method.
Using this option we can easilly check if all drawing features have been
migrated over. The plan is to remove the legacy drawing before BCon 3.
This project moves the current UV/Image editor drawing to the draw manager.
Why would we do this:
**Performance**:
Current implementation would draw each texel per time. Multiple texels could be
drawn per pixel what would overwrite the previous result. You can notice this
when working with large textures. Repeat image drawing made this visible by
drawing for a small period of time and stop drawing the rest. Now the rendering
is fast and all repeated images are drawn.
**Alpha drawing**:
Current implementation would draw directly in display space. Giving incorrect
results when displaying alpha transparent images.
This addresses {T52680}, {T74709}, {T79518}
The image editor now can show emission only colors. See {D8234} for
examples.
**Current Limitations**
Using images that are larger than supported by your GPU are resized (eg larger
than 16000x16000 are resized to 8k). This leaves some blurring artifacts. It is
a low priority to add support back of displaying individual pixels of huge
images. There is a design task {T80113} with more detail.
**Implementation overview**
Introduced an Image Engine in the draw module. this engine is responsible for
drawing the texture in the main area of the UV/Image editor. The overlay engine
has a edit_uv overlay which is responsible to draw the UV's, shadows and
overlays specifically for the UV Image editor. The background + checker pattern
is drawn by the overlay_background.
The patch will allow us to share overlays between the 3d viewport and UV/Image
editor more easily. In most cases we just need to switch the `pos` with the `u`
attribute in the vertex shader.
The project can be activated in the user preferences as experimental features.
In a later commit this will be reversed.
Reviewed By: Clément Foucault
Differential Revision: https://developer.blender.org/D8234
This should improve the issue with Scrape accumulation in concave
surfaces. When the strength of the brush is higher, the area radius is
also bigger, so the scrape plane is more stable preventing it from
accumulating displacement in the same area.
The Scrape/Fill default presets are also updated to include this
functionality.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8821
When using the left and right arrows of the number button, the button
would seem to do nothing. That is because it modified a value smaller
than what is displayed. Show more precision so the button doesn't appear
broken.
There were some problems when converted several objects at the same time, especially with the material conversion. The problems were more visible when bake an animation with several objects at the same time.
* Now the layer name include the object name.
* Reorganize how the materials are generated including object name.
* Fix color not converted to sRGB.
* Avoid triangles when generate the stroke. This fix a draw manager issue and also add more geometry to use later.
* Code cleanup.
* Don't link against Mesa OpenGL software emulation libraries from the
lib folder, they are not intended to be used for building, only install.
* Don't use static OpenMP library by default, it's usually not built to
work in shared libraries.
* Disable jemalloc on all platforms, the reason it fails is not specific
to Linux.
Differential Revision: https://developer.blender.org/D8827
When using Windows Terminal the same control codes Linux
uses to colorize the text can be used. WT can be detected
by looking at the WT_SESSION environment variable.
Differential Revision: https://developer.blender.org/D8848
Reviewed by: campbellbarton
`use_color` was defaulted to 'on' and only turned off when it was
decided the terminal didn't support this. This was the wrong
way to go about things, since that test only ran for Linux/Mac other
operating systems would get color formatting regardless if they
supported it.
The new default is `off` until the platform specific code inside
`CLG_ctx_output_set` turns it on after running some sanity checks.
This fix was part of the review of D8848 but deserved its on commit
Reviewed By: campbellbarton
When creating the boundaries for the symmetry passess, it could happen
that a symmetrical vertex is not found inside the radius of the brush.
In that case, the function to initialize a new boundary was called with
index -1 (BOUNDARY_VERTEX_NONE), so this function should also check
that the initial vertex that is going to be used to find the boundary is
a valid index.
Reviewed By: sergey
Maniphest Tasks: T79754
Differential Revision: https://developer.blender.org/D8860
When opening the outliner context menu with multiple data types
selected, the menu did not show operators related to the target of the
cursor. For example, if a modifier and object are selected, a right
click on the modifier does not show the modifier options, rather it
shows the object context menu.
Now the data type of the right-click element is used to determine the
context menu to draw. For this to work properly the active element is
now set on right click.
Part of T77408
Manifest Task: https://developer.blender.org/T77770
Differential Revision: https://developer.blender.org/D8647
The outliner context menu has options to enter and exit edit mode, but
they only show in edit mode, and they don't work. This removes the
broken entries and related code.
Part of T77408
Differential Revision: https://developer.blender.org/D8641
Add a column of icons in the left gutter of the outliner for controlling
the interaction modes of objects. When an object is in a mode other than
object mode, the mode icon will draw to the left of that object. Any
other objects that are valid to be added or swapped into the mode are
drawn with a dot to the left of the object.
Clicking the dot to the left of an object will swap that object with the
current active object in the interaction mode. For edit and pose modes,
ctrl clicking the dot will add that object to the current mode.
Clicking the mode icon next to the active object removes it and all
other objects from the current mode.
The behavior is nearly identical to the previous edit/pose mode toggling
by selecting the mesh and armature datablocks, with additional support
for all interaction modes.
Currently two undo steps are pushed to prevent an assert.
Part of T77408
Manifest Task: https://developer.blender.org/T68498
Differential Revision: https://developer.blender.org/D8641
The Eraser Point mode was erasing too much points.
This eraser existed before the new Soft erasers were created and now it was working wrongly. The eraser point mode must remove the points below the eraser cursor at the first contact. For more soft erasers, the new Soft modes can be used.
This allows subsequent redraw to work just like before.
However a better safeguard system against setting the uniforms in the
wrong shader would be nice to have.
This prevents Blender from crashing when switching
to rendered view and the grid_data is NULL in the
lightcache (due to possible corruption), by switching
to the fallback lightcache. The fix extends the solution
for possible corruption in the cube_data as well.
Fix T79737
Reviewed By: fclem
Differential Revision: https://developer.blender.org/D8835
This makes it possible to disable all the extensions when forcing
workarounds.
Also it will allow future options to selectively disable each extension
to know which one is buggy.
This is to improve debugging on older hardware that may not support
4.3 debug capabilities (like Macs).
This avoids sprinkling glGetErrors manually. This might still be needed
to find the root cause since not all functions are covered.
This overrides the functions pointers that GLEW have already init.
This is only enabled if using --debug-gpu option and the debug extension
are not available.
This also cleanup the usage of GLContext::debug_layer_support and use
wrapper to set object labels.
Modular caches for noise, particles and meshes require that additional data is baked (i.e. is resumable option). With this commit users will be explicitly asked to enable this option before being able to bake noise, particles or meshes.
This commit renames the Sequence Display Mode to Video Sequencer. THis
matches the editor name in the editor list.
This commit also adjusts the tooltip because the data show in this
dispay mode is not strickly speaking "data-blocks".
Ref D7716
This commit also changes the U shortcut to open the unwrap menu instead
of the Unwrap operator. The unwrap operator can now be accessed by
pressing U twice.
Note, these operators use the 3D Viewports selection and not the UV
Editor selection. In the future the operators should unwrap based on the
selection within that editor.
Fixes T80600
Differential Revision: https://developer.blender.org/D8834
Caused by own rB6dc7266cf1f4.
When overriding context for constraint operators (such as in constraint
tests), it could happen that context "active_pose_bone" is set, but
"pose_bone" isnt. Now check for both in ED_object_pose_constraint_list.
Load all grids when the grid display option is turned on. Otherwise, the grid from the viewport display options might not have been loaded from the cache.
It seems that using a vertex shader using both `gl_InstanceID` and
`gl_VertexID` is causing some issues on MacOS + Intel Iris.
Regression introduced by rB052538edc1fba109d3427471047611888ed13bea
The drawing of popus should be done with the window's default viewport
and scissor. Previous functions may change these, so they should be
explicitly reset.
There is a list of attributes, along with operators to add and remove
attributes. For adding, there are a few standard attributes that can be
added quickly, as well as a popup to create a custom attribute.
Ref T76659
Differential Revision: https://developer.blender.org/D8636
Instead of custom data layer with special types, using general Vector and Float
attributes.
Ref T76659
Differential Revision: https://developer.blender.org/D8635
This puts all generic float/int/vector/color/string geometry attributes in a new
.attributes property. For meshes it provides a more general API for existing
attributes, for point clouds attributes will be used as an essential part of
particle nodes.
This patch was implemented by @lichtwerk, with further changes by me. It's
still a work in progress, but posting here to show what is going on and for
early feedback.
Ref T76659
Differential Revision: https://developer.blender.org/D8200
The fix involves going over ALL the possible combinations of viewlayers
and viewports and re-sync.
I tested this with multiple windows, multiple scenes and multiple
viewlayers.
Since (for now?) the operation of syncing the local layer collections is
not too expensive, this is not so bad. In theory we could improve this
by checking if the collection the object was moved to and from is in the
scene before iterating over it. I don't think it is worthy though.
Thanks Arun Parolikkal for the initial attempt on D8342.
Final patch reviewed by Brecht Van Lommel.
This makes it error-proof to disable/enable experimental features
during the release cycles. Since it is handled by CMake it should always
work reliably now (not depending on someone turning this on and off).
Reviewed by Sergey Sharybin.
not active
Caused by {rB608d9b5aa1f1}
Prior to rB608d9b5aa1f1, the constraint was gotten using **context**
[CTX_data_pointer_get_type(C, "constraint", &RNA_Constraint) -- which is
valid for bones on hidden layers].
After rB608d9b5aa1f1, the constraint is found (or isnt) using
`edit_constraint_property_get` [this is **not** valid for bones on
hidden layers because internally `BKE_pose_channel_active` checks if the
bone is on an active layer].
Some observations:
- Every operator using `edit_constraint_property_get` doesnt work for
bones on inactive layers [delete, moveup, movedown, move to index (drag
n drop nowadays)]
-- moveup, movedown, move to index check if they could find a constraint
beforehand though (dont crash)
-- delete crashes (doesnt check if a constraint could actually be found)
- Every operator using `edit_constraint_property_get` for constraint
data doesnt work for bones on inactive layers [stretchto_reset,
limitdistance_reset, childof_set_inverse, ...]
-- these all check if they could find a constraint beforehand though
(dont crash)
This is because the poll function is using **context** to get the
constraint, the operators themselves use
**edit_constraint_property_get** which leads to inconsistent/unexpected
results.
Possible solutions were:
- [1] let the delete operator just work with the context constraint
again (like prior to rB608d9b5aa1f1) -- allows for deleting constraints
on bones in inactive layers
- [2] check if we could get a constraint -- prevents the crash, but does
**not** allow for deleting constraints on bones in inactive layers
- [3] make the poll `edit_constraint_poll_generic` be as strict as the
operators -- dont use **context** to get the constraint, but something
like **edit_constraint_property_get**
- [4] make the operators be more graceful and let them act on bones on
hidden layers -- let **edit_constraint_property_get** actually use the
same **context**
This patch implements [4], so poll an doperators are now in sync.
- prevents reported crash
- also enables operators for bone constraints on hidden layers
- also enables drag and drop reordering of constraints on hidden layers
This might be a candidate for 2.90.1? (if it is, take care to include
prior "Refactor getting constraints" refactoring commit)
Note: Adding constraints also doesnt work for bones on inactive layers
[that was the case in 2.79 as well -- it is also using
`BKE_pose_channel_active`]
Maniphest Tasks: T80464
Differential Revision: https://developer.blender.org/D8805
This is the refactoring part of D8805 (should be no functional changes).
- exposes pose-related part of former 'get_constraints()' from
interface_templates.c to new ED_object_pose_constraint_list
- rename ED_object_constraint_list_from_context -->
ED_object_constraint_active_list
Also clarify comments on both of these.
ref T80464
ref https://developer.blender.org/D8805
Blender does string based searching in many different places.
This patch updates four of these places to use the new string
search api in `BLI_string_search.h`.
In the future we probably want to update the other searches as well.
Reviewers: Severin, pablovazquez
Differential Revision: https://developer.blender.org/D8825
This adds a generic string search library in `BLI_string_search.h`.
The library has a simple to use C api that allows it's users to
filter and sort a set of possible search results based on some
query string.
Reviewers: Severin
Differential Revision: https://developer.blender.org/D8825
-1 is a valid value to pass as RNA float precision. We let the UI code
figure out an appropriate precision then. So don't fail the assert if
-1 is passed.
The point cache code needs a non NULL rbw pointer.
This could have been avoided if there was a sanity check in the convert
function, so added a check there as well.
On systems with 512kb stack this happened at around 13k points.
This happened at times with grease-pencil, although callers that
frequently use complex polygons should be using BLI_polyfill_calc_arena.
This reverts commit 637a5c964a.
I commited the previous commit because I wasn't building with openvdb.
Compiling with openvdb fix the clang-tidy errror.
This reverts commit f14d24729f.
I commited the previous commit because I wasn't building with openvdb.
Compiling with openvdb fix the clang-tidy errror.
This removes from the UI all tools with missing icons and hides them
under a "Tools with missing icons" experimental option.
We agree on not making available by default tools in master without icons.
Having this experimental flag will allow to commit new tools as soon as the
technical design and implementation is finished so development can
continue, without adding broken icons to the UI.
Reviewed By: Severin
Differential Revision: https://developer.blender.org/D8831
Add support for object data instancing. This is used when the objects
are instances, for example when duplicated by a particle system, or
instanced by the duplication system (collection-duplicating empties,
vertex/face duplis, etc.)
Since Alembic already deduplicates data, this doesn't make the resulting
Alembic files any smaller. They will be faster to write, though, when
there is a lot of instanced geometry, as the deduplication system won't
have to do any comparisons.
This instancing support is still limited, in the sense that only object
data is instanced and all transforms are still written explicitly. A
future improvement could be to support instancing entire collection
hierarchies.
Blender's Alembic importer has no understanding of these Alembic
instances yet, and will thus happily duplicate the data on import.
The USD Alembic plugin seems to have problems understanding the
instancing. There might also be other software with similar issues.
Because of this, instancing can be turned off in the exporter (it's on
by default).
* Match menu items with Add Menu (order and naming e.g. Font -> Text)
* Use Icons
* Remove ellipsis from the name (policy is to use `...` only when triggering a window/popup)
No functional changes.
Thanks @HooglyBoogly for the help!
Split `ABCHierarchyIterator::get_alembic_parent()` into two functions:
- For a given export path, find the Alembic object
- Ensure that that object is usable as parent object (Alembic uses a
specific 'top' object as parent to indicate "no parent").
The new function is `public` as it will be used in an upcoming feature,
and is required to be public then.
No functional changes.
Split the `ABCHierarchyIterator::create_data_writer()` function into two
functions. This is to prepare for the creation of writers not just by
object type, but also by goal, for example writers that reference other
Alembic data instead of writing their own (i.e. instancing).
No functional changes.
The draw manager test case initialized ghost, gpu and draw manager. This
change splits the base test case to GPU specific and draw manager
specific test case.
The GPU test base test case will be used for low level GPU tests.
Memory leak is introduced as test cases reinitializes the GPU stack.
Added a call to GPU_backend_exit to fix this.
In GPU_backend_exit the GPU backend was destroyed but the pointer wasn't
reset for reuse. This patch also clears the pointer to be reused.
Extract the mesh instancing code from the mesh writing function into a
generic 'mark as instance' function on the abstract USD writer. This will
help in supporting non-mesh instances.
No functional changes.
Refactor `ED_object_parent_set`:
- Mark parameters `ob` and `par` as `const` so that it's clear the
function doesn't assign any other value to them.
- Rename `pararm` to `is_armature_parent`; I mis-read it as `param` all
the time, and it was very confusing.
- Replace repeated `if-else` statements with `switch` statements.
- Reorder preconditions to have some simple checks first.
- Flip condition on a huge `if`-statement to return early and unindent
the remainder of the function.
This function still requires splitting up into smaller functions, but at
least this is a step forward.
No functional changes.
This patch supports instantiating object data on append/link,
reported as a bug T58304.
This is an option, available when linking/appending,
similar to the existing "Instance Collections" option.
Reviewed by @sybren
Ref D8792
Move arguments to BLO_library_link_{begin/named_part/end} into
a single parameter struct, to ensure arguments always match.
This allows is to skip tagging ID's LIB_TAG_DOIT when it's not needed,
previously it was always cleared just in case it was needed.
This also makes it possible to remove BLO_library_link_named_part_ex
which was only used when tagging was needed.
This makes the GPUContext follow the same naming convention as the rest
of the module.
Also add a static getter for extra bonus style (no need for casts):
- Context::get()
- GLContext::get()
- Use the syntactic wrap/unwrap method to make code more readable.
- Update comment about hidden struct behind opaque types.
- Cleanup GPUDrawList type.
This is part of the Vulkan task T68990.
This introduce a new GLQueryPool for managing queries in an
implementation agnostic manner.
This modify the GPU selection query to use this new object.
This also make use of blender::Vector for better code quality.
No real functionnal change.
This property was inadvertently removed from the modifier's panel and
it wasn't caught in time for the release of 2.90. Thanks to the user
"VermossomreV" for bringing this to my attention.
Differential Revision: https://developer.blender.org/D8790
Based on the original patch by Vaishnav S (@padthai), this adds
support for temperature units. Initially supported units are Celsius,
Kelvin, and Fahrenheit.
The units aren't used anywhere with this commit. Those changes should
happen in separate patches by adding PROP_TEMPERATURE to RNA property
definitions. But it should be ensured that the various solvers and
simulations actually properly use real units.
The complexity of some of the changes comes from the fact that these
units have offsets from each other as well as coefficients. This also
makes the implementation in the current unit system troublesome.
For example, entering 0C evaluates correctly to 273K, but 0C + 0C
doubles that result, because each unit value is evaluated separately.
This is quite hard to solve in the general case with Blender's current
unit system, though, so it is not handled in this commit.
Differential Revision: https://developer.blender.org/D4401
The current playback popover has some issues:
- Using labels instead of headers is inconsistent with
the rest of the interface
- Incomplete context and description for some properties
- Ugly large spacing
This commit fixes these problems by using headers.
Differential Revision: https://developer.blender.org/D8434
Every time CMake is re-run, Xcode shows a popup asking if
user wants to manage schemes automatically or manually.
Building Blender wiki page recommends managing schemes automatically.
This change sets the default behavior to "automatically" and generates
the .xcscheme files while CMake is running, instead of hogging Xcode
later on. With tests enabled, the number of schemes is 203.
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D8820
Similar to {rB0a5f7061369d53b4eac55362ad2}
but also for Xcode and Ninja multi-config.
This silences 44 pairs of warnings like:
/bin/rm -f build_full/bin/tests/BLI_ghash_performance_test
"build_full/CMakeScripts/XCODE_DEPEND_HELPER.make:42: warning:
ignoring old commands for target
`build_full/bin/tests/BLI_ghash_performance_test'"
/bin/rm -f build_full/bin/tests/BLI_ghash_performance_test
"build_full/CMakeScripts/XCODE_DEPEND_HELPER.make:3523: warning:
overriding commands for target
`build_full/bin/tests/BLI_ghash_performance_test'"
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D8815
This change allows macOS developers to use
`WITH_COMPILER_ASAN` with every generator.
`CMAKE_C_IMPLICIT_LINK_DIRECTORIES` on macOS points to
`Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.14.sdk/usr/lib`
which is not where the Sanitizer libraries are.
To link the library, rpath could be used but that seems complex,
so linker flags are passed as the documentation says. [1]
If users have `ASAN_OPTIONS=detect_leaks=1` in their environment
variables, it should be removed to avoid a feature-unsupported error
while compiling.
[1]: http://clang.llvm.org/docs/AddressSanitizer.html#usage
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D8817
With this option, self-intersections in either or both operands
will be handled properly (if both sides are piecewise winding
number constant, and maybe some other cases too).
In the Boolean tool, this flag was there already but the code
forced a unary operation in that case; this commit corrects it
to make a binary operation. This flag makes the code slower, which
is why it is an option and not an always-on thing.
This is part of the Vulkan task T68990.
This commits changes a few things:
- Rename extensions to capabilities (but left the file name untouched).
- Cubemap mip render workaround detection is rewritten using gl
commands to avoid using the GPU API before initialization.
- Put all the capabilities that are only relevant for the GL backend
inside GLContext as static variables.
- Cleanup the names of the limit variables.
- Separate all GL related workaround search inside the GL module.
Part of the vulkan implementation T68990.
Pretty straight forward. Just move the GL code inside the GLBackend and
make the GPUPlatformGlobal a class object.
Refactor the operator exec function into a few smaller functions. The exec
function was mixing up vertex-parent and non-vertex-parent code, including
incorrect comments.
No functional changes.
This implements Box Trim as a boolean based trimming too gesture in sculpt
mode. This is the intended way to remove parts of the sculpt instead of
using box mask and mask slice. It also creates new face sets for the new
faces created after the boolean operation.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8766
This was introduced in the first commit of the cloth brush. In order to
support the cloth brush deformations with subsurf modifiers, the sculpt
API was changed to return the deformed coordinates from the PBVH instead
of the mesh coordinates. When using grab active vertex and rendering the
original mesh wireframe it should render the undeformed mesh coordinates
when available.
Reviewed By: sergey
Maniphest Tasks: T79914
Differential Revision: https://developer.blender.org/D8630
Now when a mesh is partially hidden using Face Sets, the open boundary
the hidden geometry produces is also detected by the brush.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8819
This should be using the mesh_cd_ldata_get_from_mesh function in
order to get ldata from BMesh in edit mode.
Reviewed By: sergey
Maniphest Tasks: T78225
Differential Revision: https://developer.blender.org/D8818
The "Grab Active Vertex" in sculpt mode highlights the vertex and
the neighbor vertices. This was working wrong in the case when mesh
has multires modifier at sculpt level 0 and has shape keys.
The issue was caused by the wrong crazy space calculation, which was
ignoring subdivision level. This is an oversight from the initial
implementation: the modifier has no effect if the subdivision level
is 0.
Part of the Vulkan port T68990
This makes a few changes in how the data is being handled by the
backend to allow more flexibility in the future.
The overall code logic is left unchanged.
Looks like this has just not been implemented before.
Use the name matching method used in other Dope Sheet UI modes.
Maniphest Tasks: T80531
Differential Revision: https://developer.blender.org/D8824
This is part of the Vulkan backend task T68990.
There is no real change, only making some code re-organisation.
This also make the IndexBuf completely abstract from outside the
GPU module.
This allows orphan object data for example (meshes, curves, etc)
to be dropped into the 3D View from the outliner,
creating a new object instance.
Previously the only way to do this was to add the same type of object
then swap it's data through the ID selector drop-down.
