Sync the list of icons in CMakeLists.txt with UI_icons.h.
Restore, in the source file, the FUND icon that was accidentally
deleted.
Delete four old/unused icons.
See also D9715.
rBc70eb30240f8b5d5a8f2ac509f0eb585936142b5 added patch to use static
zlib on Linux. But also added flags to use the zlib in LIBDIR for
Python on macOS. That causes some shared libraries (binascii for one)
to link against libz.1.dylib which will not be there on users' systems.
Reuse the said patch for macOS also to avoid rpath issues.
Fix T85648.
Reviewed By: #platform_macos, brecht
Differential Revision: https://developer.blender.org/D10479
Constraint of new window position can be incorrect when using multiple monitors.
Differential Revision: https://developer.blender.org/D10469
Reviewed by Brecht Van Lommel
Remove the setting of Dialog window styles until we confirm expected behavior between platforms.
Differential Revision: https://developer.blender.org/D10470
Own Code
Removal of 'camera frame' around blend file thumbnail images.
Differential Revision: https://developer.blender.org/D10490
Reviewed by Brecht Van Lommel
Refactoring: WM_window_open() that can open different types of windows. 'New Window' with simplified layout.
Differential Revision: https://developer.blender.org/D10419
Reviewed by Brecht Van Lommel
Geometry nodes were not adding referenced instanced collections as
dependencies to depsgraph.
This would lead to meshes and data not being ready on evaluation in
certain cases.
This updates platform/platform_win32.cmake to support
both the old and new library names for OpenXR.
The new version links against one additional system
library and the debug library filename changed ever
so slightly.
This is a temporary workaround and can be removed
once the new lib versions have landed.
Rather than hardcoding the lib names, read
boosts version.hpp and extract the version
from there.
This will make it easier to land lib changes
in the near future.
Specular color is set to black instead of white inside the Principled BSDF
when the base color is set to fully black. This is contradictory to the sample
code of the Disney BRDF in BRDF Explorer. This patch aligns both
implementations.
Differential Revision: https://developer.blender.org/D10448
rB1f5647c07d15 introduced for the first time a unicode escape in strings
to be translated, directly extracted from C-code itself.
This revealed that this case was not properly handled by current code,
for now we work around using `raw_unicode_escape` encoding/decoding of
python.
Improvements to how window states are determined and changed.
Differential Revision: https://developer.blender.org/D10470
Reviewed by Brecht Van Lommel
Taa offset was applied on first sample. This wasn't happening before
DOF refactor.
Also fixes T85618 Wireframe Displays Strangely in Eevee (Rendered,
material Preview)
Contact shadows needed correct `gl_FragCoord.z` but this is not
correctly set for fullscreen passes. Need to pass depth using a global
variable until we get rid of `cl_eval.tracing_depth`.
This was introduced in the new geodesic distances algorithm for proportional
editing. When all faces of an edge are hidden, that edge should be considered
as loose geometry.
Initial patch by Pablo with modifications by Brecht.
Differential Revision: https://developer.blender.org/D10488
Currently the GPU module for python has different ways to handle enums.
- Organizing items in `PyC_StringEnumItems` arrays and parsing them with `PyC_ParseStringEnum`.
- Using dedicated functions for each type of enum (`bpygpu_ParsePrimType`, `pygpu_ParseVertCompType` and `pygpu_ParseVertFetchMode`).
Although apparently more efficient (especially `pygpu_ParseVertCompType`
which transforms strings into integers for simple comparison), these
dedicated functions duplicate functionality, increase the complexity of
the code and consequently make it less readable.
Reviewed By: campbellbarton
Differential Revision: https://developer.blender.org/D10456
With very large distances there were precision / overflow errors, normalize
the average albedo to avoid that. This was causing test failures on macOS
Arm, but also other architectures had slightly wrong results.
Ref T78710
The Exact modifier code had been written to avoid using BMesh but
in the initial release the modifier still converted all Meshes to
BMeshes, and then after running the boolean code on the BMeshes,
converted the result back to a Mesh.