Well, this only fixes the example if one uses the new Exact mode,
but since that is available, seems fair to call this fixed.
Since these were not closed-volume operands, the Exact mode needed
some adjustment to the threshold used for "inside-outside" tests
for the case of deciding if the cutter is inside the other shape
for a Difference.
The issue was that the screen geometry calculations scaled down the
areas, but the header would become too small. So it would be upscaled
again towards the lower screen edge (where the status-bar is at). Now we
do another pass over the geometry calculations, until all areas fit into
the screen, or until some rather arbitrary maximum is reached.
This fixes the issue for common cases. Extreme cases, with many
vertically stashed editors and a too small window to show all, are still
not too well supported. Blender keeps working but the status-bar is
still overlapped. We could deal with this better but what's there now
should be good enough.
These buffer detailed infos are not needed unless going deep into
perf. profiling, in which case you can still disable this compile option.
This makes user report log much more readable.
We cannot change the texture bind point since the interface count
4/5 textures to bind. Changing the uniform to avoid one bind
make the system thinks one texture is missing.
Avoid this by creating a dummy texture and binding it to the empty
slot.
This is to modernize the API:
- Add meaningful name to all textures (except DRW textures).
- Remove unused err_out argument: only used for offscreen python.
- Add mipmap count to creation functions for future changes.
- Clarify the data usage in creation functions.
This is a cleanup commit, there is no functional change.
# Conflicts:
# source/blender/gpu/GPU_texture.h
This is to have better error detection in debug builds.
This is not a replacement for a full check like in renderdoc but it
might catch some issues early on.
This is a massive cleanup needed for vulkan support T68990. It provides:
- More meaningful enums with conversion functions.
- Less hacky supports of arrays and cubemaps (all considered layered).
- More inline with the stateless design of vulkan and modern GL.
- Methods Fallbacks are using framebuffer functions that are wrapped
instead of implementing inside the texture module.
What is not in there:
- API change.
- Samplers support (breaks a few effects).
# Conflicts:
# source/blender/gpu/GPU_texture.h
This is in preparation of using it to clear single texture.
Also includes minor cleanups about not using tex target in
assert and adding enum operators.
This is going to be unecessary after the GPU opengl texture backend refactor.
For now add a save/restore mechanism to leave the state untouched.
Also remove some calls where the caller would bind to particular binding
point and set the shader uniform.
The code to detect non-trivial coplanar intersection sometimes
falsely said there wasn't one. This caused some coplanar intersections
to be missed. Also added a test for this case.
Caused by f04260d8c6.
Cycles' CMake defines macros with the same name as Blender, which
override the Blender ones. There's however a small difference in the
re-defined `remove_cc_flag()`, the Cycles version only takes one flag at
a time. So I guess Blender's calls to it would only result in the first
flag being removed.
Of course Cycles shouldn't override any Blender macros, but I'll leave
that up to Brecht to fix properly.
For the man rationale behind this design, see 49f088e2d0. Further,
this removes users of uiBut.a1/uiBut.a2, which is a very ugly design
choice (hard to reason about).
Note that I had to do rather ugly, specific exceptions for the number
buttons in `ui_def_but_rna()`. But once all users of a1/a2 are removed,
this special handling shouldn't be needed anymore.
I also had to move a sanity check out of the button definition. It's now
moved into a new debug only sanity checking function executed when
finishing the layout definition (block end).
* Support precompiled libraries on Linux
* Add license headers
* Refactoring to deduplicate code
Includes work by Ray Molenkamp and Grische for precompiled libraries.
Ref D8769
- Declare variables where they are initialized
- Use consistent variable and static function names
- Use helper functions more for common operations
- Remove use of BezTriple struct, use CurveProfilePoint instead
- Apply small simplifications to code in some situations
This replaces the switch case and operation types with a
SculptGestureOperation struct with callbacks.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8798
This replaces the switch case with function pointers to
extract masks or face sets.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8796
Sometimes interpolate all frames is not convenient and it's better, for example, interpolate in twos.
The new parameter allows to define the number of frame for each step, by default is set to 1 as before.
{F8812621}
This is a request of animators to improve interpolate tools.
Reviewed By: mendio
Maniphest Tasks: T80190
Differential Revision: https://developer.blender.org/D8723
239b0b
Remove redundantly nested `#if` and `#ifdef` statements.
One nested `#if 0` block was left untouched, as it's in particle code
that's no longer maintained. Furthermore, that block also has some
explanation as to the differences between the enabled & disabled parts.
One nested `#if 0` construct was completely removed, leaving only the
actually used bit of code. There was no explanation as to the usefulness
of the disabled code, and it hasn't been touched in years.
No functional changes.
With the 'Restrict' option, the gradient should be restricted to the
assigned vertex, just like the other weight paint tools.
Maniphest Tasks: T80273
Differential Revision: https://developer.blender.org/D8761
This was reported for the "Add Node" search functionality, but is
relevant in other searches as well.
So e.g. when searching for "Separate XYZ", typing "sep", then " " (with
the intention to type "X" next) would clear the search field. Now use
the same method (matching against all search words) as in F3 searching
('menu_search_update_fn') in other searches as well [searching IDs,
property objects, finding nodes,...]
This should give a much nicer search experience in general.
Note: this does not touch other searches in the Dopesheet, Outliner,
Filebrowser or User Preferences that have other search implementations.
Maniphest Tasks: T78084
Differential Revision: https://developer.blender.org/D8232
The "set_expand_from_list_data" function has to be called after updating
the panels' custom data pointers. Otherwise the wrong information is
retrieved in the best base and it will cause a crash in the worst case.
Changing the active object would change the mode, but the necessary
toolsystem update to change the cursor was not run.
Steps to recreate were:
* Duplicate default cube
* Enter sculpt mode with one of the cubes
* Disable Edit > Lock Object Modes
* Select back-and-forth between the objects
Same would happen with other modes.
CMake caches the location of the CRT runtime in the
MSVC_REDIST_DIR variable, and uses it to copy the required
dll's during the install phase.
This variable is only initialized when it does not exist.
Leading to issues when compiler updates are installed and
the compiler version slightly changes, cmake still looks
in the old location for the runtime and warns about the
files not existing.
This change fixes the issue by checking if the redist dir
exists and if not unsets it so InstallRequiredSystemLibraries
can have another go at figuring out where they live.
The current 1D Voronoi implementation for the Distance to Edge option
computes the distance to the cells instead. This patch fixes that and
compute the distance to the edge.
Reviewed By: JacquesLucke, brecht
Differential Revision: https://developer.blender.org/D8634
We do this in a couple of places, so it's worth having the logic wrapped
into a function.
Also, the only way to set the disabled hint for a button from outside of
`interface/` was through `UI_block_lock_set()`/`UI_block_lock_clear()`,
for which the usage isn't obvious when you just try to disable a single
button.
This reverts commit 17905e89a7.
Fix T80429: Transform Tools cannot be cancelled
rB17905e89 better filtered the key items that should appear in the
status bar.
But it is very restrictive since canceling is still possible in other
cases.
# Conflicts:
# source/blender/editors/transform/transform.c
This implements a Face Set Extract operator, which is similar to mask
extract. This operator uses a picker to select and Face Set in the mesh
and extract the geometry directly to a new object.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8599
We have established a convention of exposing the most important
settings from popovers, then aligning a popover to control more
detailed settings.
The file browser has powerful display / filtering settings, but
they're hidden in popovers at the moment, so it's sometimes a pain
to use them. The "display as" options are now exposed to the left of
the display settings popover, and the "filter" toggle is exposed to
the left of the filter settings popover. This convention is familiar
and intuitive for users and makes interaction faster.
Note that the "show hidden" item in the filter popover still has an
effect if filtering is disabled.
This commit also:
- Removes the icons in the "Sort By" enum
- Uses property split for the "Sort By" enum
- Very slightly increases the default width of the file browser
window to make room for the new buttons.
Differential Revision: https://developer.blender.org/D8719
This was doing a matrix inversion per vertex per stroke step and it was
unused.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8791
All brushes except for grab need delta for tip orientation in order to
work with anchored stroke, not only snake hook, which is the one that
needs it always.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8789
This implements the sculpt gesture lasso and box operators for face
sets. They work the same way as the mask gesture operator and tools.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8716
It was possible to increase the text size quite a lot, making text much
bigger than their containing widgets. These new limits makes the maximum
size of text be closer to the widget size.
Note that there needs to be some extra "wiggle room" since fonts may
have quite different glyph sizes. So you can still set the font size to
be slightly bigger than the widgets.
Addresses T80175 and T79059.
Some underlying functionality was not ready for greasepencil:
- BKE_modifiers_get_virtual_modifierlist (now introduce dedicated BKE_gpencil_modifiers_get_virtual_modifierlist)
- BKE_modifiers_is_deformed_by_armature
- checks in drawing code
- checks in (pose) selection code
A couple of changes to make this work:
- `eGpencilModifierType_Armature` has to be respected (not only `eModifierType_Armature`)
- `OB_MODE_WEIGHT_GPENCIL` has to be respected (not only `OB_MODE_WEIGHT_PAINT`) -- (now use new `OB_MODE_ALL_WEIGHT_PAINT`)
- `gpencil_weightmode_toggle_exec` now shares functionality from `wpaint_mode_toggle_exec` -- moved to new `ED_object_posemode_set_for_weight_paint`
This patch will also set the context member "weight_paint_object" for greasepencil (otherwise some appropriate pose operators wont work when in weightpaint mode)
Reviewed By: campbellbarton
Maniphest Tasks: T63125
Differential Revision: https://developer.blender.org/D8483
Condition was flipped, would allow actions on constraints coming from
library, but prevented actions on local constraints.
Mistake in rBS0b49fdd0ee0.
Maniphest Tasks: T80391
Differential Revision: https://developer.blender.org/D8793
This change doesn't impact release builds,
in general avoid having defaults depend on build options
since it means files from different builds won't match.
The Snake Hook deformation mode was using the same strength as grab (not
supporting pressure), but this deformation mode supports pressure.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8724
All filters were using prevclicx, which is in screen coordinates and
mval[0], which is in region coordinates to get the filter strength.
This fixes the issue in all filters.
Reviewed By: Severin
Maniphest Tasks: T80311
Differential Revision: https://developer.blender.org/D8776
For modifier shortcuts we added a "custom_data" field to panels.
This commit uses the same system for accessing the list data that
corresponds to each panel. This way the context is only used once
and the modifier for each panel can be accessed more easily later.
This ends up being mostly a cleanup commit with a few small changes
in interface_panel.c. The large changes in the UI functions are due
to the fact that the panel custom data is now passed around as a
single pointer instead of being created again for every panel.
The list_index variable in Panel.runtime is removed as it's now
unnecessary.
Differential Revision: https://developer.blender.org/D8559
Blender tried to free objects twice from the bullet world sometime.
First we would implicity remove all objects when recreating the bullet
world and then explicity try to remove them again from the now empty
world.
This would wouldn't crash older bullet versions, but the recent versions
will as we will try to free objects that no longer exists in the bullet
world.
Also clear the cache on deletion as the object order changes.
Fix T77181: The cache clearing will fix this issue.
Continuing the work from D8326, this commit adds a property split
layout to the add torus operator. It also puts the properties common
to all object add operators at the bottom for consistency.
Differential Revision: https://developer.blender.org/D8748
This change enables the developer option `WITH_CYCLES_NATIVE_ONLY`
for MSVC. This allows a developer to just build the cycles
CPU kernel for their specific system rather than all kernels,
speeding up development.
Other platforms have had this option for years, but MSVC lacks
the compiler switch to target the host architecture hence it
always build all kernels.
This change uses a small helper program to detect the required
flags.
Only AVX/AVX2 are tested, for the following reasons
- SSE2 is enabled by default and requires no flags
- SSE3/4 have no specific build flags for msvc
- AVX512 is not yet supported by cycles
Differential Revision: https://developer.blender.org/D8775
Reviewed by: brecht, sergey
The animated objects was not updated for each internal substep for the rigidbody sim.
This would lead to unstable simulations or very annoying clipping artifacts.
Updated the code to use explicit substeps and tie it to the scene frame rate.
Fix T47402: Properly updating the animated objects fixes the reported issue.
Reviewed By: Brecht, Jacques
Differential Revision: http://developer.blender.org/D8762
This patch moves the EEVEE depth of field shaders to eevee_shaders.c and
adds them to the eevee shaders test suite.
Reviewed By: Clément Foucault
Differential Revision: https://developer.blender.org/D8771
There was no way to reset the current file to factory settings
without reloading the preferences (which disables & re-enables add-ons),
this slows down resetting files and can complicate tests.
- Use WM_operatortype_description to get the operator description.
- Pass properties to WM_operatortype_name,
so the operator name callback is used.
- Add UI_but_operatortype_get_from_enum_menu function
to access the operator from enum menus.
- Change WM_operatortype_description to return NULL when there is no
description, use WM_operatortype_description_or_name
when either can be used.
This function was called when the modifier list changes and the panel
list has to be rebuilt. Originally I thought it was necessary to to remove
the block immediately when the panel was removed, but we can just
leave it and it will be removed later in the UI drawing process.
Removing this results in fewer string lookups.
Before, Cycles was using a shared Embree device across all instances.
This could result in crashes when viewport rendering and material
preview were using Cycles simultaneously.
Fixes issue T80042
Maniphest Tasks: T80042
Differential Revision: https://developer.blender.org/D8772
Recompute Rest Length stored in the StretchTo constraint after applying
the current pose as rest pose.
The "Apply Pose as Rest Pose" operator applies the evaluated pose as
rest pose, which means that the change in bone length from the StretchTo
constraint is applied to the rest pose. The bug was caused by the fact
that the StretchTo constraint wasn't updated for the new pose, and thus
still applied the same scale factor to the new pose, effectively
doubling its effect.
The "Apply Pose as Rest Pose" operator now forces a recompute of the
rest length cached in the StretchTo constraint data. As a result, the
length of the bone before and after the pose is applied remains the
same. The X and Z scale (perpendicular to the bone length) are reset to
1.0, as with the applied pose the bone isn't stretched or squashed any
more.
Remove the hack for library initialisation; this is no longer necessary
as the required information can be passed to the USD library after its
static initialisers have run.
This new approach is compatible with both the patched and original USD
library. This means that platform maintainers don't need to rebuild the
USD library until the next upgrade.
Manifest Task: https://developer.blender.org/T80320
Initialize the USD library when used (instead of at startup), so that
this can happen inside the IO/USD module. This makes calls to the USD
library local to Blender's USD code.
Note that failure to find the USD JSON files will now only be reported
when the USD exporter is used, and not on every startup of Blender.
This is the first step in cleaning up the way Blender patches and
initialises the USD library.
Manifest Task: https://developer.blender.org/T80320
Steps to reproduce were:
* Open Sequencer, add a sound strip
* In the sidebar, open the Adjust > Sound sub-panel
* Note the placement of the "Mono" item
The layout code would disable decorators if a property came from a
non-animatable data-block type. Doing so would mess up the alignment
where properties from different data-block types were be mixed.
This is not the case any more.
Note that when actually adding the decorator, a blank icon is inserted
to keep the alignment intact when the data-block type isn't animatable.
So the decorator is still not shown, but the alignment looks fine.
This may affect more cases. If so, and if that's an issue, the
decorators should be explicitly disabled.
This improves circles AntiAliasing, and line antialiasing.
This keeps the old drawing method (3d spheres) for the selection pipeline.
This was suggested by @harley on devtalk.
If the search menu was used for a string property, and a data-block was
selected from the search, the value set would be an invalid name. The
property would get the modified UI string, not the proper data name set.
This problem was already once solved in rB249ccab111ac, but resurfaced
in rB937d89afba36.
Now only use the modified UI string if requires_exact_data_name is not
true.
Note: the comments in rB249ccab111ac [reg. library hints and string
properties, also that pointer properties are preferred over string
properties when dealing with IDs] still apply.
Reviewers: campbellbarton
Differential Revision: https://developer.blender.org/D8759
(cherry picked from 2.90 commit
cb0b0416f4)
Keymaps must be customized by the user.
But this is not the case for hardcoded keymaps.
Also the repetition of hardcoded and user-defined keyitems may induce
the user to think they have made a mistake or it is a bug.
Differential Revision: https://developer.blender.org/D6454
Now that we use internal state info, we require drawing that uses the
GPU api to use it throughout the whole pipeline. This is in order to
track the GL state and do our own error checking.
Because depsgraph isn't rebuild for animated properies, we have to
assume that active bodies will always want to have updates from the
rigidbody simulation.
Calling: bpy.ops.wm.read_homefile(use_empty=True)
exposes invalid user-counts in versioning code.
Simplified logic for assigning materials in versioning code.
Caused by 29f3af9527.
- Direct assignment caused ID user counts to be invalid.
- The first palette would always be used,
even when the named palette searched for was found.
Also pass 'const' string to `hex_to_rgb`, avoid casting to 'non-const'.
Now the callbacks are setup for each debug context.
The formating has been reworked to be less verbose and make errors
and warnings stand out from the notifications.
Errors are most of the time sufficiently explicit in their message.
This also remove the support for AMD_debug_output which is 10 years old.
This is related to the Vulkan port T68990.
After discussion with @howardt, it seems the solver property should
always be exposed, even in lite builds. This commit removes the
ifdefs for that property and adds a warning if the "Exact" solver is
used when Blender is compiled without GMP.
These changes apply to the boolean modifier as well.
Delete `hdf5.diff`, because it's no longer used. Since Blender 2.90 the
optional support for HDF5 has been dropped, but this file accidentally
wasn't deleted.
See 0102b9d47e and 0c38436227.
No functional changes.
Now we have a better distinction of what is snap to grid and what is
snap to increments.
The code also allows the implementation of mixed snap for these modes.
This is part of the Vulkan backend task T68990.
This is mostly a cleanup, however, there is a small change:
We don't use a special Vertex Array binding function for Immediate
anymore and just reuse the one for batches.
This might create a bit more state changes but this could be fixed
easily if it causes perf regression.
# Conflicts:
# source/blender/gpu/intern/gpu_context.cc
Problem: the Blender synchronization process creates and tags nodes for usage. It does
this by directly adding and removing nodes from the scene data. If some node is not tagged
as used at the end of a synchronization, it then deletes the node from the scene. This poses
a problem when it comes to supporting procedural nodes who can create other nodes not known
by the Blender synchonization system, which will remove them.
Nodes now have a NodeOwner, which is set after creation. Those owners for now are the Scene
for scene level nodes and ShaderGraph for shader nodes. Instead of creating and deleting
nodes using `new` and `delete` explicitely, we now use `create_node` and `delete_node` methods
found on the owners. `delete_node` will assert that the owner is the right one.
Whenever a scene level node is created or deleted, the appropriate node manager is tagged for
an update, freeing this responsability from BlenderSync or other software exporters.
Concerning BlenderSync, the `id_maps` do not explicitely manipulate scene data anymore, they
only keep track of which nodes are used, employing the scene to create and delete them. To
achieve this, the ParticleSystem is now a Node, although it does not have any sockets.
This is part of T79131.
Reviewed By: #cycles, brecht
Maniphest Tasks: T79131
Differential Revision: https://developer.blender.org/D8540
This was caused by a double lock of the DRW context mutex.
This changes the logic a bit by releasing the DRW context before rendering
with BKE_sequencer_give_ibuf and restoring it after.
Critical fix for 2.91
Reviewed By: dfelinto
Differential Revision: https://developer.blender.org/D8657
Instead of calculating exact normals for all faces, just do it
for those that potentially intersect. A big improvement for dense
meshes that only intersect in relatively few places.
These were disabled in the newboolean merge commit.
This commit renables them, using the original 'FAST' solver
so that the result objects need not change.
A TODO to add more tests using the 'EXACT' solver,
though most functionality there is now covered by unit gtests.
This was affecting Mesa drivers as well as AMD pro driver. But it
might have been noticeable on other config too.
This was introduced by rBa9f2ebb21508.
This way it is way clearer what each viewport state is. There is
no more save and reset. The scissor test is also saved per
framebuffer.
The only rule to remember is that the viewport state (size and
origin) is reset for both the viewport and scissor when a texture
is attached or detached from an attachment slot.
This is related to the Vulkan port T68990.
This is a full cleanup of the Framebuffer module and a separation
of OpenGL related functions.
There is some changes with how the default framebuffers are handled.
Now the default framebuffers are individually wrapped inside special
GLFrameBuffers. This make it easier to keep track of the currently bound
framebuffer state and have some specificity for operations on these
framebuffers.
Another change is dropping the optimisation of only configuring the
changed attachements during framebuffers update. This does not give
any benefits and add some complexity to the code. This might be brought
back if it has a performance impact on some systems.
This also adds support for naming framebuffers but it is currently not
used.
This is in preparation of the Framebuffer GL backend.
This is a just changing types and moving some code.
No logic is changed... almost... it just removes the context attach.
i.e: `gpu_context_add/remove_framebuffer()`
This is not needed for now and was even disabled in release.
This is part of T68990.
Use this utility function for render-shading & weight paint modes.
This adds support for edit-mode & pose-mode where all objects in the
mode are used in this case instead of the selected objects.
Useful to perform cleanup operations on many objects at once,
also these operations weren't accessible from the search menu.
This follows the convention for other clean up menus
when editing mesh, curve & grease-pencil.
Resolves issues raised in T80011
This is useful to run in object-mode, instead of from the property editor,
note that this still only used the current object when activated from
the property editor.
This is useful to run in object-mode, instead of from the property editor,
note that this still only used the current object when activated from
the property editor.
The code that found coplanar clusters was not updating a bounding box.
Also, code that was detecting non-trivial coplanar intersects was
slightly wrong, but that would not have caused any functional problems.
This was added in 1cb7267a9f, however the behavior before this
would have failed on negative values already.
Also negative values here would fail in many other places.
Properties that display conditionally depending on other properties
should generally be lower down so the movement of buttons as
settings are tweaked is less intrusive.
This patch fixes assert on grid drawing. `for` loops are used instead
of `while` loops to make sure the number of lines is exact. The old
code draws lots of unnecessary lines offscreen. This bug is fixed as
well. See the patch for a comparison without a scissor test.
Differential Revision: https://developer.blender.org/D8745
Use correct text color when syntax highlighting is off but line numbers are on.
Differential Revision: https://developer.blender.org/D7337
Reviewed by Hans Goudey
Display negative zero floats as regular zero. Does not alter underlying value.
Differential Revision: https://developer.blender.org/D4795
Reviewed by Brecht Van Lommel
At first sight this code should not build at all
but due to the use of macro's that look like functions
this seemingly has no issues building.
Clang-format alerted me to this strange bit of code
by placing the `if` on it's own line for some reason.
added the missing brackets, and clang-format is happy
again.
`ssize_t` is not a standardized type (it's a posix type)
given the line in question here is calculating the size
of a memory allocation there's no logical way this
should ever be negative.
I do not know this code too well and was unsure if
`mdverts->totweight` could ever be < 0, so I protected
it with a clamp, just in case.
This is for design task T67744, Boolean Redesign.
It adds a choice of solver to the Boolean modifier and the
Intersect (Boolean) and Intersect (Knife) tools.
The 'Fast' choice is the current Bmesh boolean.
The new 'Exact' choice is a more advanced algorithm that supports
overlapping geometry and uses more robust calculations, but is
slower than the Fast choice.
The default with this commit is set to 'Exact'. We can decide before
the 2.91 release whether or not this is the right choice, but this
choice now will get us more testing and feedback on the new code.
Problem is again with the embedded data, we want to make those local
together with their owner ID, but sometimes we are actually dealing with
copies here, which are inheritently already local.
Code did not considered that possibility before, leading to access to a
NULL `lib` pointer.
This should also be back-ported to 2.83 LTS release.
Maniphest Tasks: T80104
Differential Revision: https://developer.blender.org/D8731
The CPP Shader class does not initialize the interface attribute.
What will crash when deleting the shader.
Reviewed By: Clément Foucault
Differential Revision: https://developer.blender.org/D8740
We need to have transforms from passive objects if they are animated or
driven by parent relations. This is not immediately obvious as the
object transform matrix will still be available, it is just one frame
behind in some cases.
Fixed dependency cycles if there is a constraint between two rigid
bodies. Because bullet keeps track of its simulated bodies, we do not
need to supply objects transforms as bullet should already have them.
I need combine these two fixes because otherwise we will get depsgraph
warnings that nodes are missing that it expects to be there.
Reviewed By: Sergey, Jacques
Differential Revision: http://developer.blender.org/D8732
Issue was with our dear posebones again... when applying overrides we
keep the same address/pointer for the IDs themselves, (which avoids us
the need to remap their usages), but their inner data is often
re-allocated.
Therefore, we need once again to go over armature objects and invalidate
their posebone pointers.
This should also be back-ported to Blender LTS 2.83.
Maniphest Tasks: T80078
Differential Revision: https://developer.blender.org/D8734
A test case that compiles all the GLSL shaders for workbench, gpencil, overlay and some
of eevee. Compilation is still platform dependent, but when run on a test-farm
with different hardware we will be able to detect GLSL compilation
errors early on.
The test will be compiled when `WITH_GTEST` and `WITH_OPENGL_DRAW_TESTS`
are On.
For eevee only the shaders inside eevee_shaders.c are included. EEVEE has some shaders
located inside the submodule. They aren't accessible to the outside and aren't added
to the test case. We should see how we want to add them. For the test cases it is better
to move them to eevee_shaders.c, but for eevee perspective it is better to keep them in
the submodule. Keeping them in the submodule could lead to situations that is harder to test.
as the shader could already have been initialized.
Reviewed By: Clément Foucault
Differential Revision: https://developer.blender.org/D8667
There were some missing UBO bindings, what asserted in debug mode.
This patch fixes this by binding the missing UBO's
Reviewed By: Clément Foucault
Differential Revision: https://developer.blender.org/D8742
In recent refactoring {a9f2ebb21508} an issue was introduced that the
opengl rasterizer would be disabled when only writing to a stencil
buffer.
This fix adds stencil writing to the write mask and set it. This makes
the write map not evaluate to GPU_WRITE_NONE and the rasterizer will be
enabled in `GLStateManager::set_write_mask`.
Reviewed By: Clément Foucault
Differential Revision: https://developer.blender.org/D8743
The CPP Shader class does not initialize the interface attribute.
What will crash when deleting the shader.
Reviewed By: Clément Foucault
Differential Revision: https://developer.blender.org/D8740
Caused by rB49f59092e7c8: Curves: Implement Handles for selected points
only
Changes from deselecting all were not considered as changes anymore.
Maniphest Tasks: T80182
Differential Revision: https://developer.blender.org/D8744
This is part of T76372.
It adds the `blend_write`, `blend_read_data`, `blend_read_lib`
and `blend_read_expand` which correspond to the various
steps when reading and writing .blend files.
Having these callbacks allows us to decentralize the blenloader
code a lot more. This has the affect that code related to any
specific ID type is less scattered.
Reviewers: mont29
Differential Revision: https://developer.blender.org/D8670
We forgot to update this code as part of D3108. I'd like to include this in 2.90,
it's entirely broken now so can't really get any worse.
Differential Revision: https://developer.blender.org/D8738
More localized variables, avoid ugly 'offset by one' index usage in
favor of explicit `INDEX_UNSET` define, etc.
No behavior change expected from this commit.
Checks to preserve the active spline on duplication
required an active vertex too.
Now having no active vertex doesn't prevent duplicate
from keeping the spline active.
Reviewed by: @mano-wii
Ref D8729
Left walk navigation while the scene collection is active would collapse
the subtree which shouldn't be allowed. This adds another check to
`outliner_item_openclose` to prevent collapsing the scene collection.
Introduced in rBb077de086e14.
Because the subtrees in Data API mode are empty for performance reasons,
it was impossible to move through the tree with walk navigation. This
adds an exception to allow walk navigation to expand subtrees in that
mode.
custom split normals data
Clearing custom split normals would get rid of the CD_CUSTOMLOOPNORMAL
layer - but editing data `lnor_spacearr` would be kept.
Adding a CD_CUSTOMLOOPNORMAL layer (if none exists yet) should be done
in `edbm_average_normals_exec` / `BKE_editmesh_lnorspace_update` /
`BM_lnorspace_update` / `BM_lnorspacearr_store`. The thing is that if
the editing data `lnor_spacearr` would still be valid after `Clear
Custom Split Normals Data`, blender would happily call
`BM_lnorspace_rebuild` instead. Doing that without a CD_CUSTOMLOOPNORMAL
layer is asking for trouble.
Now clear lnor_spacearr on `Clear Custom Split Normals Data` as well.
Thx @mont29 for feedback here.
Maniphest Tasks: T80159
Differential Revision: https://developer.blender.org/D8730
Currently the panel handler loops through every block and every button
for every single panel. This commit moves that check to happen a single
time at the beginning.
We need to only collapse or expand the first panel under the cursor
rather than all of them. Note that whether the parent panel or
the subpanel is first depends on the order of the uiBlocks in the
region's list.
The code for handling panel events was much more complicated than it
needed to be. This commit removes some unecessary function calls and
variables, reduces indentation levels by returning early, and does
some other general cleanup.
Somehow the panel category drawing functions ended up in the middle
of the region event handling code. This commit moves them to their
own section next to the rest of the drawing code.
If a different object was active, clicking on a linked armature's pose
in the Outliner would enter Pose Mode for it.
This would actually cause a failed assert, but in release builds the
armature would just enter pose mode.
Steps to reproduce were:
* Link in armature object
* Activate a different object
* In the Outliner, un-collapse the armature object
* Activate Pose Mode by clicking on its pose there
MSBuild on windows currently spews a warning about
buildinfo.h_fake not being generated.
For build info we use a non existing file to trigger a
custom_command on every build, which has worked well for
years now, however in recent versions of MSBuild it has
started issuing warnings about files that should be
generated but are not.
CMake is actually aware of this being a problem and states
in the documentation that "If the output of the custom command
is not actually created as a file on disk it should be marked
with the SYMBOLIC source file property."
This change fixes the build warning by properly marking the
buildinfo.h_fake as symbolic resolving the warning.
Need to make sure node factories are initialized prior to the dependency
graph allocation.
The regression was initially introduced in 5b021dff41
Thanks Brecht for testing!
It is now possible to build against a shared embree library.
Before it was only possible to build against static Embree libraries.
Reviewed By: Brecht
Differential Revision: http://developer.blender.org/D8702
MSVC already builds with the /std:c++17 flag but for
'reasons' [1] MSVC still gives the wrong value for the
__cplusplus define.
This change sets an additional cxx flag on supported
compilers to allow the compiler properly identify
C++17 support.
This resolves 2 warnings coming out of bullet about
the register keyword being deprecated.
[1] https://devblogs.microsoft.com/cppblog/msvc-now-correctly-reports-__cplusplus/
OSL requires RTTI to be off, this is done with the /GR- flag for
MSVC, however /GR is in the default CXX flags leading to warning
D9025 : overriding '/GR' with '/GR-'
which cannot be suppressed.
/GR is on by default and this flag is not required, so removing
it from the default CXX flags makes it possible later use /GR-
without generating warnings.
The changes in rB70151e41dc02 broke subtree expansion in the Data API
display mode because the closed subtrees are empty lists. Move the empty
subtree check from `outliner_item_openclose` to the walk navigation
code to prevent the issue.
`IDTypeForeachCacheFunctionCallback` lists the `flags` parameter
as `uint`, having these functions use `eIDTypeInfoCacheCallbackFlags`
results in the following warning when building with MSVC:
warning C4028: formal parameter 4 different from declaration
This change resolves this warning by changing the parameter to
the appropriate type.
The outliner already uses the alternating row theme color as an
overlay for every other row. This uses the same color for the file
browser, instead of a hardcoded shading.
The file browser background color is slightly tweaked to match the
outliner, and the Blender Light theme is updated to use a lighter
background color like the outliner.
Reviewed by: Hans Goudey, Julian Eisel
Differential Revision: https://developer.blender.org/D8717
This adds a new `--debug-exit-on-error` flag. When it is set, Blender
will abort with a non-zero exit code when there are internal errors.
Currently, "internal errors" includes memory leaks detected by
guardedalloc and error/fatal log entries in clog.
The new flag is passed to Blender in various places where automated
tests are run. Furthermore, the `--debug-memory` flag is used in tests,
because that makes the verbose output more useful, when dealing
with memory leaks.
Reviewers: brecht, sergey
Differential Revision: https://developer.blender.org/D8665
when you call the BLI_array_free macro with a const pointer you get a
warning when the macro calls `MEM_freeN` (warning C4090: 'function':
different 'const' qualifiers)
This was warning originating from
`smart_uv_project_calculate_project_normals` in `uvedit_unwrap_ops.c`
Normally we resolve these with a non const cast at the callsite
but given BLI_array_free is a macro not a function this is not
an option here and it has to be resolved in the macro.
Since the switch to TBB, the threads value specificies the maximum number of
CPU cores used while rendering, it is not longer possible to use more threads
than cores. Change the tooltip to make this more clear.
The `ctrl` key was mapped to recursive bone selectable and visibility
toggling. This changes the key to `shift` to be consistent with objects
and collections. Also adds an explanation to the tooltip.
Part of T77408
Differential Revision: https://developer.blender.org/D8650
Previously the left and right arrow keys would close and open the active
tree element, but a subsequent key press would not select up or
down the tree as is common in tree-based interfaces.
Walking left and right now does a selection action after opening or closing
the active tree item. For example, a right key press on a closed element
will open it's subtree, and an additional right key press will select
the first child element.
A left key press anywhere in a subtree will first close the active
element if it's subtree is expanded. Walking left again will select the
parent element.
Part of T77408
Differential Revision: https://developer.blender.org/D8650
The main CMakeLists.txt specifies C++17, quadriflow tries to add C++14
flags leading to the following warnings when building with MSVC
Command line warning D9025 : overriding '/std:c++17' with '/std:c++14`
This change removes the C++14 flags, and fixes a build error caused
by the removal of `std::unary_function` in C++17 in the .obj loader
(which isn't used by blender)
Reviewed By: zeddb
Differential Revision: https://developer.blender.org/D8720
For bullet we compile at /W0 for MSVC but we did not
remove the standard /W3 flag. Leading to the following
warning:
Command line warning D9025 : overriding '/W3' with '/W0'
This change removes the W3 flag for bullet to get rid
of the warning.
When Blender is started in fullscreen mode from the command line,
or if the fullscreen state is saved in the startup file, all temporary windows
will also open in fullscreen mode. When closing the fullscreen File Browser,
Blender would either crash or parent window becomes black.
This does not happen if the Blender switches to full screen manually.
`NSWindowCollectionBehaviorFullScreenPrimary` should be set for windows that
can enter full-screen mode. Otherwise macOS will turn the wrong window into
full-screen.
Similar fix: rB4b39de677d20
Differential Revision: https://developer.blender.org/D8708
Reviewed by: Julian Eisel
This refactors Box Mask and Lasso mask making both functions share the
same code. After this change it should be easier to add new
functionality, new gesture tools or implement new gesture modes.
No functional changes.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8707
Use BKE_mesh_wrapper API access to access mesh coordinates
for modifier evaluation.
Call BKE_mesh_wrapper_ensure_mdata when binding
since it's a one off operation.
Regression from deaff945d0.
Reviewed by: @brecht
Ref D8709
This was an oversight in rB83e3d25bcae3.
Basically we still have the "hair" and "point_cloud" entries for the
context. However they were ifdef'ed.
Note this would mostly happen in 2.90 since we always build without hair
and particles there.
Differential Revision: https://developer.blender.org/D8712
- Cycling between part of the boundary & the entire UV boundary.
- Include pole vertices in the selection.
Edge loop selection was rewritten to use BMesh connectivity data.
Further changes to interface_handlers.c to avoid a block of variable
declarations at the beginning of functions. Also use const in some
situations. I only made changes where the variable's intended scope
was obvious.
- Reduce variables scope
- Use some const prefixes
- Initialize variables at declaration
- Use comparison to boolean false instead of 0
Differential Revision: https://developer.blender.org/D8678
This fix a GL_INVALID_VALUE error on startup due to 0.0f max line width.
Also moves the max anisotropy filter to the sampler creation.
This reduces code fragmentation.
Grease pencil modifiers already had defined outliner icons, but had
never been included in the tree. This adds the modifiers and the shader
effects to the tree.
Part of T68498
The previous fix loaded the pixels so existing tiles were not overwritten.
However the Cycles render buffer is expected to be scaled by the number of
sample, which was not taken into account.
This is not ideal in that previews could have a mismatched number of samples
between multiple objects, though the result will be correct. The better solution
would be to bake all objects together per tile, rather than one after the other.
But that is a bigger change than we can do in 2.90.
Differential Revision: https://developer.blender.org/D8704
This prevents UV layer mix up in
MeshBatchCache.cd_used/cd_needed/cd_used_over_time which depends on the
extraction method.
One object's mesh can be accessed with MR_EXTRACT_MESH, another object
that uses the same mesh can use MR_EXTRACT_BMESH based on
(Object.mode & OB_MODE_EDIT), this causes a problem as the edit-mesh
and the mesh aren't always in sync, the custom data layers wont
necessarily match up, causing T77359.
Reviewed by @jbakker, @brecht
Ref D8645
This should not cause any problem since the depth test is required
in order to draw to the depth buffer and this is not altered by
this change.
To be on the safe side, we still restor the mask after altering it.
This workaround addresses T79533 and T79038 for Blender 2.90. The
solution isn't clear and needs more research and work. In order to
continue with blender 2.90 release it was proposed to add this work
around.
It has been tested with the test files provided in the reports.
Reviewed By: Dalai Felinto, Clément Foucault
Differential Revision: https://developer.blender.org/D8695
This workaround addresses T79533 and T79038 for Blender 2.90. The
solution isn't clear and needs more research and work. In order to
continue with blender 2.90 release it was proposed to add this work
around.
It has been tested with the test files provided in the reports.
Reviewed By: Dalai Felinto, Clément Foucault
Differential Revision: https://developer.blender.org/D8695
Use invert_m4_m4_safe_ortho when joining objects so zero scaled axis
doesn't cause all points to be scaled to zero.
Instead geometry is left un-scaled on degenerate axes.
Report a warning in this case since users may want to adjust the
active objects scale.
Unlike invert_m4_m4_safe, this calculates zeroed axes.
Useful when we need to use the inverse of an objects matrix,
keeping the valid axis, only filling in the zeroed ones.
This implements Snake Hook as a deform type for the cloth brush. This
brush changes the strength of the deformation constraints per brush step
to avoid affecting the results of the simulation as much as possible. It
allows to grab the cloth without producing any artifacts in the surface
and create more natural looking folds than any of the other deformation
modes.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8621
In 2.83 and previous versions there was a bug that was causing boundary
vertices to be detected incorrectly that was preventing the smooth brush
to work on boundaries if there was a pole on them.
In 2.90 the boundary vertex detection was fixed, but it was still using a
simplified version of the algorithm without any boundary smoothing. This
patch implements a similar smoothing algorithm to what I think it was
the intention of 2.83 and previous versions, but working correctly.
Reviewed By: sergey
Maniphest Tasks: T80008
Differential Revision: https://developer.blender.org/D8680
This is quite embarassing... it was returning the base instance instead of
the correct vao. No wonder that it was causing crash and at most drawing
issues.
This skips the subdivision operation if the mesh has no loops, avoiding
the crash.
Reviewed By: sergey
Maniphest Tasks: T80039
Differential Revision: https://developer.blender.org/D8696
The return value of scene_intersect must be checked, the isect struct members
can't be assumed to be initialized if that returns false.
Differential Revision: https://developer.blender.org/D8692
The problem occurs when a deforming modifier is added to the object
after the MeshSequenceCache modifier. We should only consider the cached
velocities if the MeshSequenceCache modifier is the last one on the
object and we also need to check for the correct vertex count before
adding the motion vertex attribute.
Ensure that time source always is in the dependency graph, allowing to
tag the graph for time update prior it was fully built.
Collaboration of Philipp Oeser, Jacques Lucke and myself.
Thanks everyone :)
`TH_UV_OTHERS` is a theme option that isn't hooked to anything since
blender 2.80. This patch will remove the option and related code.
Reviewed By: Campbell Barton
Differential Revision: https://developer.blender.org/D8669
The code uses OpenGL functionality, so is to be linked against
OpenGL libraries.
This makes it easier to integrate with cycles using CMake.
Differential Revision: https://developer.blender.org/D8371
This was caused by motion blur camera movement tagging the view as
invalid and thus resetting the temporal sampling.
Critical fix for 2.90. Need second look, but quite confident. This function
is only called once when Motion blur is off.
Reviewed by: jbakker
Differential Revision: https://developer.blender.org/D8676
This operator is dependent on mouse position (if the Use Cursor Position
option is used). The Channel property is irrelevant/unused in this case.
So it is not optimal displaying this property when calling this from the
menu (or even using the shortcut with default settings).
Now use a custom UI in the Adjust Last Operation panel in this case.
The properties are now drawn in relation to another then (Channel
underneath Use Cursor Position) next to some other minor layout
improvements.
Thx @HooglyBoogly for feedback (also providing UI code)
Maniphest Tasks: T79872
Differential Revision: https://developer.blender.org/D8625
This replace `GPU_clear()` by `GPU_clear_color()` and `GPU_clear_depth()`.
Since we always set the clear value before clearing, it is unecessary
to track the clear color state.
Moreover, it makes it clearer what we clear the framebuffer to.
When the OSL_ROOT variable is set this is ignored
by findpackage on cmake < 3.12. CMake 3.12 and up
also ignore it and warn about it. This change
tells cmake it is OK to use the variable and
stop warning
This changes the state of occlusion queries quite a bit. Now scissors is
explicitely disabled and we enabled color write.
I still don't understand why we now need this. This patch is just trial and
error on an affected setup. Note that on the same computer, renderdoc
was not able to capture the regression (the regression did not manifest
during capture).
Regression likely introduced by rB5f414234ddea
Mistake in b077de086e. I did the same fix for a few operators there,
but missed the object "Join" one.
The joining operator changes the layer content. So it must send a
notifier for that.
Before b077de086e that didn't cause a noticeable issue, because the
Outliner happened to listen to other notifiers (active/selection
changes) the operator sent and fully rebuilt its tree in response. Now
missing these notifiers can be more problematic, since we try to avoid
more rebuilds.
Added comments to the notifier types to avoid at least this pitfall.
Since stone age of lib_query, the code would iterate over the current
list of active sequences (from a meta strip e.g.), and not over the
whole list of those.
This is a critical issue as it means in some cases (editing a meta strip
typically), some ID pointers would be missed/ignored by this foreach
looper, which is now at the center of most of our ID management code.
This caused a bug here at the studio, leading to loss of all sound IDs
used by sound strips when editing and undoing inside a meta strip, since
ID refcounting would not happen properly on strips using sounds outside
the meta-strip context during file reading of the undo steps.
To be backported to 2.83.
Differential Revision: https://developer.blender.org/D8671
This is in preparation of vulkan backend. We move all opengl
functionnalities behind an abstract class.
This also cleansup the "dynamic" ubo create and rename it to
`GPU_uniformbuf_from_list()`
Contains, no functional change.
Part of T68990 Vulkan support.
This follows the GPU module naming of other buffers.
We pass name to distinguish each GPUUniformBuf in debug mode.
Also remove DRW_uniform_buffer interface.
Remove `exec()` and `eval()` calls from `WM_OT_properties_edit` where
possible. This not only results in simpler, cleaner code, but also
removes the necessity for `repr(value)` to evaluate to a Python
expression that in turn evaluates to `value` again.
No functional changes.
Reviewed By: campbellbarton
Differential Revision: https://developer.blender.org/D8658
Split the depsgraph allocation into a separate function
`BKE_scene_ensure_depsgraph()`. Parameters are only passed to those
functions that actually need them. This removes the the "if that boolean
is `false` this pointer is allowed to be `NULL`" logic and more cleanly
decouples code.
No functional changes.
Reduce nesting by flipping conditions and returning early. It's now much
clearer that it's actually a linear function (rather than a nested one).
No functional changes.
For the sanity checks to work we don't actually need to check other
scenes. So this function can be simplified so that it does not require
a `Main *`.
Mistake in 5cc08510e0.
- Use const in some places where applicable
- Use consistent variable names in some places
- Use "r_" prefix for return arguments
- Declare variables in smaller scope where they are used
We can avoid the rather expensive outliner tree rebuilds and only redraw
if nothing but the selection or active item changes. This should give a
bit of speedup for heavy scenes.
For this to work I had to correct a few notifiers, some were only
sending selection/active change notifiers that actually did things like
adding objects. I also added a more precise notifier type for when the
active collection changes. At the notifier subtype/action level we're
not even close to running out of bits, so this should be fine.
Also had to correct a wrong notifier check (was using `&` rather than
`==`).
Code dealing with object copy of master collection was bugy in case one
of the new object copy would get a name lesser than the original object,
leading to new copy being inserted before original one in lists.