This change skips that. Most of the work here is in getting the
Custom Data layers right. The approach taken is to merge default
layers from all operand meshes into the final result, and then
use the original verts, edges, polys, and loops to copy or interpolate
the appropriate custom data layers from all operands into the result.
When opening a Blender version for which there are no preferences, the splash
shows a button like "Load 2.92 Settings". Using this could cause a memory leak
of the storage for recently opened files.
The "quick setup" dialog is actually a 'menu', and the "splash screen" block
contains the UI_BLOCK_LOOP flag which causes the buttons' text to align
to the left, however, usually regular buttons have centered text.
As a workaround, add the UI_BLOCK_QUICK_SETUP flag which prevents
the text from being left-aligned.
Differential Revision: https://developer.blender.org/D10486
Reviewed by: Julian Eisel
Support Python 3.10a5 or 3.9x with support explicitly enabled.
- Enable Python's postponed annotations for Blender's RNA classes
types registered on startup.
- Using postponed annotations has implications for how they are defined,
since they must evaluate in the modules name-space instead of the
classes name-space. See changes to annotations in `release/scripts`.
- Use `from __future__ import annotations` at the top of the module
to ensure the script will run with Python 3.10.
- Old logic is kept since it could be used if PEP-649 is supported.
Resolves T83626
Ref D10474
This adds an approximation of inverted AO by reversing the max horizon
search (becoming a min horizon). The horizons are correctly clamped in
the reverse direction to the shading and geometric normals.
The arc integration is untouched as it seems to be symetrical.
The limitation of this technique is that since it is still screen-space
AO you don't get other hidden surfaces occlusion. This is more
problematic in the case of inverted AO than for normal AO but it's
better than no support AO.
Support of distance parameter was easy thanks to recent AO refactor.
Fix regression with roughness not masking reflections when not using
Screen Space raytracing.
The trick was to only evaluate one planar per pixel, the one with
the most influence. This should not be too limiting since this is what
we do for SSR.
Also change evaluation order do not apply occlusion on planars probes.
- Fix noise/banding artifact on distant geometry.
- Fix overshadowing on un-occluded surfaces at grazing angle producing "fresnel"
like shadowing. Some of it still appears but this is caused to the low number
of horizons per pixel.
- Improve performance by using a fixed number of samples and fixing the
sampling area size. A better sampling pattern is planned to recover
the lost precision on large AO radius.
- Improved normal reconstruction for the AO pass.
- Improve Bent Normal reconstruction resulting in less faceted look on
smoothed geometry.
- Add Thickness heuristic to avoid overshadowing of thin objects.
Factor is currently hardcoded.
- Add bent normal support to Glossy reflections.
- Change Glossy occlusion to give less light leaks from lightprobes.
It can overshadow on smooth surface but this should be mitigated by
using SSR.
- Use Bent Normal for rough Glossy surfaces.
- Occlusion is now correctly evaluated for each BSDF. However this does make
everything slower. This is mitigated by the fact the search is a lot faster
than before.
Previously, methods like `Span.drop_front` would crash when more
elements would be dropped than are available. While this is most
efficient, it is not very practical in some use cases. Also other languages
silently clamp the index, so one can easily write wrong code accidentally.
Now, `Span.drop_front` and similar methods will only crash when n
is negative. Too large values will be clamped down to their maximum
possible value. While this is slightly less efficient, I did not have a case
where this actually mattered yet. If it does matter in the future, we can
add a separate `*_unchecked` method.
This should not change the behavior of existing code.
Expanded enum items like this usually have centered text, but there are
limitations in the popup code that break this here. Add a workaround for this
limitation.
Differential Revision: https://developer.blender.org/D9854
Reviewed by: Julian Eisel
This patch adds a to_curve method to the Object ID. This method is
analogous to the to_mesh method. The method can operate on curve and
text objects. For text objects, the text is converted into a 3D Curve ID
and that curve is returned. For curve objects, if apply_modifiers is
true, the spline deform modifiers will be applied and a Curve ID with
the result will be returned, otherwise a copy of the curve will be
returned.