Maniphest Tasks: T79931
Differential Revision: https://developer.blender.org/D8656
Allow adding keys outside of the NLA strip, when Sync Length is
activated. Before this, an animator would have to exit tweak mode,
adjust the strip length, then enter tweak mode again.
Now the animator can freely insert and modify keys outside of the strip
bounds. They will never have to touch the strip bounds directly.
Reviewed By: sybren
Differential Revision: https://developer.blender.org/D7533
This moves the allocation for filter to separate functions, removes all
repeated calls to RNA_enum_get an fixes some other code quality issues.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8639
This updates the view navigation origin after modifying the pivot
position.
Reviewed By: sergey
Maniphest Tasks: T79924
Differential Revision: https://developer.blender.org/D8632
Cloth deform brushes (both in cloth brush or other tools with cloth
simulation targets) should not have their spacing as a value relative
the brush radius to avoid affecting the simulation speed when changing
the brush radius.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8620
This adds the orientation modes to the Cloth Filter.
Similar to the mesh filter, they set the orientation of the axis when
limiting the forces.
When using the gravity mesh filter, the orientation also sets the gravity
direction. In world orientation, cloth will always fall towards the
ground plane. In view mode, cloth will always fall down relative to the
view.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8618
Change the way Sync Length works so that it keeps keyframes at the same
key as before the sync. Now when the user exits tweak mode, the overall
NLA result is unchanged.
Reviewed By: sybren
Differential Revision: https://developer.blender.org/D7602
Popovers which are aligned with properties should turn themself
(and the panel they open) inactive when the corresponding property
toggle is turned off. A notable exception is the snap settings
popover, as snapping can still be used when the toggle is off.
Differential Revision: https://developer.blender.org/D8584
This reverts {rB1693a5efe91999b60b3dc0bdff727473b3bd00bb}
and implements an alternative solution.
The old patch had the problem that the depsgraph would always
evaluate at the current frame of the original scene (even when
`DEG_evaluate_on_framechange` was used). Now it is possible
to evaluate the depsgraph at a specific frame without having to
change the original scene.
Reviewers: sergey, sybren
Differential Revision: https://developer.blender.org/D8616
Add `-Wno-maybe-uninitialized` option to `CMAKE_CXX_FLAGS_RELEASE` and
`CMAKE_CXX_FLAGS_RELWITHDEBINFO` variables on GCC/Linux.
In Release builds GCC's `-Wmaybe-uninitialized` warning is unreliable,
and thus causes noise that can drown out other warnings. These warnings
are now silenced in release mode builds.a
Debug builds seem fine, so flags for debug builds are not touched by
this commit.
No functional changes.
Reviewed By: Sergey
Differential Revision: https://developer.blender.org/D8615
This reverts commit 52f40bcff2.
Apologies for the noise. I caught a problem with this that I hadn't before. I will
commit later after thorough testing.
This will re-link all usages of a library override data-block
(including all of its override dependencies) to its reference linked
IDs, and delete those liboverrides.
As usual, it is available in the ID sub-menu of the outliner context
right-click menu.
Part of T76555.
RNA update function would only update objects from 'main' instantiated
collection, not those from sub-collections.
This should be comitted to 2.90 (and backported to 2.83 too).
Maniphest Tasks: T79935
Differential Revision: https://developer.blender.org/D8654
RNA update function would only update objects from 'main' instantiated
collection, not those from sub-collections.
This should be comitted to 2.90 (and backported to 2.83 too).
Maniphest Tasks: T79935
Differential Revision: https://developer.blender.org/D8654
Continuing the work of eb9055a572, remove remaining unecessary
variables and arguments that were related tabbing and horizontal
alignment of panels. For example, "vertical" was always true, and
removing that exposed other unused variables.
Horizontal panel alignment hasn't been used for years, and we have no
plans to use it in the future. It adds a fair amount of complexity to
the panel code which makes adding features take longer.
This code removes the X closing flag, and all of the logic / variables
unused without it.
This commit includes a file subversion bump.
Differential Revision: https://developer.blender.org/D8601
Using C++ exceptions in Blender is difficult, due to the large
number of C functions in the call stack. However, C++ data
structures in blenlib should at least try to be exception safe,
so that they can be used if someone wants to use exceptions
in some isolated area.
This patch improves the exception safety of the Vector, Array
and Stack data structure. This is mainly achieved by reordering
some lines and doing some explicit exception handling.
I don't expect performance of common operations to be affected
by this change.
The three containers are supposed to provide at least the
basic exception guarantee for most methods (except for e.g.
`*_unchecked` methods). So, resources should not leak when
the contained type throws an exception.
I also added new unit tests that test the exception handling
in various cases.
Available from the usual ID submenu in the Outliner context menu.
The goal of this operator is to re-create the override from the linked
data, while preserving existing defined overrides.
This allows to update local opverrides when relations between datablocks
are changed in source library linked data.
Part of T76555.
This splits the volume related data (properties for rendering and attributes) of the Mesh node
into a new `Volume` node type.
This `Volume` node derives from the `Mesh` class since we generate a mesh for the bounds of the
volume, as such we can safely work on `Volumes` as if they were `Meshes`, e.g. for BVH creation.
However such code should still check for the geometry type of the object to be `MESH` or `VOLUME`
which may be bug prone if this is forgotten.
This is part of T79131.
Reviewed By: brecht
Maniphest Tasks: T79131
Differential Revision: https://developer.blender.org/D8538
Now by default the selection mode is All keyframes types.
Also removed the icon for All option and renamed from `All Types`to `All`.
UI review by @pablovazquez
Users may input an invalid scene/view-layer combination via Python. We
can easily detect that and fail gracefully.
Besides checking input parameters in RNA, also assert the
scene/view-layer combination is valid in lower level Depsgraph lookup
function (called by statistics query).
This RNA/BPY function was removed in c08d847488. For understandable
reasons really-- getting scene statistics from a string displayed in the
status bar is not exactly the best design. But we have committed to not
changing the RNA API too much for the 2.90 release, so we would like to
keep this functionality.
Differential Revision: https://developer.blender.org/D8604
Reviewd by: Julian Eisel
This RNA/BPY function was removed in c08d847488. For understandable
reasons really-- getting scene statistics from a string displayed in the
status bar is not exactly the best design. But we have committed to not
changing the RNA API too much for the 2.90 release, so we would like to
keep this functionality.
Differential Revision: https://developer.blender.org/D8604
Reviewd by: Julian Eisel
During alpha the user preferences > experimental featuers are available
to prevent merge issues and allow developers to seek feedback.
This needs to be manually turned off when we branch for beta, otherwise
the RNA of the incomplete features will be exposed.
I only skipped a few loops in the monstrous ui_handle_menu_event
function. Also, I only changed variable names where necessary to
prevent redeclarations.
Differential Revision: https://developer.blender.org/D8586
This make use of the GLStateStack functions for:
- `GPU_blend()`
- `GPU_blend_set_func()`
- `GPU_blend_set_func_separate()`
The goal is to unify them using an explicit state setting.
This will remove the need to use obscure blend functions
Now error printing only display the line related to the error.
We also put char marker if present.
Example:
```
-- Shader Compilation Errors : MAMaterial --
10414 | node_fresnel(, facingnormal, viewposition, tmp34);
| ^
| error: syntax error, unexpected ',', expecting ')'
----------------------------------
```
This mainly removes the bmain argument, which can be
retrieved from the graph itself.
Also, I removed some outdated/unnecessary comments.
Reviewers: sergey, sybren
Differential Revision: https://developer.blender.org/D8614
This adds a new brush property called "Deformation Target" which
controls how the brush deformations is going to affect the mesh data. By
default is set to Geometry, which makes the brushes displace the
vertices. When set to Cloth Simulation, the deformation of the brush is
applied to the cloth solver constraints, so the simulation is
responsible to apply the final deformation. This allows to add cloth
simulation effects to other sculpt tools with minor modifications to their
code.
This patch enables Cloth Simulation deformation target for Pose and
Boundary brushes, which are tools that are already designed to work in
low poly counts and produce large deformations. This allows creating the
most common cloth effects, like bending and compressing folds, without
relying on collisions.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8578
Same concept as the Multires Displacement Eraser Brush but implemented
as a mesh Filter. This allows to delete the displacement of an entire
area uniformly.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8608
This enables the invert mode in the smooth brush as Enhance Details.
The operation is similar to the Sharpen Filter intensify details parameter,
which consist in applying the laplacian smooth displacement in the opposite
direction calculated using the original coordinates.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8509
The evaluated mesh is now only recomputed when the required data layers
are missing. Previously the evaluated mesh was re-evaluated incorrectly,
which caused the stippled result and the failing assert.
Since now only the evaluated mesh is used, and never a temporary mesh,
there is also no more need to keep track of whether the mesh needs
freeing or not.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8574
Issue introduced on fe045b2b77.
Since the stereoscopy compositing (anaglyph, ...) is only done for
viewports the VSE preview and compositor need to use viewports.
Reviewed by: dfelinto
Differential Revision: https://developer.blender.org/D8472
Issue introduced on fe045b2b77.
Since the stereoscopy compositing (anaglyph, ...) is only done for
viewports the VSE preview and compositor need to use viewports.
Reviewed by: dfelinto
Differential Revision: https://developer.blender.org/D8472
Previously, the XYZ deform axis of the Mesh Filter were limited to
object space (which is the default for sculpt mode). Now it is possible
to limit the XYZ displacement in Local, Global or View space.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8582
For 2.90 release this should not be exposed in the RNA API.
In master this needs to be ON by default, that's all.
Differential Revision: https://developer.blender.org/D8589
The Pass struct is now a Node and the passes are moved from the Film
class to the Scene class.
The Pass Node only has `type` and `name` as sockets as those seem to be
the only properties settable by exporters (other properties are implicit
and depend on the pass type).
This is part of T79131.
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D8591
Final releases (including beta) should strictly show features that are
finalized to prevent loss of data, old API clanging around, and the
overall quality of the product (Blender) presented.
Note that rendering should never be affected by user preferences, so
this is only changing things in the UI level.
Development note: This is reset experimental UI on file load.
Also note: to hide RNA (needed for hair and particles) will be done as a
separate patch.
Differential Revision: https://developer.blender.org/D8606
After drawing, the batch could be reset and draw again using other
parameters.
Avoid having to track references by setting the batch pointer to NULL after
drawing.
This uses the same concept of the Mesh Filter but for applying the cloth
filter forces, so now it can be limited to a specific axis.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8567
This adds an option to the Multires modifier to sculpt directly on the
base mesh while previewing the displacement of a higher subdivisions
level. What this does it considering Multires as a regular modifier
without exposing the grid displacement to sculpt mode.
This allows to see the propagation happening in real time, which enables
to use complex tools like Cloth or Pose in much higher resolutions and
without surface noise and artifacts.
Reviewed By: sergey, Severin
Differential Revision: https://developer.blender.org/D8555
Functions `mesh_create_eval_final_view()` and
`mesh_create_eval_final_render()` were doing the exact same thing,
except for a hack introduced in d3eb9dddd6 (2012-10-08, Better fix for
T32846: dupligroup messes up particle instancing on rendering) that
appears to be no longer necessary. Besides that, these functions had
confusing names. Their functionality changed over time, and whether to
do for-render or for-viewport evaluation is now actually determined by
the depsgraph evaluation mode. This means that the `..._render` function
could evaluate a mesh with viewport settings, and vice versa.
The functions are now merged into `mesh_create_eval_final()`, and the
hack has been removed. The `OB_NO_PSYS_UPDATE` flag has been removed
entirely (instead of keeping it around as deprecated flag), because it
was always only temporarily set on objects during mesh evaluation and
thus not saved to the blend file.
No expected functional changes as far as users are concerned.
One should be able to edit overridable IDProps values, but never their
settings/definitions.
Note that being able to add new IDProps to overrides is still a TODO.
Reported by Josephbburg (@Josephbburg) over blenderchat, thanks.
Even the Cloth Brush grab works like a regular grab brush, it makes
sense to support spacing just in this brush in order to prevent creating
more brush steps that update the simulation. This way, it is possible to
create grab brushes that update the simulation constantly while grabbing
(using the dots stroke mode) or brushes that only update the simulation
when the cursor moves (using spacing).
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8568
This code was left here after the refactor, it was doing nothing and it
was causing an assert.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8579
The Face Set visibility operator was using the last active Face Set
updated by the paint cursor, so when the paint cursor is not used (when
using a filter or a transform tool), the active Face Set was not updating
and it was hidding the wrong Face Set based on the last cursor position
with a brush tool active. Now the Face Set Visitility operator has an
invoke callback wich forces a active vertex and face set update
regardless of the active tool, so it should always work correctly.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8580
This uses a single row for the three axis instead of a row per axis.
Reviewed By: sergey, Severin
Differential Revision: https://developer.blender.org/D8581
This was a TODO in the code. Previously the Face Set datalayer was
deleted and recreated with a constant ID of 0. Now the datalayer is
preserved and set to the SculptSession after slicing the mask and a new
ID is calculated for the new faces that the slicing operation produced,
so they can be easily isolated for further tweaking.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8583
This will be used for new features like supporting cloth deformation in
other brushes and tools outside of the cloth brush code.
No functional changes.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8602
This moves `Session::get_requested_device_features`,
`Session::load_kernels`, and `Session::update_scene` out of `Session`
and into `Scene`, as mentioned in D8544.
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D8590
Some constraint-specific operators, like set/clear inverse matrix of
childof constraint, are also valid on original, linked/overridden
constraints.
Similar change to what was done to modifiers poll function a few days
ago.
Reported by Josephbburg (@Josephbburg) over IRC, thanks.
The description is self-explanatory: enable SSE optimizations in the
FFmpeg library. They were disabled from the very beginning of the
dependency builder, possibly due to portability concerns.
FFmpeg does perform runtime check for available CPU microarchitectures,
so the codecs will still run on older hardware, but they will run way
faster on newer hardware.
For example, re-encoding 3405 2560x1376 frames on Xeon E5-2699 V4 CPU
went down from 313sec to 210.
Differential Revision: https://developer.blender.org/D8594
Changing line-width shouldn't scale cursor motion. Related to T79787.
Use dpi_fac for scaling curve error threshold & number button drag
threshold calculation.
Use 'dpi_fac' instead of 'pixelsize' to scale input sensitivity
based on the interface scale.
Also use dpi_fac for view zoom operator.
Thanks to @ISS for investigating.
Multithreaded implementation can provide significant performance gain
on CPU's that can handle more parallel tasks.
On my 8 core system I got about 3.5x faster processing speed.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8585
Define `RNA_def_property_ui_range` for head and tail transform properties.
Step of 10 should match original behavior when tweaking value.
Reviewed By: Severin
Differential Revision: https://developer.blender.org/D8513
Original sequence lookup failed, becase name changed in another thread.
Fix is same as 0471349c90 - stop prefetching before changing content
of seqbase.
I have covered more cases, and added assert so it is more obvious that
issue is in lookup, and it shouldn't fail.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8592
The fix for T75936 made it possible to export invisible objects to
Alembic. This commit applies the same approach to the USD exporter.
The USD and Alembic code is slightly different in terms of where in the
exported file the visibility attribute is stored. In USD the visibility
is used to prune the scene graph, and thus there are only two options:
"hidden" and "inherited". Setting the visiblity of a node in the scene
graph to "hidden" immediately hides all its children. To allow hidden
parents with visible children, the visibility is stored on the object
data (so the geometry/camera/lamp/etc) instead.
Add a test that checks invisible objects are iterated over by the
`IO::AbstractHierarchyIterator` class, when a suitable depsgraph is given.
No functional changes.
Add a new depsgraph builder class that includes invisible objects and
use that in the Alembic exporter.
Alembic supports three options for visibility, "visible", "inherited",
and "hidden". This means that parents can be hidden and still have
visible children (contrary to USD, where invisibility is used to prune
an entire scene graph subtree). Because of this, the visibility is
stored on the transform node, as that represents the Object in Blender
and thus keeps the Alembic file as close to Blender's own structure as
possible.
Reviewed By: Sergey
Differential Revision: https://developer.blender.org/D8595
Rename test `alembic_tests` to `alembic_export_tests`, so that its name
is consistent with the Python file containing the tests,
`alembic_export_tests.py`.
No functional changes.
The paint cursor now uses a single call to
SCULPT_cursor_geometry_info_update and the Sculpt API to update both the
radius and the SculptSession data, so random access needs to be
initialized for dyntopo before using API functions.
Reviewed By: sergey
Maniphest Tasks: T79597
Differential Revision: https://developer.blender.org/D8489
When copying between a set RNA runtime property and an unset one, code
would not behave properly.
We have to also consider NULL 'override apply' callback pointer as a
valid 'use default override apply function' case.
Instead of checking for the length of a list, just handle the error that
occurs when the length is incorrect.
No functional changes to any actual test.
Currently, the search operator in the node add menu NODE_MT_add is drawn
even if no node categories exists. This patch only draws the operator if
at least one node category passes the poll.
This patch is needed because some add-ons use custom search operator and
do not register node categories. In this case, it is undesirable to have
a search operator drawn that do nothing and is not used. One such add-on
is Animation Nodes.
Reviewed By: Jacques Lucke
Differential Revision: https://developer.blender.org/D8576
The BLI_bitmap had the potential to eat a lot of memory as it was a
dense map between all the vertices in the mesh.
Now we instead use BLI_edgeset so that we only store data when needed
(creating a spare map instead of a dense).
In EEVEE the AO renderpass influenced other render passes. Until now the
pass wasn't selectable when AO was disabled in the scene to remove these
render artifacts.
This patch allows rendering EEVEE AO pass without enabling it in the
scene. It does this by binding a fallback texture that is used by the
surface shaders.
Reviewed By: Clément Foucault
Differential Revision: https://developer.blender.org/D7956
This commit renames 'execute' to 'run' because:
- This follows Python's "PyRun" which these functions wrap.
- Execution functions can use either exec/eval modes,
making naming awkward (for future API refactoring).
This was committed as a temporary workaround in 82150f5641
as release builds were failing (only debug builds worked).
This adds `stdio.h` to the header which is now split into a file that
contains more specialized functionality.
Also move function body inside BPY_python_backtrace,
removing PyC_StackPrint as we have PyC_StackSpit() for
similar functionality that can be called from a debugger.
Checks for header alignment didn't account for tool-header & header
having different alignment.
There is no reason to use a lookup function on the area
(ED_area_header_alignment) as we already have region.
Check the regions alignment directly, remove access functions.
The existing header flip detection didn't account for mixed
tool-header and header flipping & visibility between space-types.
Now alignment syncing handles any combination of header,
tool-header & footer flipped state, in a way that can be extended
to other region types in the future.
By default boost::python in debug configuration links the release
python libraries. Which leads to loading issues with the produced
modules in a debug blender (which does use the debug python libs)
Bjam has an option to switch this to the debug libraries but when
you switch that on it changes the library names for *all* boost
libraries, even ones that don't have anything to do with python
and even the release libraries.
therefore an alternative way has been chosen and we're dealing
with this by adding a define rather than asking bjam to do it.
Many of these popovers could use a design pass in 2.91, but for 2.90 we
don't want to change any UI strings at this point, so the best way to
solve the cutoff text is to widen the popovers. Sadly this won't affect
popovers when other languages besides English have longer strings, but
solving that is a much larger task.
Another benefit is that tweaking sculpt / paint brush options feels much
less cramped with slightly wider popovers.
I only know of one string that is still slightly cutoff by default with
this patch, the "Max Element Distance" property of the pose brush in the
sculpt mode brush settings popover. But I didn't think it was worth
widening that popover more to deal with that one case.
Differential Revision: https://developer.blender.org/D8575
We have to explicitly enable fixed world space lighting. This was in
fact already done, but overridden by the code to sync the 3D View
shading settings to the VR view.
Mistake in cb578ca104. Before that, the extension vector was static,
to make sure the extension name strings wouldn't get destructed when
leaving the function. I didn't think that was an issue and couldn't
recreate one, because until the previous commit we wouldn't actually
add any extensions to the vector on Windows (the system I tested
with).
Use C++17's `std::string_view` now, which avoids the string copies
`std::string` creates for itself and thus its destruction when leaving
the local scope.
The OpenXR debug extension was disabled on Windows as a workaround. This
was an old leftover from when there was only the Windows Mixed Reality
runtime on Windows. The debug extension didn't work for it and we didn't
have a way to disable it just for Windows Mixed Reality.
Now it seems to work though, so we remove the workaround. If specific
runtimes still have trouble with the extension, we can disable it
specifically for these runtimes now.
Windows only workaround. I'll have to investigate Linux separately.
Steam's OpenGL compatibility is still new and doesn't work for us yet
(neither does it for standard OpenXR examples from what I've heard and
seen myself). We can work around that by falling back to our DirectX
compatibility layer.
Note that this DirectX compatibility still doesn't work for some
systems, see T76082.
Implementation note: Since the graphics binding extensions have to be
enabled before we can find out which runtime is in use (e.g. SteamVR vs.
Oculus, etc), we can now enable multiple graphics binding extensions but
settle for a single one to use later.
Once the SteamVR OpenGL backend works, we can remove this workaround
again.
Fixes T78267.
On VR session start with positional tracking disabled, the pose would
have an offset applied but it was supposed to start exactly at the
landmark position.
Issue is that we applied the offset to cancel out the position offset
reported by the OpenXR runtime incorrectly. We only want to do that if
positional tracking is enabled, because if not we don't even apply the
runtime's position offset. So we'd cancel something out that wasn't
there.
The new reverse iterators behave as the reverse iterators for contains from
the standard library. Have a look at the tests to see how to use them.
Using them will hopefully become easier with ranges in C++20.
A Vector can now be constructed from two iterators, which is very common
in the standard library.
New Vector.insert methods allow adding elements in the middle of a vector.
These methods should not be used often in practice, because they has a linear running time.
New Vector.prepend methods allow adding elements to the beginning of a vector.
These methods are O(n) as well.
* Avoid deep copy of vectors (technically more than a cleanup).
* Use `std::make_unique` for allocating unique pointers, rather than
manual `new`.
* Use `std::optional` for optional by-value return values, rather than
C-style `bool` to indicate success + return-argument.
* Use references rather than pointers for non-optional arguments.
* Avoid manual `new`/`delete`. Use `std::unique_ptr` for local scope
bound lifetime.
* Use C++ `nullptr` rather than C's `NULL`.
* Remove unnecessary friend declaration.
These changes are generally considered good practise and move us more to
a "modern C++" style. We can still go much further of course.
See https://isocpp.github.io/CppCoreGuidelines/CppCoreGuidelines.
Since DATA_PT_sculpt_vertex_colors has its own poll() we need to call
the poll() of MeshButtonsPanel as well
Differential Revision: https://developer.blender.org/D8563
This feature was removed during the refactor, but it's needed for this type of strokes to display the line using the stroke color without checking if the stroke is enabled or not in the material.
This color is used only for these special strokes.
This patch changes openvdb from a static to a dynamic library.
this is in preparation for enabling pyopenvdb at some point
in the future.
Differential Revision: https://developer.blender.org/D8282
Reviewed by: brecht
Many of the icons had lots of interior geometry left over from
subdivision. In these cases we should remove the interior geometry
and leave the object with the modifier for the future.
This icon mimics the details of the cloth brush icon while using the
frame style extablished for the other "filter" tools.
Differential Revision: https://developer.blender.org/D8467
This icon mimics the details of the cloth brush icon while using the
frame style extablished for the other "filter" tools.
Differential Revision: https://developer.blender.org/D8467
In this mode the preview image is always using the most of the preview
area space: it is scaled to fit, preserving aspect ratio. This makes it
possible to always have maximum of the preview region even after resize
of other areas.
This mode is enabled by default, is available in the View -> Zoom to Fit
menu. It is enabled when View All (Home key) is used, and is disabled
when manual navigation ([panning, zooming) is performed.
There is no versioning code, which means existing files will open as-is,
but new projects will have this option enabled.
Ref T78987
Maniphest Tasks: T78987
Differential Revision: https://developer.blender.org/D8549
Allows to hook per-space code which is to be run on view navigation.
This is required to have zoom-to-fit implemented in the sequencer.
There might be more cases where the clalback is to be called from,
but it could be easier to address those on the case-by-case basis
when its needed.
This changes the drawing paradigm a bit. The VAO configuration is done
JIT-style and depends on context active shader.
This is to allow more flexibility for implementations to do optimization
at lower level.
The vao cache is now its own class to isolate the concept. It is this
class that is reference by the GLContext for ownership of the containing
VAO ids.
We instead use a handle reference counter on the GPUVertBufs used by
the instancing batches. This make sure that if an update happens on the
GPUVertBuf used to contruct the batch, they will never have the same
memory address than the previously allocated ones (since they are still
pending deletion thanks to the refcounter).
This avoid the linear search to update the GPUBatch in the case a
batch is deleted (which was even a bad option since they could be only
cleared)
A handle refcount is here to avoid freeing of the GPUVertBuf datablock
if it is still referenced somewhere else.
This does not prevent deleting the actual data. This is to avoid too
much zombie data usage.
This is in order to avoid most hacks inside `draw_instance_data.c`.
It was impossible for drivers to use shape key properties, modifiers
generate a new mesh. After mesh evaluation the shape keys are no longer
necessary, and because of this the `key` pointer was not copied. As
drivers work on evaluated data, however, they do need this `key`
pointer.
This commit makes the `key` pointer available in evaluated meshes, but
this is somewhat dangerous. There was an explicit reason why the key on
result was kept at null pointer: to have the evaluated mesh in a
consistent state. Assigning this pointer makes it potentially
inconsistent, as the evaluated mesh and the original shape key may have
different topologies.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D7785
It was always possible to set it to zero by typing in the value.
This new soft limit is more consistent with the fluid cache
and the Scene.frame_start property.
This was a regression in deaff945d0 which skips copying a mesh.
Dupli-verts/faces were not updated to account for this.
This supports iterating over edit-mesh vertices & faces,
since falling back to a full copy (as we do in some places)
will be slow while transforming geometry.
This commit looks as if it would change behavior with orcos,
since any edit-mesh deformation causes them to be assigned.
However in practice there is no functional change, details in comments.
Replace the evaluated mesh with VertexDupliData.mvert since only
vertices are used. This makes dupli-vert similar to how dupli-face
was already working.
When using constant falloff symmetry was working fine because the same
deformation is applied twice on the same vertices. When using no
constant falloffs, the deformation is different between symmetry passes,
so vertices need to be separated by symmetry areas to get the right
deformation from its symmetry pass without being overwriten by the next
one.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8541
This adds the boundary_falloff_type and boundary_offset to control how the
falloff of the Boundary Brush is applied.
Boundary Origin Offset is the same concept as the Pose Origin offset in
the Pose Brush. It is a multiplier that adds extra length to the brush
radius to locate the deformation pivot further from the boundary without
affecting the falloff.
The Falloff type includes Constant (previous default), brush radius, loop
and loop and invert. Loop and Loop and Invert can be used to create
deformation patterns in a mesh.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8526
If the File Browser was used in regular editor mode (e.g. not through an
operation like File > Open), the operator that usually opens files and
directories wouldn't execute. We need to keep two operators for
double-click in the keymap so selecting and opening works in all cases.
Caused by c606044157.
This adds support for panel categories to the instanced panel system
used for modifiers and others. The change is pulled from D7997 where
it is needed for FCurve modifiers, but it is unused now.
The change is simple and basically amounts to checking the panel
category where it was overlooked before.
This brush deletes displacement information of the Multires Modifier,
resetting the mesh to the subdivision limit surface.
This can be use to easily delete parts of the sculpt or to fix
reprojection artifacts after applying a shrinkwrap.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8543
This this adds the option of building boost::python
in the libs builder, in preparation for future
dependencies that require it (ie pyopenvdb)
disabled by default, can be enabled with the
`WITH_BOOST_PYTHON` cmake option.
Differential Revision: https://developer.blender.org/D8212
Reviewed by: brecht
Previously the way to use this operations was using the shortcut Ctrl +
W and Ctrl + Alt + W, which was not very discoverable and it was
limiting the amount of options that can be added to the operator.Now the
same functionality of the operator is available as a tool, which will
make easier to add other editing operations and options without adding
more entries to the keymap.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8545
The object "delta_" rna variables were not added to the depsgraph search
and thus it would not trigger updates of the object during animation
playback.
The object "delta_" rna variables were not added to the depsgraph search
and thus it would not trigger updates of the object during animation
playback.
DRW_render_set_time is calling RE_engine_frame_set will in turn calls
BKE_scene_camera_switch_update.
To workaround this, we get the original camera object at render init and
get the evaluated version from it after each time change.
This change how motion data are indexed inside the ghash.
We follow cycles closely now and use evaluated ID pointers.
By removing the hack, it fixes T78561 (No Motion Blur on linked objects)
This fix issues with instanced geometry and modifiers. Since the
depsgraph will duplicate the objects when they have different modifiers,
the evaluated object are garanteed to be unique.
If one of the operands of a binary operator a float, integer
operand gets promoted to float as well.
Differential Revision: https://developer.blender.org/D8552
For a locked shapekey, a SculptSession's orig_cos / deform_cos /
deform_imats are not initialized (they only are when
deform_modifiers_active is true -- this in turn is only true for
shapekeys if they are //not// locked [and for deforming modifiers of
course])
We can just update that keyblock with sculpt_update_keyblock() in case
of a locked shapekey
Maniphest Tasks: T79622
Differential Revision: https://developer.blender.org/D8499
Caused by rBfffba2b6530.
In above commit, the buttons rnaindex [-1 for entire arrays like colors]
was not used anymore for the entire function body of
`ED_autokeyframe_property` (previously in `ui_but_anim_autokey`).
Replacing the index with 0 (in the array case) is indeed necessary for
`BKE_fcurve_find_by_rna_context_ui`, prior to rBfffba2b6530 this was
taken care of by using `ui_but_get_fcurve` instead [which did this
internally].
But using an index of 0 (instead of -1) for the entire function body of
`ED_autokeyframe_property` fails for the array part later, namely
`insert_keyframe` needs -1 here.
Now just replace the index for //finding the FCurve//, but use the
original for //inserting the keyframe//.
Could be backported to 2.83 LTS.
Reviewers: campbellbarton, sybren
Subscribers:
The previous algorithm was not using all of the requested grids to build a mesh
around the volume due to limitations regarding the use of a dense buffer to
gather information about the volume's topology. This resulted in artefacts during
rendering.
The mesh generation is now done by merging all of the input grids and using the
resulting grid's topology to create the mesh. The generation of the mesh
is still done in index space as before, and the vertices are converted to object
space by using the merged topology grid indexToWorld transform.
To be able to merge the grids together we have to make sure that their transformation
matrices and their index spaces match, thus, if they do not match we simply resample
the grids. This behaviour should tackle one other limitation of the current algorithm,
which is that only one transformation matrix was used to generate the final mesh.
If we do not have an OpenVDB grid for the requested volume data, we generate
a temporary OpenVDB grid for it.
Differential Revision: https://developer.blender.org/D8401
Now the one vertex defines the position of the UV chart, while rotation and
shape is still determined automatically.
Initial patch by Willis (wlssirius).
Differential Revision: https://developer.blender.org/D8484
This is particular important because 2.90 will coexist with 2.83 LTS, so
we should warn the users of potential loss of data when going from 2.90
to 2.83 and back.
Differential Revision: https://developer.blender.org/D8488
Those where assuming we always get a valid modifier data from context,
which is not always true...
Also fix similar issue with shortcuts as reported in T79635.
It is unfortunate that we cannot get active modifier from context when
operator is called from a shortcut, but we'd need an event for this to
work... So for now forbid any modifier operation of liboverride objects
in that case.
'VIEW2D_OT' operators were not respected in WM_keymap_guess_opname().
This was seemingly done on purpose (see comment "Op types purposely
skipped for now"), but dont really see the reason for doing so.
Since the "View2D" keymap is not bound to a specific spacetype, we can
still find it using WM_keymap_find_all() [and passing 0 as spacetype].
Reviewers: Severin
Subscribers:
Vert-only mesh is valid input for the skin modifier (displays isolated
cubes), prevent error message about missing root vertex in that case.
Maniphest Tasks: T79700
Differential Revision: https://developer.blender.org/D8533
saving with 'Remap Relative' option
Caused by rBf7386b97571e.
Logic in BKE_bpath_traverse_main calls the callback multiple times [as
often as there are images in the strip].
Prior to above commit, the callback was
'bpath_relative_convert_visit_cb' [this one did not have this problem -
since it returned early if the path was already made relative once]
After rBf7386b97571e though, the 'bpath_relative_rebase_visit_cb' is
used [this one should not be entered multiple times, it would modifiy the
directy again and again].
Luckily, we have a flag (BKE_BPATH_TRAVERSE_SKIP_MULTIFILE) that can be
used to prevent this (this will take care of only calling the callback
once in BKE_bpath_traverse_main for the VSE case)
Could be backported to 2.83 I think.
Maniphest Tasks: T79676
Differential Revision: https://developer.blender.org/D8536
This commit generally moves variable declarations to the smallest scope
the variables are used in. This makes the code more readable by
making it clearer when variables are used and by removing the block
of variable declarations at the top of each function.
This change makes the generated points a light dimmer than selectable points.
Before:
{F8765593}
After:
{F8765585}
Maniphest Tasks: T79683
Differential Revision: https://developer.blender.org/D8515
This options allows to perform Face Sets operations while preserving the
mesh visibility. Edit hidden face sets is enabled by default in order to
expand the visible area of the mesh with the grow/shrink operator, but
this can be changed in the keymap per edit operation as more operations
are supported.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8029
Old and new collections are the same data in Master collection case
here, so we cannot consider the `gobject` listbase of `collection_old`
as always immutable.
This commits:
- Updates some libraries to latest officially supported versions:
** Numpy: 1.17.5
** OCIO: 1.1.1
** OIIO: 2.1.15
** OSL: 1.10.10
** OIDN: 1.2.1
- Re-enables some distro packages (like OSL, OIIO, OCIO...).
- Add missing 'CMake cleanup commands' for generated CMake update
command, for Embree, OIDN and OpenXR.
- Generalizes using min/max versions of accepted libraries, if no
package can be found in specified range then it is built from sources.
The later point should help keeping things a bit in better conditions,
although current maximal accepted versions are somewhat arbitrary guess
currently.
The preview of points was only done when the edge is wire.
Now the preview of points is done when the edge is not connected to any
quad.
Also to avoid edge slide in this case, all new vertices created in this
specific case are not selected.
Differential Revision: https://developer.blender.org/D7457
Is not only the values of translation/scale/rotation which are to be inverted,
but also the order of operations.
Differential Revision: https://developer.blender.org/D8518
This reverts commit 9adedb2605. It changes
how duplis work, and by that altered how Alembic and USD files are
written. This was signalled by a failing Alembic unit test.
Now the operators are split to define different default values. This is also required for the future Bezier primitive tools.
This allows to show in the Topbar the number of subdivisions. Before this value was totally hidden and it was number of Edges. The wheelmouse can be used to override the value while running, but does not change the default value.
{F8766270}
All operators share same code.
Also, fixed some bad practices done with Toolbar in python.
Reviewed By: mendio, pepeland
Differential Revision: https://developer.blender.org/D8506
Mainly caused by moving tests to source folder, which brings
new compiler flags.
Also, no need to extend include directories (to the non-existing paths,
btw) since this is done by the CmakeList.txt which is in the blenlib folder.
Solves possible pointer-based comparison fiasco.
Another nice outcome of this is that topology cache will now be
preserved throughout the undo system. For example, undo of object
transform will not require topology cache to be re-created.
Differential Revision: https://developer.blender.org/D8493
The file was including BKE. It was a required fix for another
compilation error a while back. It is a bad level include, and
seems is not needed anymore.
The problem here is that the baking code uses tiles to exchange pixel data with
the renderer since a recent-ish refactor, but the code that sent data to the
renderer did not initialize the bake result pixels.
Therefore, when the baking process for the second object started, Cycles
received empty tiles and sent them back as-is if the second object did not
cover them.
By initializing the tiles with the result of the previous bakes, we avoid this
problem.
We can't exactly follow what we do for macOS here. On Windows special
characters can be inserted with Ctrl+Alt. So make sure we expect UTF-8
characters when Alt is held.
Mistake in 87062d4d67.
Since C++17 there is also std::string_view, which is similar to StringRef.
This commit does a couple of things:
* New implicit conversions between StringRef and std::string_view.
* Support std::string_view in blender::DefaultHash.
* Support most of the methods that std::string_view has.
* Add StringRef::not_found which can be used instead of -1 in some places.
* Improve/fix the description at the top of BLI_string_ref.hh.
And make them part of the blender_test runner. The one exception is blenlib
performance tests, which we don't want to run by default. They remain in their
own executable.
Differential Revision: https://developer.blender.org/D8498
This adds a property that checks the normals of each vertex against the
view direction to decide if they should be masked (similar to the
"Front Faces Only" option works for brushes.
Reviewed By: sergey
Maniphest Tasks: T77637
Differential Revision: https://developer.blender.org/D8448
This brush includes a set of deformation modes designed to deform and
control the shape of the mesh boundaries, which are really hard to do
with regular sculpt brushes (and even in edit mode). This is useful
for creating cloth assets and hard surface base meshes.
The brush detects the mesh boundary closest to the active vertex and
propagates the deformation using the brush falloff into the mesh.
It includes bend, expand, inflate, grab and twist deform modes.
The main use cases of this brush are the Bend and Expand deformation
modes, which depend on a grid topology to create the best results.
In order to do further adjustments and tweaks to the result of these
deformation modes, the brush also includes the Inflate, Grab and
Twist deformation modes, which do not depend that much on the topology.
Grab and Inflate are the same operation that is implemented in the
Grab and Inflate tools, they are also available in the boundary brush
as producing deformations with regular brushes in these areas is very
hard to control.
Even if this brush can produce deformations in triangle meshes and
meshes with a non-regular quad grid, the more regular and clean the
topology is, the better. Most of the assets this brush is intended to
deform are always created from a cylindrical or plane quad grid, so it
should be fine. Also, its algorithms can be improved in future versions
to handle more corner cases and topology patterns.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8356
Proper handling of View Layers for the VR session was never implemented.
Now the View Layer of the VR session follows the window the session was
started in.
Note that if this window is closed, we fallback to another window. This
is done to avoid the overhead it would take to maintain a separate
depsgraph for the VR view. Instead we always share some already visible
View Layer (and hence the depsgraph).
The problem is caused by a lack of prediction in the `isect_line_segment_tri_v3`
that incorrectly confirms some intersections of coplanar segments to the triangle.
The solution is to use another algorithm to detect intersections.
This also resulted in a slight improvement in the performance:
- 1min 17sec to 1min 6sec in my test file
Differential Revision: https://developer.blender.org/D8500
Root of the issue was not fixed in 2.90, only hidden by the fact that we
now re-read much less data during undo's that we used to, when some new
datablock gets added or removed.
This is not an ideal solution (as usual when dealing with data pointers
shared across data-blocks), but it's decent enough. thanks a lot to
@brecht for it!
To be backported to 2.83 too.
We want the session to start exactly at the landmark position, with
no additional offset. Some runtimes (e.g. Windows Mixed Reality) may
give an initial non-[0,0,0] position at session start though.
Also add a comment explaining the purpose of the eye offset variable.
There would always be an unintended offset applied. Per design there
should not be any offset when changing VR Landmarks, the view should
just jump exactly to the Landmark.
Due to the recent changes, we don't have to add, but substract the eye
offset we apply to get the wanted behavior.
Mistake in 607d745a79.
We want the session to start exactly at the landmark position, with
no additional offset. Some runtimes (e.g. Windows Mixed Reality) may
give an initial non-[0,0,0] position at session start though.
Also add a comment explaining the purpose of the eye offset variable.
There would always be an unintended offset applied. Per design there
should not be any offset when changing VR Landmarks, the view should
just jump exactly to the Landmark.
Due to the recent changes, we don't have to add, but substract the eye
offset we apply to get the wanted behavior.
Mistake in 607d745a79.
This does not fix all the cases in the bug report, because there are multiple
different issues. Only the first two are fixed. The third is probably a known
issue for now.
Before this patch, the rigid body simulation was always done after modifiers
are evaluated, because to perform the simulation, the final geometry of the
object was required. However, the geometry is not required in all cases,
depending on the selected collisions shape.
This patch changes it so that when the simulation does not need the
evaluated geometry, the simulation will be done before the modifiers
are evaluated. This gives the modifiers access to the simulated positions.
When the rigid body simulation does depend on the evaluated geometry,
it will still be performed after modifiers are evaluated.
The simulation will be performed after modifiers are evaluated, iff
the collision shape is "Convex Hull" or "Mesh" and the source is set
to "Deform" or "Final".
Reviewers: sergey
Differential Revision: https://developer.blender.org/D8487
In last set of refactoring patches, code implementing this feature
has been accidentally removed.
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D8449
This just set a global object responsible for allocating new objects
in a thread safe way without needing any GPUContext bound to this
thread.
This also introduce the GLContext which will contain all the GL related
functions for the current context.
# Conflicts:
# source/blender/gpu/intern/gpu_context.cc
This just set a global object responsible for allocating new objects
in a thread safe way without needing any GPUContext bound to this
thread.
This also introduce the GLContext which will contain all the GL related
functions for the current context.
For the main rationale behind this design, see 49f088e2d0. Further,
this removes users of uiBut.a1, which is a very ugly design
choice (hard to reason about).
Part of T74432.
Fix package name missmatch in a few module files. IE "ALEMBIC" was
defined in the file but the find_package commands used "Alembic"
Some modules state that they set and use the _LIBRARY variable but the
do in fact not do this. Removed these comments from those files.
Enable Clang-Tidy's `readability-function-size` rule and add a few
`NOLINT` markers to explicitly silence warnings for three functions.
These functions are huge and would IMO benefit from splitting up, but
are hard to without intimate knowledge of the code.
At least by enabling the rule, we can start tweaking the values and
refactoring other functions that bubble up as being too long/complex.
No functional changes.
Add a `NOLINT` marker to explicitly silence a warning from Clang-Tidy's
`readability-function-size` rule for the `node_type_base()` function.
This function is indeed huge, but that is because a lot of macros are
expanded. Before expansion things are still not small, but still
understandable & expandable.
No functional changes.
In the 3D View, the "Animation" keymap is not used, but the mode
specific ones. So the shortcut editing code should use these too, just
like the default keymap does.
This adds a `NOLINT` marker to explicitly silence a warning from
Clang-Tidy's `readability-function-size` rule for the `incircleadapt()`
function in `delaunay_2d.c`.
No functional changes.
This adds `NOLINT` markers to explicitly silence warnings from Clang-Tidy's
`readability-function-size` rule for versioning functions. Technically
these could be refactored and split up into smaller bits, but generally
they are hardly ever looked at once they're a few releases old.
No functional changes.
This addresses warnings from Clang-Tidy's `readability-function-size`
rule in the `source/blender/python` module.
It's just `BPyInit_bgl()` that's been split up into one or two smaller
functions per OpenGL version.
No functional changes.
For the main rationale behind this design, see 03b122e2a18df. Further,
this removes users of `uiBut.a1`, which is a very ugly design
choice (hard to reason about).
Part of T74432.
For the main rationale behind this design, see 03b122e2a18df. Further,
this removes users of `uiBut.a1`/`uiBut.a2`, which is a very ugly design
choice (hard to reason about).
Part of Part of T74432.
The current on-size-fits-all `uiBut` creates quite a mess, where it's
hard to reason about which members are free for use, under which
conditions they are used and how.
`uiBut` also has members that aren't used at times, violating the "don't
pay for what you don't use" principle.
To address this, we want to move to typed buttons, where `uiBut` is just
a base struct and each type extends it as needed. That structures data
better and type specific data is only available if it's actually used by
a button type.
Two trade-offs:
* Many casts to the derived type have to be done.
* Sometimes we change the button type after it's created. So I had to
add logic to reallocate the button for use with the new, possibly
derived struct. Ideally that wouldn't be needed, but for now that's
what we have.
Part of T74432.
Differential Revision: https://developer.blender.org/D7610
Reviewed by: Brecht Van Lommel, Campbell Barton
This addresses warnings from Clang-Tidy's `readability-else-after-return`
rule. This should be the final commit of the series of commits that
addresses this particular rule.
No functional changes.
This addresses warnings from Clang-Tidy's `readability-else-after-return`
rule in the `source/blender/draw` module. Not all warnings are addressed
in this commit.
No functional changes.
Clicking on the column header is supposed to enable sorting by this
column, or switch the sort order if already enabled.
The double-click event would be blocked by the `file.execute()`
operator, which is not supposed to act if the user clicked outside the
file list.
This addresses warnings from Clang-Tidy's `readability-else-after-return`
rule in the `source/blender/blenlib` module. Not all warnings are
addressed in this commit.