The goal of this addition is to allow the developer to access the splines
of text objects and to get the result of modifier applications which was
otherwise not possible.
Reviewed By: Brecht
Differential Revision: https://developer.blender.org/D10354
Previously the same functions were used to both set and get the active
state for outliner tree elements. This has quite a few problems.
- It's hard to tell when data is changed or simply read
- It prevents using `const`
- The code is full of if statements, making it longer and less readable.
This commit replaces the `tree_element_type_active` and
`tree_element_active` functions with `_get` and `_set` variants. One
has const arguments and returns the active state, the other deals only
with setting the state. While this refactor results in slightly more
lines of code, the result is much better in my opinion.
This commit also removes unused variables from arguments of the affected
functions.
Differential Revision: https://developer.blender.org/D10232
This commit adds a simple string input node, intended for use in the
attribute workflow to make using the same attribute name in multiple
places easier. The node is function node similar to the existing vector
input node.
Ref T84971
Differential Revision: https://developer.blender.org/D10316
This makes the nodes look more consistent with the rest of the UI
by using the same split with right aligned labels as the property editor
and elsewhere. Additionally, for consistency, the "Type" part of some
dropdowns is removed. It already wasn't displayed everywhere, and
it gets redundant quite quickly.
The "Hide Collection" operators assigned to the number keys in edit mode
trigger a redraw of the outliner, but as an optimization, they do *not*
trigger a rebuild of the tree.
This optimization is valid because unlike the collection exclude toggle,
the heirarchy is not affected by collection visibility. However, it means
that currently you must trigger a rebuild to get the correct "grayed out"
status after using the operator.
Rather than trigger a rebuild in this case to solve the bug, this patch
moves the decision for whether to gray out the text of a tree element
to the draw step rather than the build step. This means that any change
to the corresponding properties doesn't require a full tree rebuild.
Note that changing the "hide_viewport" property from the outliner still
causes a tree rebuild. I think that's because of the checks in
`outliner_collection_set_flag_recursive_fn`.
That could be optimized in the future.
Differential Revision: https://developer.blender.org/D10240
Add the Simple subdivision option to Geometry nodes, as a new node
instead of part of the existing subdivision node because of future
backend changes to the Simple option. (See T85584)
https://developer.blender.org/D10409
This reverts commit 7ee518cf70.
Causes T84869. The initial issue is less bad than this. We are looking into
alternative fixes for 2.93 (see D9854 and D9853).
This adds fuzzy search functionality to various input fields in Blender
where one can choose a value from many existing values (e.g. the
Vertex Group property in the Displace modifier).
Differential Revision: https://developer.blender.org/D10446
Following with the changes included to interpolate strokes of different number of points, a full review has been done in the interpolation tools.
* Interpolate now is a tool and not an operator. It was not logic to have this tool as a button.
* Interpolate tool parameters have been moved to topbar.
* Interpolate popover has been removed from topbar and interpolate `Sequence` operator has been moved to grease pencil menu.
* Interpolate Sequence now include a Redo panel.
* Interpolate tool now allows to select the strokes by pairs. This allows to interpolate any stroke with any stroke and not as before that it was only possible by drawing order. If no stroke is selected, the interpolation is done as before.
* Now is possible interpolate again if a previous keyframe exist. Before, it was impossible to interpolate two times in same frame and this made impossible to do the interpolation by groups of frames.
* New automatic option to `Flip strokes` if the stroke and end are not in the right position. Also the flip can be set manually for corner cases.
* Cleanup of menus related to interpolate.
* Fixed some bugs and removed parameters from scene because now all are tool or operator contained.
* Some code cleanup and function renames.
This commit also includes the some codebase to future implementation of the concept `Vertex Active` that now does not exist in grease pencil.
This patch implements the list panel system D7490 for FCurve modifiers.
The UI layouts are updated to make use of subpanels and to be consistent
with the rest of the interface, and easier to understand.
See the differential revision for screenshots.
This commit also significantly cleans up the FModifier UI code, and
improves, mainly by replacing the old button creation code is with
the newer interface API using RNA. In turn there is a bit of complexity
added because each FModifier has a separate panel.