No functional changes.
This replaces header include guards with `#pragma once`.
A couple of include guards are not removed yet (e.g. `__RNA_TYPES_H__`),
because they are used in other places.
This patch has been generated by P1561 followed by `make format`.
Differential Revision: https://developer.blender.org/D8466
Follow path seems to not be catched by `BKE_object_moves_in_time`.
For this reason, we cache all transforms for all object and check
ourselves if an animation occurs. This is almost what cycles does.
We also fix the rigid body case if the rigid body use deformation.
The Pose FK mode assings the rotation origin to the boundary of the last
visited face set in the floodfill operation. In some cases, the topology
of the model may make the flood fill operation to visit a face set as the
first one (assinging it to target) and visit it again as the last one
(assinging it to origin). This will make the pose brush to default the
origin and target to the brush location and not to the face sets as it
considers that there is only one possible boundary.
This adds a GSet to ensure that a particular face set is not visited
twice in the flood fill, fixing these cases.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D7984
This makes possible to choose between a local and a global simulation
when the cloth brush is used. Local simulation is the current default.
When global simulation is enabled, the cloth brush simulates the entire
mesh without taking any simulation limits into account.
This was possible before by setting the simulation limits to 10 (the
current maximum value allowed) so the entire mesh was inside the limits,
but this was a hack as the limits scale with the radius and there should
not be any limitation on how big the simulated area can be to be able to
simulate an entire object. This also allows to make a more clear
distinction between cloth brush presets that are intended to be used in
local areas to add details or globally to generate the base shape of the
mesh.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8481
This patching duplicates the first frame of the layer if the first frame number is not equals to the scene first frame number.
Related to T79567
Differential Revision: https://developer.blender.org/D8486
Some minor cleanup in the patch.
The paint_draw_cursor function was handling the cursor drawing for 2D
and 3D views of all paint modes, calculating the brush radius, updating
the SculptSession data and updating and drawing all sculpt cursor
overlays for different tools. It was almost impossible to understand when
and what was being drawn and in which state the GPU matrix was.
Now everyting is organized into different functions, with clear
separation between modes, sculpt tool overlays and different drawing
setups. Update and drawing functions are also separated (this allows to
skip one PBVH query on each cursor drawing).
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8206
The current code is accessing this from outside the gpu "namespace". As
such it should be accessing GPU_ functions, not gpu_ functions.
This is also a place to centralize the XXX message that will be
addressed upon refactor. So we can reuse this call in other places that
need the same temporary workaround.
Groundwork for upcoming fix (D8472)
In an upcoming bugfix we'll use OpenVDB data structures directly to build mesh
for sparse OpenVDB volumes, loading them OpenVDB grids earlier and removing any
references to Blender data structures makes that easier.
This also makes changes to Blender volumes to support this, so Cycles can take
ownership of a grid without Blender having to keep its own reference to it.
This should also be useful in a future Python API.
Ref D8401
Small tweaks to make labels and texts more correct, consistent and
polished.
Reviewed by: Aaron Carlisle, Julian Eisel
Differential Revision: https://developer.blender.org/D8346
Having includes in debug builds makes it possible to accidentally
break release builds.
Avoid this by moving calls to other modules out of BLI_assert.h
into BLI_assert.c
When adding multiple force fields of different types they are all called
"Field", making it difficult to tell them apart.
Lights were already named based on their type. This follows the light
code.
New names:
- Force
- Vortex
- Magnet
- Wind
- Guide
- TextureField
- Harmonic
- Charge
- Lennard-Jones
- Boid
- Turbulence
- Drag
- Fluid
- Field
Reviewed by: Julian Eisel
Differential Revision: https://developer.blender.org/D8420
Similar to T58668, labels were not aligned when multi-editing widgets
that are not center-aligned.
Reviewed by: Hans Goudey, Julian Eisel
Differential Revision: https://developer.blender.org/D8441
Align items in the edit mesh context menus (reducing padding), for
consistency with other menus.
The root layout of menus doesn't add the padding, for sub-layouts
`align` has to be enabled.
{F8749633}
Reviewed By: Julian Eisel
Differential Revision: https://developer.blender.org/D8480
This seems to be caused by a change to the logic of movieclip_get_gputexture_ptr in rB97b597c.
Differential Revision: https://developer.blender.org/D8469
This implements collisions in the solver of the cloth brush/filter. It
uses the scene colliders as a regular physics simulation.
There are still some parameters (friction, distance to the surface...)
that can be exposed as properties in later patches.
Thanks to Sebastian Parborg for helping me with the implementation.
Reviewed By: sergey, zeddb
Differential Revision: https://developer.blender.org/D8019
The panels are rebuilt when a modifier is removed so the button handlers need
to properly finish. By adding a context argument to the panel_delete function
this will happen properly.
This adds a curvature smoothing and intensify details properties to control
the result of the Sharpen Mesh Filter.
Curvature smoothing removes high frequency details from the precalculated
sharpen data, so the filter result has much smoother surfaces and cleaner
sharpen lines;
Intensify details displaces the vertices of creases and valleys in the direction
opposite to its neighbors average, so it intensifies high frequency details
in those areas, producing more noisy and sharp shapes:
Both this properties can be used in combination to achieve a good balance of
high and low frequency details depending on the shape and the desired result.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8447
Previusly it was using the / character, which implies that there is one
deformation mode and and invert deformation mode like in the other two
deformation, but squash and stretch is only one deformation mode and
does not have an invert mode.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8463
The scale deform mode includes rotation by default, so when when scaling
down a part of the models it becomes harder to control as the effect of
the rotation less predictable (similar to using trackball rotation in a
very small radius). This locks the rotation of the segment, so parts of
the model can be scaled down in a more predictable way.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8465
Steps to reproduce were:
* Open File Browser
* Create a new directory
* Cancel renaming with Esc
File selection flags were modified during drawing when the rename button
got removed, but the file name label drawing wasn't checking the
modified state.
The file.execute() operator is the one that actually opened directories
and files, not file.select() with the "open" option, as it was assumed
when changing the keymap to double-click for opening. It only acts on
the current selection though, so we have to ensure the selection is set
on the first click.
Now, some touch-pads have a delay until they register a click event, so
the double-click would be registered instead, before the selection is
set. Always select on mouse-down now and remove the unnecessary select
operator call on double-click.
The crash is caused by the fact that a NULL Object pointer is passed to
calculate the transform orientation, which has been set to normal.
A check has been include to detect the same.
Differential Revision: https://developer.blender.org/D7951
Use `td->iloc` as the coordinates of the transformed element.
It is more accurate and other transformation modes, such as scale, also
operate on `td->iloc`.
The icons are label buttons. Usually these are not editable and can not
become active. These are draggable ones though (so dragging files can be
dragged by dragging the icon) which creates an exception to this rule.
So hovering the icon would activate its label and when executing the
rename operator via shortcut it wouldn't get exited properly. This broke
the invariant of only allowing a single active button at a time.
Added an assert to check that invariant now.
Letting the code to activate the text button ensure any currently active
button is exited seems sensible.
When disconnecting links for defaulted node group inputs, recurse
into the nested node group nodes, instead of checking the socket
flag. Otherwise the behavior is confusing and differs from Cycles.
Differential Revision: https://developer.blender.org/D8455
Disabled buttons would incorrectly toggle state when a drag toggle
passed over them. This adds a check to prevent a drag toggle on disabled
buttons.
Differential Revision: https://developer.blender.org/D8476
The `rna_Scene_ray_cast()` function tried to find the current depsgraph. To
this end, it required the scene, the view layer, and bmain. Scene has a cache
of per-view-layer depsgraphs, to speed up switching between view layers. This
cache does not contain render depsgraphs, and evaluated view layers also don't
have a depsgraph here.
When a suitable depsgraph cannot be found, a new depsgraph is created. However,
this depsgraph is not evaluated, and has an unexpanded scene pointer with a
`NULL` `view_layer`. Using this then crashes Blender. Also, there was no way
for the code to get the render depsgraph.
The solution is to pass the depsgraph to the `ray_cast()` function, instead of
the view layer. This avoids the depsgraph lookup, and also works correctly when
rendering.
Some add-ons also need updating, which I'll do in the `addons`
repository soon.
Reviewed By: Sergey
Differential Revision: https://developer.blender.org/D8475
This uses the vertices per grid instead of quads to set the limit of
grids per PBVH Node. This should create more leaf nodes in lower
subdivisions levels where the duplicates count is high, producing more
uniform performance across different levels.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8454
There were two issues. First, I made a mistake when I switched from unsigned
to signed integers. Second, two classes with the same name were defined in
separate files. Those classes are in an anonymus namespace now, so that they
don't leak into other files.
ad4928a171 disabled alignment for too many cases. Still try to avoid
aligning many items, to avoid thousands of redundant alignment
calculations. But now we're much more picky adding an sub-row with
alignment.
On windows, spacebar would be passed as UTF-8 text input, despite the
control key being pressed. On macOS, there already was an explicit
exception for this (command key in this case), on Linux XInput already
handled this case for us.
Note that Alt should still allow text input, for special character
sequences.
Issue also happened in the Text Editor if a text data-block was set.
Optimization was disabled in this function to work around a bug in MSVC, use
a different solution that does not come with such a big performance regression.
In Blender 2.90 EEVEE materials were refactored that introduced crashes on Intel
GPUs on Windows. The crash happened in the `local_context_workaround` that temporary
stored compiled materials in a binary form to reload it in the main GL context.
It has been tested that the workaround isn't needed anymore for HD6xx GPUs, but it
is still needed for HD4000.
After several unsuccesfull fixes we came to the conclusion that we could not support
the local context workaround and needed to come with a different workaround. The idea
of this patch is that in these cases there is only a single context that is used for
rendering. Threads that uses these contextes are guarded by a mutex and will block.
Impact on User Level:
* Due to main mutex lock the UI freezes when rendering or baking or feel less snappy
Reviewed By: Clément Foucault, Brecht van Lommel
Differential Revision: https://developer.blender.org/D8410
As far as I can see, this problem was introduced with the gpu_free_unused_buffers() changes in rB97b597c.
The code checks the returned pointer to stop looping, but the last iteration will return the last GPUTexture and set the queue to NULL, meaning that the next pop() will crash.
Differential Revision: https://developer.blender.org/D8468
Some implementation have different maximum texture size.
This patch avoid crash when texture allocation fails when:
- trying to bake a lightcache too big for the OpenGL imeplementaion.
- loading a cache from file that is too big for the OpenGL imeplementation.
Face Sets where only set and updated on the PBVH after starting a sculpt
tool. In order to preserve the visibility they store when changing
levels, they need to be updated and sync also on PBVH creation
Reviewed By: sergey
Maniphest Tasks: T78665
Differential Revision: https://developer.blender.org/D8225
Duplicates of a grid corner adjacent to an edge which are on the
adjacent grid of the same face were not added when requested.
Needed for D8356 to work, it may also fix some other bug in Multires.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8385
The cloth brush has a defined simulated area with a falloff. In the falloff
area (the area between the dashed white circle and the exterior white
circle), simulation properties change in order to fade out the
simulation deformation effects towards the boundary.
With some brushes and stroke types (like anchored strokes with pinching
or grabbing with full strength), it is possible to apply more force than
what the boundary falloff can compensate, so the simulation breaks when
this happens.
This option pins the falloff area with softbody constraints, This
produces a much better deformation falloff and it is no longer possible
to move the vertices near the simulation boundary, so the simulation
won't break no matter the strength of the forces applied inside the
simulated areas.
This is an option as it is particularly useful for some brushes to add
localized details, but for brushes that are supposed to deform the
entire mesh (like the grab brush in D8424), this can add unwanted
softbody constraints that affect the simulation result.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8435
Box mask is not a selection, so it should not be part of the select
operator. This allows to add more sculpt mode specific functionality and
properties and to share more code with the lasso mask operator in a
later refactor.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8456
When switching workspaces we need to have an unused screen layout that
we can activate. The other window now showed the only available screen
layout in fullscreen though.
Usually when there's no unused screen layout we duplicate an existing
one, but that code didn't respect the fullscreen case properly.
This also tries to clean up the logic a bit, but things are still rather
complicated to follow.
Changes in this code are always risky. Of course things worked fine in
my tests, but I wouldn't be surprised if something breaks.
Filtering Collections when also filtering object children would only
display the parent object. Add a check for the NO_CHILDREN filter before
creating the object-parent hierarchy.
The winding order of the faces changes when flipping the faces.
This lead to the loop indices changing as well.
Now we take this into account when restoring and flipping the custom
normals. Before the normals would be swapped.
The winding order of the faces changes when flipping the faces.
This lead to the loop indices changing as well.
Now we take this into account when restoring and flipping the custom
normals. Before the normals would be swapped.
The problem in this case was that the flag for active fire was not set. With hidden flow sources the flag was not updated in update_flowsflags().
The solution for this is to take the active field from the config cache file.
Previously a for loop with two iterations was used to calculate the
3D segment locations for the input number and the higher power of 2.
Splitting off the inside of the for loop to a separate function makes
the code more readable.
This commit also includes a simple timer for bevel, enabled with a
define. Interestingly, the cleanup in this commit happended to give
a 3% speedup on a Ryzen 3700x for a bevel calculation with 64
segments.
Same exact system as instancing on vertices for a mesh. More powerful ways
of instancing will be added at some later point, but this makes the basics
works and is consistent with other geometry types.
The tests were broken because of an extra call to
RNA_def_struct_name_property which set the name of the velocity
attribute to be the string used to identify CacheFiles in
bpy.data.cachefiles.
When changing the material while the properties editor temporarily isn't
visible (e.g. because another editor is in full-screen or a different
workspace is active), the preview wouldn't be updated on changes.
Always trigger a material preview update on screen layout or editor type
changes.
When changing the material while the properties editor temporarily isn't
visible (e.g. because another editor is in full-screen or a different
workspace is active), the preview wouldn't be updated on changes.
Always trigger a material preview update on screen layout or editor type
changes.
The data member `new` was conflicting with the `new` keyword
when `BKE_screen.h` was included in C++ files.
Reviewers: sergey
Differential Revision: https://developer.blender.org/D8459
Caused by rB63ee3db96107.
While above commit corrected the clip offset, it also removed logic to
ensure a marker on a particular frame. This is needed though, otherwise
changes on a particular frame are applied to the marker being returned
by 'BKE_tracking_marker_get' which can be a completely different marker
if none exist for that frame yet.
This patch partly reverts rB63ee3db96107 and reintroduces the framenr
for the MarkerUpdateCb and uses that to ensure a marker on that frame.
Candidate for backporting to 2.83 LTS?
Reviewers: sergey, jacqueslucke
Subscribers:
Before: If the current frame is out of the cache start/end range, the viewport will show the fluid as it was on the last frame that was still in the cache frame range.
Now: If the current frame is out of the cache start/end range, the viewport will show no fluid at all (even if there are cache files present for this frame).
This fix is related / in response to T79423.
The decomposed transform would have consists of nan values if the input
transform had zero scale.
Now the decomposition will check for zero scale, and if it is detected
then the result will be ensured to be finite. Additionally, rotation
value will be copied from previous/next time step to help avoiding
obscure interpolation.
The latter step can become more comprehensive than the current simple
implementation.
Differential Revision: https://developer.blender.org/D8450
This will make it easier & cleaner to make custom-built depsgraph (for
example for exporting invisible objects to USD or Alembic, see T75936).
No functional changes.
Reviewed by: sergey
Differential Revision: https://developer.blender.org/D8423
This flag specifies that even when the socket is not connected,
the node should not display the input field for the constant input
value. This is useful for inputs like Normal, which have special
handling for the missing input case and don't use a constant value.
Currently there is no way to change this flag from Python, and
through UI it can only be done by re-creating the socket.
This patch exposes the flag through RNA and UI, makes sure it
is properly updated when changed, and adds special handling to
ensure that it is correctly set when creating a node group from
a node set that includes reroute nodes.
Differential Revision: https://developer.blender.org/D8395
Instead of assuming the node width, compute the actual right bound.
Otherwise BSDF nodes tend to be too wide, and intersect Output.
Differential Revision: https://developer.blender.org/D8395
Avoid the conversion for the hook modifier as riggers
may need to enable this modifier in edit-mode.
The speedup of the modifier alone is significant since the hook
operation is simple compared to conversion that took over 99.9%
of the time in my tests, however the overall speedup was around to 1.6x.
The bevel code initialized the CurveProfile with the "higher power of 2"
segments after the normal number of segments, leaving the widget in an
incorrect state after the calculation. A simple fix is to re-order the
initializations, doing the input number second.
'Default' actually means 'Keep Existing' and it makes more sense to
display the more descriptive label.
Differential Revision: https://developer.blender.org/D7614
Fixes T78346.
The shortcut display and change code is context sensitive. To make it
work correctly the context needs to be set properly.
When executing operators from the dropdowns, the active region is the
header, but the shortcut handlers are set for the main region. So make
sure that is used instead.
This also sets the main region active for context menu operators, where
this issue shouldn't be present. Doing it anyway shouldn't hurt though
and fixes this issue in case somebody displays the context menu in the
header as dropdown too.
Instead of clearing forces at the end of the simulation step, they will now be cleared before writing to them, i.e. at the beginning of a step.
Also cleaned up minor areas that I looked at while making this change.
Following work done in 2.83, the resolution control is now a real
level-of-detail parameter. It is now useful to be able to set the
resolution for display independently from render. This is true for
both mesh generation and mesh deformation modes.
For compatibility with old scenes, resolution is retained and is the
render resolution. Old modifiers loaded have the value of resolution
also applied to viewport resolution. This allows newer modifiers to
be used in older versions without trouble
Differential Revision: https://developer.blender.org/D8336
The text colors set by the general widget state function
(`widget_state()`) would always be overriden by the menu-back text
colors to avoid contrast issues. This would only respect the selected
state, not other states.
Address this now by changing the input theme colors to use the menu-back
ones, rather than overriding after the fact (calling `widget_state()`).
Changing the surface distance through the flow type is inappropriate here. It had been added to ensure that liquids / smoke use a different emission value.
Now the value will only be changed when changing from a gas to a liquid emitter or, vice-versa, when changing from a liquid to a gas emitter.
There is a parameter typo in Python API document about bpy.props.StringProperty.
This patch fixes this.
Also this patch should apply to 2.90 branch as well.
Reviewed By: Grische, mont29
Differential Revision: https://developer.blender.org/D8430
Code required the tabs to be placed in an aligned region. Code should
work fine even for unaligned regions though, so I don't see a reason to
forbid this.
This patch adds the ability to render motion blur from Alembic caches.
The motion blur data is derived from a velocity attribute whose name has
to be defined by the user through the MeshSequenceCache modifier, with a
default value of ".velocities", which is the standard name in Alembic
for the velocity property, although other software may ignore it and
write velocity with their own naming convention (e.g. "v" in Houdini).
Furthermore, a property was added to define how the velocity vectors
are interpreted with regard to time : frame or second. "Frame"
means that the velocity is already scaled by the time step and we do not
need to modify it for it to look proper. "Second" means that the unit
the velocity was measured in is in seconds and so has to be scaled by
some time step computed here as being the time between two frames (1 /
FPS, which would be typical for a simulation). This appears to be
common, and is the default behavior.
Another property was added to control the scale of the velocity to
further modify the look of the motion blur.
Reviewed By: brecht, sybren
Differential Revision: https://developer.blender.org/D2388
The hardcoded age limit is now gone. The behavior can be implemented
with an Age Reached Event and Kill Particle node. Other utility nodes
to handle age limits of particles can be added later. Adding an
Age Limit attribute to particles on birth will be useful for some effects,
e.g. when you want to control the color or size of a particle over its
life time.
The Random Float node takes a seed currently. Different nodes will
produce different values even with the same seed. However, the same
node will generate the same random number for the same seed every
time. The "Hash" of a particle can be used as seed. Later, we'd want
to have more modes in the node to make it more user friendly.
Modes could be: Per Particle, Per Time, Per Particle Per Time,
Per Node Instance, ...
Also a Random Vector node will be useful, as it currently has to be
build using three Random Float nodes.
Find free slot first for sound strips then for movie strips.
This patch also fixes issue where all strips were added to channel 2 by default.
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D8339
As creator.c is used for the 'main' function,
avoid obscure declarations being placed so prominently.
- Move Blender as a Python module declarations into it's own section.
- Move USD declaration to an 'extern' just before it's called.
Previously, there were only particle-birth and time-step events.
Now the solver can handle custom events. On the user level
this does not change anything yet. This feature of the solver
will be used by an upcoming Age Reached Event node and
possibly others. When this node exists, I can finally remove
the hardcoded maximum particle age.
The abbreviation 'init' is brief, unambiguous and already used
in thousands of places, also initialize is often accidentally
written with British spelling.
Yet another update to ensure that pointer variables are always up to date, i.e. those referencing the memory allocated by Mantaflow. Outdated pointers usually show up through flickering in viewport.
The weight must be added only to the real points, not to the autogenerated points by modifiers. This affects, not only to subdivide, but to any modifier that generate points.
In f2b04302cd the pointer update was refactored. It was sufficient to update pointers just in replay mode at the end of a step since the 'ensure()' functions from manta_fluid_API.cpp had their own pointer update call. These were removed in 51f4bee5a5, however, and so in order to still have some sort of update, the given update call needs to be available to all cache types.
Required for the new boolean code, disabled by default
until all platforms have landed the libs and the boolean
code actually lands in master.
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D8384
Caused by c7aa0f9d74.
Since above commit, BKE_image_user_frame_calc requires an image (not
just the iuser) to get the framenumber.
Cycles used to call this with NULL image (in `image_user_file_path` and
`image_user_frame_number`), now pass the image as well.
Maniphest Tasks: T79111
Differential Revision: https://developer.blender.org/D8439
A collection of multiple changes that had been living in my stash:
- Use nullptr instead of NULL in C++ files
- Removed unused/deprecated functions from headers
- Cleared animatable flag in cache UI
- Comment cleanups
Use static_cast() where possible and refresh pointers for every frame when in replay mode. The latter is particularly important as this seems to have caused the issue where smoke in the viewport was flickering when loading data from pointers after loading them from disk for the frame before (e.g. when resuming a bake job).
Add our own copy of the gtest discovery scripts from CMake a few reasons:
* Use the very latest version which supports PRE_TEST for Windows
* Fix usage of [] symbols in file paths that fail with the zsh shell
* Disable asan leak checker when discovering tests
This means Windows also no longer requires the very latest CMake 3.18.
The root of the issue comes to the fact that sub-data pointers were
used to match strips before/after copy-on-write. The undo system might
re-use sub-data pointers after re-allocating them, making it so that,
for example, pointer used by sound strip is later re-used by video
strip.
This fix takes an advantage of recently introduced per-sequence UUID
and uses it to match sequences before/after copy-on-write.
Will trigger code paths which makes sure UUIDs are generated and
are unique.
Enabled with --debug-depsgraph-uuid (which is also implied by
--debug-depsgraph).
This is the first step for having sequences covered with session UUID
with the goal to remove code which uses original sequence pointer to
match sequences.
Currently this UUID is maintained on file load, allocation and leaf
duplication function.There are more cases to cover and ensure UUID
is re-generated or re-used when needed. It will be done as follow-up
development.
Allows to use pre-defined structure for session UUIDs in all data
structures which needs it: pose channels, sequencer strips, modifiers.
The goal of all this is to have a reliable way of matching original
and copy-on-written versions of data, so that it's possible to
preserve runtime caches.
button
This was reported for the FCurve modifier restrict ranges, but might fail
elsewhere, too. Reason is that the post_but has its range (hardmin/
hardmax etc) set before the updates to the active button take place, so
changes here dont end up on the post_but (even though the RNA range
function is properly called for a new defined button - new one is not the
same as the post_but though).
Now update the ranges on the post_but when that gets active.
Fixes T78763
Maniphest Tasks: T78763
Differential Revision: https://developer.blender.org/D8265
These were added in rB146473f08335e8cb774ccaf1baad82a1d308fbe1 however
there were a few errors with the labels:
1. Underscore in label
2. Abbreviation when not needed
This bumps the minimum requirement for cmake from 3.10 to 3.18 on windows
if `WITH_GTESTS` is enabled.
Reviewed By: sergey brecht sybren campbellbarton
Differential Revision: https://developer.blender.org/D8405
Collection instance datablocks were not filtered out when only showing
the selected object. They were treated as a collection (which should
show when filtering objects). Adds a case to check if the parent is an
object.
This patch adds a new compound shape entry to the shape selection
dropdown. It also corrects wrong inertia calculation for convex hulls,
that resulted in strange behavior for small objects.
The compound shape take the collision shapes from its object children
and combines them. This makes it possible to create concave shapes from
primitive shapes. Using this instead of the mesh collision shape is
often many times faster.
Reviewed By: Sergey, Sebastian Parborg
Differential Revision: http://developer.blender.org/D5797
Description of `USER_HEADER_SEARCH_PATHS` build setting:
"
This is a list of paths to folders to be searched by the compiler
for included or imported user header files (those headers listed
in quotes) when compiling C, Objective-C, C++, or Objective-C++.
Paths are delimited by whitespace, so any paths with spaces in
them need to be properly quoted. See Always Search User Paths
(Deprecated) (ALWAYS_SEARCH_USER_PATHS) for more details
on how this setting is used. If the compiler doesn't support the
concept of user headers, then the search paths are prepended to
the any existing header search paths defined in Header Search
Paths (HEADER_SEARCH_PATHS).
"
http://help.apple.com/xcode/mac/current/#/itcaec37c2a6
Xcode doesn't use `HEADER_SEARCH_PATHS` for auto-complete. Only the
header files in the same directory as the current file are suggested.
CMake as of now correctly sets `SYSTEM_HEADER_SEARCH_PATHS` and lumps the
rest in `HEADER_SEARCH_PATHS`. The standard way is to use
`USER_HEADER_SEARCH_PATHS` & `SYSTEM_HEADER_SEARCH_PATHS` and let
`HEADER_SEARCH_PATHS` be used as a fallback for compilers which do not
distinguish between `<*.h>` and `"*.h"` syntax.
So set `USER_HEADER_SEARCH_PATHS` to the include paths specified
in the `CMakeLists.txt` files of all targets.
Caused by {rBa1a333a1e92e}
`BKE_sequencer_prefetch_get_original_sequence` cant get a sequence in
`BKE_sequencer_cache_put`.
rBa1a333a1e92e moved multiple NULL checks (including the one checking
seq) above BKE_sequencer_prefetch_get_original_sequence (this should
never return NULL really, but this is for another fix).
So solution suggested by @ISS is to just stop prefetching all together
before changing content of seqbase.
Maniphest Tasks: T79357
Differential Revision: https://developer.blender.org/D8421
- add the use of DRWShaderLibrary to EEVEE's glsl codebase to reduce code
complexity and duplication.
- split bsdf_common_lib.glsl into multiple sub library which are now shared
with other engines.
- the surface shader code is now more organised and have its own files.
- change default world to use a material nodetree and make lookdev shader
more clear.
Reviewed By: jbakker
Differential Revision: https://developer.blender.org/D8306
Previously, deform brushes were modifying the final positions in the
simulation directly, which was causing all sorts of artifacts in the
deformed area and problems with other features of the solver.
Now these brushes deform a separate array of positions and the solver
adds constraints to them, so the real vertices are moved when solving
the constraints. This prevents those artifacts and gives the brush a
much better behavior.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8424
The cloth brush builds the constraints when the stroke starts usign the
current state of the mesh. This means that deformations profuced by the
simulattion will accumulate after multiple strokes as it will always
start from the previous deformed state. While this is useful in many
cases, for other uses it is convenient to always simulate the same
initial shape, but applying different forces to it.
The persistent base options work like the persistent base in the layer
brush and allows the cloth brush to not accumulate deformation after
each stroke. When enabled, constraints are created for the shape stored
in the persistent base instead of from the current state of the mesh.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8428
Fix several issues in CLOG code:
* In `clg_str_reserve`, allocated memory may be bigger than requested
one, do not assign the latter back to `cstr->len_alloc`.
* `clg_str_vappendf` was mis-interpreting returned value from
`vsnprintf`, and completely mixing total allocated memory and extra
needed amount of memory to allocate...
Simplified code of `clg_str_vappendf` to only have allocating code
handled in one place, makes things easier to follow too.
Think this should also be beckported to 2.83.
This is part of T79273 where separating a stroke would result in
multiuser gpencil data for the original object.
Real underlying issue seems to be that gpencil multiuser objects (also
the ones created by Alt+D duplicating) have a problem evaluating
modifiers correctly [this will need further investigation].
Not sure if this is a limitation of D5470?
This patch only corrects the usercount on separating [which already fixes
the scenario reported because singleuser gpencil modifiers work
correctly].
Note: we could have also called `ED_object_add_duplicate` with the
`USER_DUP_GPENCIL` dupflag -- that would have taken care of the usercount
--, but then the whole following logic in `gpencil_stroke_separate_exec`
would need to change from **adding** layers/frames/strokes to
**removing** these.
Part of T79273
Maniphest Tasks: T79273
Differential Revision: https://developer.blender.org/D8419
This is in order to disolve GPU_draw.h into more meaningful code blocks.
All the Image related function are in `image_gpu.c`.
All the MovieClip related function are in `movieclip.c`.
The IMB module now has a connection with GPU. This is not strickly
necessary and the code could be move to `image_gpu.c` if needed.
The Image garbage collection is also ported to `image_gpu.c`.
This property adds constraints to the simulation using the initial
location of the vertices, making it behave like a soft body. The
strength of these constraints can be modified with the brush parameter.
This makes some deformation modes more subtle and predictable, making it
possible to use the cloth brush to add surface detail in a more
controllable way without loosing completely the original shape of the
mesh.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D7845
This uses mesh's runtime mutex for both `BKE_mesh_runtime_looptri_ensure`
(was using its own global RW mutex before), and `BKE_mesh_wrapper_ensure_mdata`
(was not protected at all before).
This is more like a band-aid than a proper fix, as mentioned in the report
proper fix would be for the modifier to request those data (the relevant
BVHTree, which would implicitely also call the tow others) through flags,
just like it does for regular CDData layers. But this is a much bigger
refactor to be done outside of bugfix scope.
Reviewed By: sergey
Maniphest Tasks: T78285
Differential Revision: https://developer.blender.org/D8415
Previously, gravity was only applied in the real brush radius, not in
the whole simulation radius. For most deformation modes, applying
gravity to the entire simulation instead of just to the brush radius and
scaled by the radius (like a regular sculpt brush) makes more sense.
After this fix and with the cloth collisions patch applied, it is possible
to do things like this with the cloth grab brush.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8406
As the comment says, anchored stroke can't rely on the first stroke
iteration for creating the simulation data. Probably lost in a cleanup.
I also made that anchored stroke doesn't restore the mesh state in the
cloth brush, so it can create the simulation effect.
Reviewed By: sergey
Maniphest Tasks: T79054
Differential Revision: https://developer.blender.org/D8348
The final render will use scene resolution in this case.
For example, when Color Input is plugger to preview and composite output
nodes, final render will flood-fill the final image which is a size of
scene resolution with this color. Before this fix the node preview was
empty. After this fix the node preview will be flood-filled with the
color.
Fixes T78586
Differential Revision: https://developer.blender.org/D8263
Were two issues:
- Divider was calculated in integer domain, causing rounding issues in
general case, and causing singularity in a corner case when input is
smaller than the preview size.
- The resolution was scaled down by 1 pixel for no obvious reason.
it's good to have an option to ' pin' a mode to the brush, to use that mode always, independent of the current viewport selected mode.
{F8723224}
Reviewed By: pepeland
Differential Revision: https://developer.blender.org/D8399
81a002
This was fine in those cases with current code, but that kind of
assumption is always risky, and an open door to hard-to-track bugs when
code changes in the future...
This is also a bit of code cleanup, reorganisation.
Tried to be DRYed but avoid too much code change to (hopefully) minimize
breakage.
- GPU: remove TEXTARGET_CUBE_MAP, this is no longer used in the codebase.
- GPUTexture: Move compressed texture upload to gpu_texture.cc
- GPUTexture: Add per texture Anisotropic filtering switch
Before, when extrude a point, the extruded point is considered as the first point of the stroke, but this was not logic.
Now, the extrude point is considered as last.
Related to T79313
Use 'emboss' instead of 'draw_type' as enum, layout & functions use
the term emboss.
This issue was noted by @Poulpator in D8414, as `dt` is also an
abbreviation for delta-time.
Build a temp matarray storing materials from obdata and source object
(depending on slots 'allocation' of source object), and assign those to
targets.
Also remove limitation of 'using same obdata is forbidden', just never
edit obdata materials in that case...
Certainly not perfect, but already much better than existing code.
The fisheye camera setup causes the edges of the image to not shoot primary rays. This was not
respected by OptiX because of an optimization that tried to reduce conditionals around trace calls.
Removing that does not seem to have an impact on performance anymore however and it fixes
the issue.
A face must have area in order to be possible calculate interpolation weights.
The same to the reference UVs.
But the new faces created with the extrude operator, have no area (before moving).
The solution was to get the loop data from some neighbor face.
Differential Revision: https://developer.blender.org/D8278
Keeping face attributes connected is now optional.
Keeping UV's connected is useful for organic modeling, but bad for
architectural.
Differential Revision: https://developer.blender.org/D8360
This bumps the minimally required CMake version from 3.5 to 3.10, as
discussed in D8405.
Since D7649 landed, there has been the `gtest_discover_tests()` call to
discover individual unit tests in `bin/tests/blender_test`. This
function was introduced in CMake 3.10.
Since there were no complaints about this incompatibility, I suspect
that a newer version is already in use by the majority of the
Blender-building people.
When an object is moved only by the rigid body physics system, the
function `BKE_object_moves_in_time()` will incorrectly return `false`.
This commit adds a comment to make this behaviour more explicit.
No functional changes.
The root cause was that `BKE_object_moves_in_time()` incorrectly returns
`false` when an object is moved by the physics system.
This also fixes the same issue in the USD exporter.
This was done for the regular dopesheet in rB9ccc73ade8a2, this just
makes that label consistent across all anim views.
Differential Revision: https://developer.blender.org/D8255
rna_EffectorWeight_path() needs to point to "domain_settings" (instead
of "settings"), was a missing change when switching the FluidModifier to
mantaflow.
Maniphest Tasks: T79264
Differential Revision: https://developer.blender.org/D8398
Setting the environment variable `PXR_PATH_DEBUG` non-empty will make the
USD library print the directories it uses to find its JSON files. This can
aid in debugging when this unit test fails. Now the failure message also
tells you about this.
No functional changes.
Move the common entries (View and Area) into a static method to be
called from other menus to avoid duplicating the New Collection and ID
Paste operators.
Previously, mesh boundaries were relaxed as any other vertex, which was
causing artifacts and unwanted deformation. In order to prevent this,
the mesh filter was using the automasking system to lock the boundary
vertices, which was hacked into the tool. For the brush, the only
solution was to enable boundary automasking to lock those vertices
in plance.
Now the relax vertex function slides the boundary vertices along the
mesh boundary edges, relaxing all the topology correctly while
preserving the shape of the mesh. The automasking hack in the relax
mesh filter was also removed as now vertices slide correctly along
the boundary.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8350
All these data arrays are created for a specific topology, so they should be
freed and updated when the PBVH rebuilds. Previously, this was only
happening when freeing the SculptSession, but it also needs to happen in
BKE_sculpt_update_object_before_eval to avoid reusing out of date data.
Reviewed By: sergey
Maniphest Tasks: T79074
Differential Revision: https://developer.blender.org/D8357
When there is no color layer available,
BKE_sculpt_update_object_for_edit creates a new one and tags the mesh
with ID_RECLAC_GEOMETRY, so this layer is inmediatly available when the
tool starts. This also deletes the PBVH and when it is created again in
BKE_sculpt_update_object_after_eval, the pmap is not initialized, making
the tool crash.
This moves the color layer creation to a separate function outside
BKE_sculpt_update_object_for_edit, which now runs after the color
layer is available, so it won't need to update again and the pmap will
still be available when the tool is used.
Reviewed By: sergey
Maniphest Tasks: T78242
Differential Revision: https://developer.blender.org/D8135
The buildbot uses a separate `CMAKE_INSTALL_PREFIX`. This means that
the unit test could not find its USD JSON files in the build directory.
Using `${CMAKE_INSTALL_PREFIX}` instead of `$<TARGET_FILE_DIR:blender>`
solved this.
This commit reverts the "noindex" part of the original commit.
using noindex made it imposible to link to a specific property.
The original warnings do not pose an issue so until a proper solution is
found I am reverting this commit.
This reverts commit 953c232db3
MSVC does need the wholearchive flag but it was not set,
so no tests were actually linked into the binary.
Reviewed By: sybren
Differential Revision: https://developer.blender.org/D8404
Some modes were working by mere chance in that ugly 'reversed' case, but
the to/from selection modes were not properly swapped...
Should also be safe for 2.83.
In the situation where the precompiled libraries are used on Linux +
GCC, a version of GCC older than 9.3 is guaranteed to cause problems.
This just implents a fatal error message when we know it doesn't make
sense to continue. We could do more checks and add some warnings, but
it's very likely that these will be ignored amongst the other noise.
Reviewed By: sergey, brecht
Differential Revision: https://developer.blender.org/D8396
The following nodes work now (although things can still be improved of course):
Particle Birth Event, Praticle Time Step Event, Set Particle Attribute and Execute Condition.
Multiple Set Particle Attribute nodes can be chained using the "Execute" sockets.
They will be executed from left to right.
Showing the Python error without any explanation is often
not enough information and doesn't hint that the error was in the
user input.
The error report from a invalid expression such as '..1' used to be:
('invalid syntax', ('<string>', 1, 1, '..1'))
Now reads:
Error evaluating number, see Info editor for details: invalid syntax
Address issue raised by T78913.
The result of syntax errors read poorly in reports,
resulting in cryptic and unhelpful information.
Change PyC_ExceptionBuffer_Simple only to extract the initial text,
making syntax errors when entering invalid numeric expressions into
buttons easier to follow.
This operator cleanup any frame that is equal to the previous one. This is very handy when convert a mesh animation to Gpencil and the mesh is static for several frames.
Differential Revision: https://developer.blender.org/D8149
This makes `RNANodeQuery::construct_node_identifier()` more strict in
its matching of certain property names.
The downside of this approach is that it's not possible any more to use
`"rotation"` and expect a match for `"rotation_euler"` and friends, so
the list of strings to test against is now 3x as long.
Reviewed By: sergey
Maniphest Tasks: T79121
Differential Revision: https://developer.blender.org/D8375
The complete lack of a transform panel was confusing enough to spawn
this question: https://blender.stackexchange.com/q/169074/599
Displaying a message instead of nothing is more consistent with the
behavior of the mesh transform panel anyway.
Reviewed By: Blendify
Differential Revision: https://developer.blender.org/D8390
New option that lets users the define the maximum number of fluid particles that will be allowed in the simulation. This can come in handy, for example, to ensure that the particle count will not exceed the hardware capabilities, or to avoid excessive amounts of particles in a scene.
New option that lets users the define the maximum number of fluid particles that will be allowed in the simulation. This can come in handy, for example, to ensure that the particle count will not exceed the hardware capabilities, or to avoid excessive amounts of particles in a scene.
This is handy to add support for enums used in both C and C++
files. This removes the need to typecast each time for every
operation.
Only support bitwise operators for now.
Operator logic is limited to iterating over selection and executing
same code as python API does.
Functional changes:
- No attempt to preserve effects is made. Dependant effects are deleted.
- No attempt to change meta strip boundaries.
Partially fixes T73828
Reviewed By: campbellbarton
Differential Revision: https://developer.blender.org/D6892
Add recursion check before assigning strip as a mask for modifier.
Same check is used for recursion check when reassigning effect input, so it
should not be possible to create recursion at all.
Sequencer was not initialized yet, and RNA update function tried to clean up cache.
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D8323
This operator automates the following steps:
1. Create a point cloud object.
2. Create a simulation data block.
3. Add a small particle simulation to the node tree.
4. Add a Simulation modifier to the point cloud object.
5. Reference the particle simulation from the modifier.
You have to go back to frame 1 to start the simulation.
The simulation is not yet cached and cannot be rendered.
The bounding box of the point cloud object is enabled for now,
because otherwise it is hard to select the object.
The OptiX kernels are compiled for target "compute_sm_52", which is only available on second
generation Maxwell GPUs, so disable support for older ones.
A particle action is some function that is triggered by some event.
Right now, users cannot control this. There is just a
randomize-velocity on-birth action. So the direction of spawned
particles is slightly randomized now.
This also adds a new integer attribute called "Hash" which is
useful for a number of things. Mainly for generating random numbers
for a specific particle. The ID of a particle is not necessarily a good source
of randomness.
When definining static variables that own memory, you should
use the "construct on first use" idiom. Otherwise, you'll get
a warning when Blender exits.
More details are provided in D8354.
Differential Revision: https://developer.blender.org/D8354
This is a convenience wrapper for `Map<Key, Vector<Value>>`.
It does not provide any performance benefits (yet). I need this
kind of map in a couple of places and before I was duplicating
the lookup logic in many places.
This is a continuation of fix for T78307. Turns out instancing do not
work at all, so enforce single widget drawing on macOS and Intel GPU.
It was also reported that certain AMD and Mesa driver suffer from
similar issue, so disabled instancing for this configuration as well.
Differential Revision: https://developer.blender.org/D8374
When "ID Data" -> "Make Single User" is chosen with selected elements
that are not curves, there is a crash. This fix ensures that the id in the
callback function is an Action.
Once the base pose was changed (e.g. by changing the active landmark), we'd always run the logic to reset to the base pose. That would mess up the final viewer pose.
Think this only got exposed through 607d745a79.
When interpolate, the temp frames are tagged, and later removed. If for any reason any stroke was tagged in other area of Blender (this tag is used for temp tagging), the stroke could be removed by error when run interpolation.
In a previous fix, the tag was cleared before, but only for the frame range of the interpolation. Now, the clear is done for all frames.
Also, instead to clear for each stroke sampling, now it's done only once at invoke time.
Pretty straight-forward, although it does not rebuild relationships
(this is a heavy process we do not want to automate for now, will be a
separate operator in near future).
Fix T78179: Library Overrides - Additional geometry not updating without reloading scene.
Once again, we need lots of unatural twisting and convolutions to get
those nasty things to work...
Note that in previous versions of Blender, this would not crash, but was
still broken (shapekeys were never actually reloaded from libraries).
This is a change I pulled from the property-search-ui branch,
where I have to use the list of tabs to search the inactive tabs
and it makes more sense to use the array directly.
It is also an improvement to have this fundamental code to the
properties editor in the editor code rather than an RNA callback.
There are no functional changes.
Differential Revision: https://developer.blender.org/D8368
The extra depsgraph relations that were added to prevent threading
issues during evaluation (rB4c30dc343165) caused a considerable slowdown
on complex scenes with many drivers (T77277, T78615). This commit
improves this as follows.
Only the following drivers are considered for execution serialisation:
- Drivers on Array elements, and
- Drivers on Boolean or Enum properties.
Relations between drivers of the same arrays are added blindly, i.e.
without checking for transitive or cyclic relations. This is possible as
other relations will just target the `PROPERTIES_ENTRY` or
`PROPERTIES_EXIT` nodes.
Checking whether a driver is on an array is first done by checking
`array_index > 0`, and then falling back to resolving the RNA path to an
RNA property and inspecting that.
The code also avoids circular dependencies when there are multiple
drivers on the same property. This not something that is expected to
happen (both the UI and the Python API prevent duplicate drivers), it
did happen in a file (F8669945, example file of T78615) and it is easy
to deal with here.
Reviewers: sergey
Subscribers: mont29
Comment update
This means that we delete all override properties except for those over
ID pointers *if* the assigned pointer matches the linked data hierarchy.
Then we reload affected datablocks.
Regression from d6cefef98f
This also fixes an unreported issue where finding an exact match
wasn't being detected for items that contained an ID prefix.
A simulation data block has an embedded node tree, which requires
special handling in a couple of places. Some of those places were
missing beforehand.
This also adds a relation to make sure that the simulation is evaluated
after animations on the embedded node tree are evaluated.
High quality emitters need to maintain state themselves. For example,
this it needs to remember when it spawned the last particle.
This is especially important when the birth rate is changing over time.
Otherwise, there will be very visible artifacts.
It is quite likely that other components of the simulation need their own
state as well. Therefore, I refactored the `SimulationState` type a bit,
to make it more extensible. Instead of using hardcoded type numbers, a
string is used to identify the state type. Also, instead of having switch
statements in many places, there is a new `SimulationStateType` that
encapsulates information about how a specific state is created/freed/copied/...
I removed the integration with the point cache for now, because it was
not used anyway in it's current state.
* Changing to a landmark moves the view exactly to it, rather than
keeping the current position offset.
* Disabling positional tracking moves the viewer back to the landmark
position.
This is a more predictable and practical way to use landmarks. See
feedback in T71347.
On the code side, I did some cleanup so the logic flow is more clear.
Note: This is entirely untested. I currently don't have access to a
device. There might be issues, tomorrow I'll hopefully get feedback.
This adds extra deform modes to the slide mode of the Topology
Slide/Relax brush (both slide and smear are almost identical).
This is useful to move topology to a specific area to add more localized
details
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8349
This allows to use pen pressure modulation in hardness, wet mix, wet
persistence, flow and density, as well as inverting the modulation (more
pressure, less density...). With this, it is possible to create brushes
that mix paint or apply a new color based on the pressure.
Reviewed By: sergey, campbellbarton
Differential Revision: https://developer.blender.org/D8267
This matches the change that was done to the bevel modifier so that the
interface for the modifier, the active tool, and the operator are consistent.
This commit extends the refactor to the bmesh implementation too, so
that the parameters in the implementation don't stray too far from what
is exposed.
Tests are adjusted and still pass.
Both to_v and form_v need to be included. From_v is needed to include
the active vertex when there is only one vertex in the radius of the
brush, to_v needs to be included to add all grids duplicates
Reviewed By: sergey
Maniphest Tasks: T79056
Differential Revision: https://developer.blender.org/D8347
The custom smooth functions for bmesh and meshes where removed and
replaced by a generic smooth function using the sculpt API, which needs
to initialize the bmesh indices in order to be used
Reviewed By: sergey
Maniphest Tasks: T79007
Differential Revision: https://developer.blender.org/D8333
Implementation of lerp without a function requires repeating one of
the arguments, which is not ideal. To avoid that, add a new function
to the driver namespace. In addition, provide a function for clamping
between 0 and 1 to support easy clamped lerp, and a smoothstep function
from GLSL that is somewhat related.
The function implementations are added to a new bl_math module.
As an aside, add the round function and two-argument log to the
pylike expression subset.
Differential Revision: https://developer.blender.org/D8205
Object sockets work now, but only the new Object Transforms and the
Particle Mesh Emitter node use it. The emitter does not actually
use the mesh surface yet. Instead, new particles are just emitted around
the origin of the object.
Internally, handles to object data blocks are passed around in the network,
instead of raw object pointers. Using handles has a couple of benefits:
* The caller of the function has control over which handles can be resolved
and therefore limit access to specific data. The set of data blocks that
is accessed by a node tree should be known statically. This is necessary
for a proper integration with the dependency graph.
* When the pointer to an object changes (e.g. after restarting Blender),
all handles are still valid.
* When an object is deleted, the handle is invalidated without causing crashes.
* The handle is just an integer that can be stored per particle and can be cached easily.
The mapping between handles and their corresponding data blocks is
stored in the Simulation data block.
Adjusting the layout for this modifier is the final part of the modifier UI
project for 2.90. This layout exposes the most important information,
the levels and subdivision controls, by default, putting other contols
in collapsed subpanels.
Note that there is empty space for the "Delete Lower" button
that is still planned for 2.90. And there will also eventually be more
items added to the "Shape" panel, maybe for 2.90.
Differential Revision: https://developer.blender.org/D8187
Operators are one of the last places in Blender to use older UI designs
that don't fit in with recent style conventions. This commit updates
these custom operator UI callbacks for consistency and clarity.
Some of the code is also simplified a lot. Some of the older operator layouts
were much more complex (in terms of code) than they needed to be.
See the differential revision for a before and after screenshot
of each operator.
Differential Revision: https://developer.blender.org/D8326
Currently if you drag and drop an item from the outliner elsewhere in
the Blender window, the outliner will scroll the entire time, even if the
mouse is far away. This commit adds optional behavior for the edge pan
operator that makes it only act if the mouse is close enough to the region.
Differential Revision: https://developer.blender.org/D8193
When overriding a whole linked collection from a single object in
3DView, said 'reference' object would be removed by code. We only want
that behavior when overriding from an instanciating Empty.
It looks like the code left this as a todo, but the basic solution is to
add an extra parameter to BKE_bose_rest to check whether bones
are selected before reseting them.
I also corrected the operator description which said it acted on only
selected bones even when there is an option to turn that off. The
"act on selected" is generally implied for Blender's operators anyway.
Differential Revision: https://developer.blender.org/D8319
The function is called for all writers, not just 'object' writers.
Furthermore, it's called by the function `release_writers()`, so now the
name is consistent with that as well.
No functional changes.
The unit tests for `bf_io_common` didn't actually link against
`bf_io_common`, so when both USD and Alembic were disabled, nothing
would link against that library and building the tests would fail.
Any time FFmpeg was used to get a picture Blender would crash
when FFmpeg 4.3 is used. This affects thumbnails, VSE, Clip Editor.
Caused by a bug in FFmpeg which lead to crashes when unaligned
buffer was passed to sws_scale(). It got fixed later on in FFmpeg,
but for portability and compatibility reasons still nice to avoid
crash, especially since it's not so difficult to do.
FFmpeg ticked number is #8747
The FFmpeg Git hash with the fix: ba3e771a42c2
Differential Revision: https://developer.blender.org/D8355
Previous ones were extremely confusing and innacurate (probably
inherited from older versions of scene collections?).
Also, use named fields in struct initialization, much much safer and
cleaner than anonymous, positionned-based thing.
Added support for transforming only origins with greasepencil objects.
The new functions is based on BKE_gpencil_transform. That is why there is FIXME statements in there.
Reviewed By: Campbell, Antonio
Differential Revision: http://developer.blender.org/D8303
Internally UV selection considered close UV's to be connected.
While this could be convenient in some cases,
it complicates logic for more advanced selection operations that
need to check when UV's should be considered part of the same vertex
since simple threshold checks would give different results depending
on the order of UV's tested.
Users must now run "Merge by Distance" instead of relying
on this selection threshold.
There was a weird looking gap between the checkbox and the "Motion Tracking"
label. Plus, the label could not be clicked to change the value, unlike
usually.
Issue is that the row is actually a sub-panel header. The checkbox being drawn
with the draw_header() callback, and the label being added as separate item by
the popover panel code. This adds a hack so the checkbox can add the panel
label itself (the popup drawing skips adding the label then). That addresses
mentioned issues.
The sky will appear brighter than before by default. To compensate for this,
lower exposure in the Film panel. The default altitude was also changed from
90 to 15 degrees.
Patch contributed by Marco with the help of Ryan Jones.
Differential Revision: https://developer.blender.org/D8285
When bl_use_gpu_context is set, an OpenGL context will be available for
OpenGL based render engines.
Differential Revision: https://developer.blender.org/D8305
This patch changes the discovery of pre-compiled kernels, to look for any PTX, even if
it does not match the current architecture version exactly. It works because the driver can
JIT-compile PTX generated for architectures less than or equal to the current one.
This e.g. makes it possible to render on a new GPU architecture even if no pre-compiled
binary kernel was distributed for it as part of the Blender installation.
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D8332
Trying to get shape key pointer after having unlinked its owner from
Main data-base is rather useless... So those shapekeys ended up never
being deleted.
This refactor is in response to reports in which the adaptive domain with noise caused a crash (e.g. T79009). It should also fix issues where the smoke appeared to be cut off when using the adaptive domain together with noise. It is also possible that some of these changes improve the lines issue from T74559.
This was caused by `BPy_*_ALLOW_THREADS` being used when it shouldn't.
Implemented the simple fix suggested by @brecht :
> The simplest solution may be to ensure that Python stuff is only done
> when called through the RNA API, and not when Eevee calls it directly.
That project cannot be opened correctly ayway, it has recursive
collections intanciating themselves...
But at least now we have a check at startup to detect and 'fix' those
nasty cycles in collections.
Those are useful when you have to create containers with static
storage duration. If those would use Blender's guarded allocator,
it would report memory leaks, that are not actually leaks.
Analyzing the cuts, the points shown in the viewport (indicating the
previous and current cuts) do not correspond to the final cuts.
Sometimes a point, even snapped to a vert, is a point that cuts an edge and
sometimes a point, even snapped to an edge, is a point that cuts only the face.
This is because the detection of snapping vertices from mouse is different
from the "snap" that detects vertices that are cut.
So small projection inaccuracies can result in detection failures.
The solution for this is simply to confirm the cuts whose vertices
indicate the `prev` ou and `cur` point.
The tolerance distance does not need to be calculated in these cases.
This updates the usage of integer types in code I wrote according to our new style guides.
Major changes:
* Use signed instead of unsigned integers in many places.
* C++ containers in blenlib use `int64_t` for size and indices now (instead of `uint`).
* Hash values for C++ containers are 64 bit wide now (instead of 32 bit).
I do hope that I broke no builds, but it is quite likely that some compiler reports
slightly different errors. Please let me know when there are any errors. If the fix
is small, feel free to commit it yourself.
I compiled successfully on linux with gcc and on windows.
Custom driver functions need access to the dependency graph that is
triggering the evaluation of the driver. This patch passes the
dependency graph pointer through all the animation-related calls.
Instead of passing the evaluation time to functions, the code now passes
an `AnimationEvalContext` pointer:
```
typedef struct AnimationEvalContext {
struct Depsgraph *const depsgraph;
const float eval_time;
} AnimationEvalContext;
```
These structs are read-only, meaning that the code cannot change the
evaluation time. Note that the `depsgraph` pointer itself is const, but
it points to a non-const depsgraph.
FCurves and Drivers can be evaluated at a different time than the
current scene time, for example when evaluating NLA strips. This means
that, even though the current time is stored in the dependency graph, we
need an explicit evaluation time.
There are two functions that allow creation of `AnimationEvalContext`
objects:
- `BKE_animsys_eval_context_construct(Depsgraph *depsgraph, float
eval_time)`, which creates a new context object from scratch, and
- `BKE_animsys_eval_context_construct_at(AnimationEvalContext
*anim_eval_context, float eval_time)`, which can be used to create a
`AnimationEvalContext` with the same depsgraph, but at a different
time. This makes it possible to later add fields without changing any
of the code that just want to change the eval time.
This also provides a fix for T75553, although it does require a change
to the custom driver function. The driver should call
`custom_function(depsgraph)`, and the function should use that depsgraph
instead of information from `bpy.context`.
Reviewed By: brecht, sergey
Differential Revision: https://developer.blender.org/D8047
c08d847488 incremented the patch version instead of the file subversion
for versioning code when adding new options. This commit resets the patch
version and instead bumps the file subversion.
Status Bar can show scene statistics, memory usage, version, etc set by context menu. Part two of T75672.
Differential Revision: https://developer.blender.org/D7557
Reviewed by Julian Eisel
Support setting vert/edge/face selection, using the sticky option
without performing a second loop over all faces to flush selection.
Existing selection code didn't take advantage of BMesh connectivity
since the logic is from before BMesh was included.
Ogl default is 4 but for almost all cases, blender use tightly
packed format. This avoid confusion and state change for the
common case.
The only case that __might__ need alignement is DDS loader
(untested) so leaving this as it is.
The panel looks out of place with the rest of Blender's UI and the
text is cropped. With property split turned on and a few smaller
tweaks these issues are fixed.
| Before | After |
|{F8700181}|{F8700183}|
Differential Revision: https://developer.blender.org/D8322
When the playhead drawing moved to an overlay, a check was added to keep
it from drawing with a locked interface. This is necessary for some overlays,
but not this one, so this removes the check, making it the responsibility of
the editor.
A context function is added to make that check easier in the future.
Differential Revision: https://developer.blender.org/D8313
Adjusted the fluid build system so that plugins that depend on numpy can be compiled as well.
Note that in this commit numpy support is still disabled. It can be enabled by re-running the Mantaflow update script with USE_NUMPY=1 and enabling WITH_MANTA_NUMPY in extern/mantaflow/CMakeLists.txt. This will happen in a future commit.
No longer including unused dependencies. Should numpy IO be needed at some point, the Manta source update script can be configured so that the required dependencies are included again.
Instead of depending on static initialization order of globals use
static variables within functions. Those are initialized on first use.
This is every so slighly less efficient, but avoids a full class of problems.
I removed bf_blenkernel from `nodes/CMakeLists.txt` again (added it yesterday),
because now this was causing me unresolved symbol errors... Without it, cmake
seems to link the libraries bf_simulation, bf_blenkernel and bf_nodes in the right
order. Not sure if that is just luck or if it is guaranteed.
It was possible to fix the issue by using cmakes `LINK_INTERFACE_MULTIPLICITY`,
but that is probably bad style.
When using link duplicated objects it could happen that one object is
calculating the GPUBuffers and the second object is marking these
buffers invalid. This introduces threading issues.
This patch fixes this by combining the surface and surface per material
batches. Most likely the surface per material batches are used and when
requested you will most likely need the surface batch for the depth
tests and overlays.
During tests it slightly improves performance as batches aren't thrown
away without using it.
After this patch we can add a quick path for meshes with one material
and two materials.
Alternative approaches that have been checked:
- sync extraction per object: reduced performance to much (-15%)
({D8292})
- post checks: reduced the threading issues, but didn't solve it.
- separating preparation and execution of the extraction ({D8312})
Reviewed By: Clément Foucault
Differential Revision: https://developer.blender.org/D8329
This also introduces the `blender::nodes` namespace. Eventually,
we want to move most/all of the node implementation files into
this namespace.
The reason for this file-move is that the code fits much better
into the `nodes` directory than in the `blenkernel` directory.
Only the delete shortcut applies here, although the move up and down
operators can optionally be assigned in the keymap.
See rB1fa40c9f8a81 for more details and rB5d2005cbb54b for the
grease pencil modifier panel implementation, which is the same.
Some refactoring of the constraint delete operator was necessary,
including adding an invoke function.
Differential Revision: https://developer.blender.org/D8238
I got undefined reference errors on the `NodeMFNetworkBuilder::get_default_fn`
function under some circumstances. This symbol is definitely defined in bf_blenkernel.
The error seemed a bit undeterministic and was probably caused by some incorrect
link order. I don't get the error with this change.
CMake, when it's configuring the project, runs the `blender_test` test
runner (if it exists from a previous build) to discover which tests it
contains. At this time none of the tests themselves are run, so it's not
that useful to run ASAN and have it break things when there are memory
leaks.
This commit disables ASAN by injecting `ASAN_OPTIONS="detect_leaks=0"` in
the environment variables.
It is not enough to use `set(ENV{ASAN_OPTIONS} "detect_leaks=0")` in
`tests/gtests/runner/CMakeLists.txt`, as it wouldn't be passed to the child
process.
This commit is a followup of {D7649}, and ports the USD tests to the new
testing approach. It moves test code from `tests/gtests/usd` into
`source/blender/io/common` and `source/blender/io/usd`, and adjusts the
use of namespaces to be consistent with the other tests.
I decided to put one test into `io/usd/tests`, instead of
`io/usd/intern`. The reason is that this test does not correspond with a
single file in that directory; instead, it tests Blender's integration
with the USD library itself.
There are two new CLI arguments for the Big Test Runner:
- `--test-assets-dir`, which points to the `lib/tests` directory in the
SVN repository. This allows unit tests to find test assets.
- `--test-release-dir`, which points to `bin/{BLENDER_VERSION}` in the
build directory. At the moment this is only used by the USD test.
The CLI arguments are automatically passed to the Big Test Runner when
using `ctest`. When manually running the tests, the arguments are only
required when there is a test run that needs them.
For more info about splitting some code into 'common', see
rB084c5d6c7e2cf8.
No functional changes to the tests themselves, only to the way they are
built & run.
Differential Revision: https://developer.blender.org/D8314
Reviewed by: brecht, mont29
Not sure if these conversions are a good idea. However, we have them
in Cycles, so they be available in the simulation node tree for consistency
reasons.
For still images, always return 0 for the current frame number. This ensures
Cycles can detects that the image did not change.
Based on patch by Vincent Blankfield.
Differential Revision: https://developer.blender.org/D8242
Function names will be updated in a separate commit.
This will be the place for the new particle system and other
code related to the Simulation data block. We don't want
to have all that code in blenkernel.
Approved by brecht.
This commit introduces a new way to build unit tests. It is now possible
for each module to generate its own test library. The tests in these
libraries are then bundled into a single executable.
The test executable can be run with `ctest`. Even though the tests
reside in a single executable, they are still exposed as individual
tests to `ctest`, and thus can be selected via its `-R` argument.
Not yet ported tests still build & run as before.
The following rules apply:
- Test code should reside in the same directory as the code under test.
- Tests that target functionality in `somefile.{c,cc}` should reside in
`somefile_test.cc`.
- The namespace for tests is the `tests` sub-namespace of the code under
test. For example, tests for `blender::bke` should be in
`blender::bke:tests`.
- The test files should be listed in the module's `CMakeLists.txt` in a
`blender_add_test_lib()` call. See the `blenkernel` module for an
example.
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D7649
Having a sound strip in the VSE caused a missing relation error to be
logged on the console. This was caused by the AUDIO depsgraph component
not having an entry node. This commits adds that node, and sets up
relations correctly.
Differential Revision: https://developer.blender.org/D8290
Reviewed By: Sergey
Enabling all `make deps` dependencies with the exception of Embree and OIDN.
After that, Blender can be compiled on an Apple Silicon Mac just like on any
Intel based Mac. There are still compiler warnings that need to be
investigated and there are probably a couple of bug still to be discovered
and to be fixed.
Most patches to the dependencies are simple and are about disabling SSE and
setting the proper architecture to compiile for. Notable exception is Python,
where I back ported a yet to be accepted PR for upstream Python:
https://github.com/python/cpython/pull/21249
Cross compiling or buliding a Universal Binary is not supported yet.
The minimum macOS target version for x86_64 remains at 10.13, the target
for arm64 is 11.00.
Differential Revision: https://developer.blender.org/D8236
Key-map display was doing thousands of redundant alignment operations.
Set the spacing instead as align was only set to use zero spacing.
This would have prevented the crash reported by T78636.
Displaying user preferences search crashed on macOS when the search
contained a common character such as 'E'.
This caused alignment to 'alloca' too much memory.
Replace with a heap allocation fallback.
This happened because of typo in seq_dupli() when duplicating effect data.
Instead of duplicating data to new sequence, it was duplicated into original.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D8295
Also includes outline overlays. Removes the temp overlay drawing
We make the geometry follow camera like billboards this uses less
geometry. Currently we use half octahedron for now. Goal would be
to use icospheres.
This patch also optimize the case when pointcloud has uniform radius.
However we should premultiply the radius prop by the default radius
beforehand to avoid a multiplication on CPU.
Using geometry instead of pseudo raytraced spheres is more scalable as
we can render as low as 1 or 2 triangle to a full half sphere and can
integrate easily in the render pipeline using a low amount of code.
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D8301
The vertex colors node was using the M_COL attribute type but Sculpt
Vertex Colors use CD_PROP_COLOR
Now the Vertex Color node also fallbacks to legacy vertex colors if
Scultp Vertex Colors are not enabled as experimental.
Reviewed By: brecht
Maniphest Tasks: T78369
Differential Revision: https://developer.blender.org/D8185
When the mesh is linked, the materials can not be available or be the same assigned to mesh. Now, if the mesh is linked, a simple two materials conversion is used.
To get the full list of materials, the mesh must not be linked.
Also checked some indexes to be sure never get a wrong value and that materials are not created again and again.
`CD_LOCATION` was only used temporarily due to the lack
of a better alternative. This also removes the name from
`CD_LOCATION` again, because at most one layer of this
type should exist.
This issue was produced by a hack in the sculpt mode code from 2.80
when the sculpt API for connectivity info was not available.
The smooth brush was the only brush that needed connectivity info,
so there were 3 different smooth functions with the connectivity
queries implemented for dyntopo, meshes and grids. The mesh version
of smoothing was checking the number of connected faces to a vertex
to mask the mesh boundaries, which was not covering all cases and
was hardcoded in the smooth function itself.
This patch removes all those legacy functions and unifies all
smooth functions into a single one using the new API and the
automasking system. In order to achieve this, there were needed
some extra changes:
- The smooth brush now does not automasks the boundaries by default,
so its default preset needs to be updated to enable automasking
- The mesh boundary info is extracted once and cached in a
bitmap, similar to the disconnected elements IDs. This makes
boundary detection work as expected in all cases, solving a lot
of known issues with the smooth brush. In multires, this info is
extracted and cached only at the base mesh level, so it is much
more memory efficient than the previous automasking system.
- In order to keep the brushes responsive as they were before,
the automasking system can now skip creating the cache when it
is not needed for the requested options. This means that for
high poly meshes and simple automasking options the brushes
won't lag on start.
Reviewed By: sergey
Maniphest Tasks: T78747
Differential Revision: https://developer.blender.org/D8260
The flood fill operation was setting the mask using to_v, so in the first
iteration when the floodfill callback was using the active vertex as
from_v and any other neighbor as to_v, the mask for the active vertex
was never set.
Now the mask is set using from_v and it checks if it should continue
propagating to the next neighbor using to_v.
Reviewed By: sergey
Maniphest Tasks: T77417
Differential Revision: https://developer.blender.org/D8294
Those checks have been added to clang tidy within the last year.
They fail when I use a clang tidy version I built from source.
Reviewers: sergey
Differential Revision: https://developer.blender.org/D8302
option(WITH_CLANG_TIDY"Use Clang Tidy to analyze the source code (only enable for development on Linux using Clang, or Windows using the Visual Studio IDE)"OFF)
mark_as_advanced(WITH_CLANG_TIDY)
endif()
@@ -428,6 +440,14 @@ if(WIN32)
option(WITH_TBB_MALLOC_PROXY"Enable the TBB malloc replacement"ON)
endif()
# This should be turned off when Blender enter beta/rc/release
if("${BLENDER_VERSION_CYCLE}"STREQUAL"release"OR
"${BLENDER_VERSION_CYCLE}"STREQUAL"rc")
set(WITH_EXPERIMENTAL_FEATURESOFF)
else()
set(WITH_EXPERIMENTAL_FEATURESON)
endif()
# Unit testsing
option(WITH_GTESTS"Enable GTest unit testing"OFF)
option(WITH_OPENGL_RENDER_TESTS"Enable OpenGL render related unit testing (Experimental)"OFF)
@@ -498,9 +518,22 @@ if(CMAKE_COMPILER_IS_GNUCC OR CMAKE_C_COMPILER_ID MATCHES "Clang")
-fno-sanitize=alignment \
")
if(NOT MSVC)# not all sanitizers are supported with clang-cl, these two however are very vocal about it
Some files were not shown because too many files have changed in this diff
Show More
Reference in New Issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