Although reordering of FModifiers was not implemented before, we get
drag and drop basically for free here, so it is also included.
As noted in some older to do tasks, FModifiers aren't evaluated
in perfect order, which may be a point of improvement for the future.
Differential Revision: https://developer.blender.org/D7997
The problem was introduced fixing task T85035.
As the frame was set again when render, if there was a time modifier, the frame was not remaped to the right frame number.
Commit rBf448ff2afe7a accidentally added menu entries that were removed.
`use_zoom_to_fit` is duplicated in menu, others were moved to overlays menu.
Remove these entries.
Reviewed By: ISS, campbellbarton
Differential Revision: https://developer.blender.org/D10453
Previously, functions would have to ask for every geometry type explicitely.
Using a vector is return type is fine. In practice this will probably never
allocate because of the small buffer optimization in vector.
If the axes are aligned in auto pivot mode then the rotation axis would be (0,0,0).
We now fall back to the x axis in this case. If that fails, we fall back to the y axis.
Differential Revision: https://developer.blender.org/D10466
Previously, the density was set per point. That implies that when a
point has a non-zero weight, points might be distributed in all
connected polygons. By specifying the density per corner, this
limitation is removed. Note, per-point density maps (such as vertex
groups) can still be used. They will be adapted to the corner domain
without loss of information.
Differential Revision: https://developer.blender.org/D10461
Instead of raising an expection a warnign is generated instead. This
fixes the issue where `['hair', 'pointcloud']` are disabled for release
builds. In the future a better solution would be to generate the context
map dynamically but this would require refactoring of the API: D9988
This fixes the issue where `['hair', 'pointcloud']` are disabled for
release builds.
In the future a better solution would be to generate the context map
dynamically but this would require refactoring of the API: D9988
Fixes T80364
Differential revision: https://developer.blender.org/D10468
This is needed to support Python 3.10's `typing.get_type_hints`,
to access the name-space used when creating the class.
Also added a docstring for execfile.
When `Preferences → Editing → New Objects → Align To` is set to `3D
Cursor`, the fluid domain added by `Object → Quick Effects → Quick
Liquid` is aligned to the 3D cursor. This shouldn't be the case, since
these aren't new objects created directly by the user.
Differential Revision: https://developer.blender.org/D10467
Crash happened in versioning code on NULL dereference in function
seq_convert_transform_crop() for Strip crop and transform
fields.
Strips created after rB1fd7b380f4cf were assumed to have crop and
transform always initialized, but this wasn't the case. This has been
fixed in 2.90, but not in versioning code.
Initialize these fields if they are not initialized already.
Reviewed By: campbellbarton
Differential Revision: https://developer.blender.org/D10463
E.g. steps to reproduce:
* Enter Vertex Paint mode
* In the tool settings, right-click > "Edit Source"
When creating a number slider via `layout.prop(..., slider=True)`, the UI code
would reallocate the number button to be a number-slider button. That's because
we now actually have different button data-structures for these (see
e6f0b60c2e). The edit source code stored data based on the button pointers,
which didn't get updated after changing the type. The fix just adds this
updating.
"Kind" is a bit less generic than "Info" for me. Especially, it implies
that the struct does not contain the name of a specific attribute
(for me anyway).
Caused by rB85fe12071ad7.
When looking at a render pass in viewport and move the camera, there's a
"flash" effect in this AOVs test file:
{F9753476}
{F9753473}
The cause seems to be that taa_current_sample when rendering
(DRW_state_is_image_render()) is ok because it has been incremented in
EEVEE_temporal_sampling_draw but taa_current_sample is wrong when we are
not rendering.
D10375 by Ulysse Martin (youle) with clang format changes.
Reviewed By: jbakker, lichtwerk, also blessing from fclem
Differential Revision: https://developer.blender.org/D10375
This allows accessing per-point attributes on the corner domain,
which can be useful e.g. when adding per-point displacement
to per-corner uv coordinates.
Also it is required to make an upcoming patch work well, that
makes the Point Distribute node use density weights per corner
instead of per point, giving the user more precise control over
the distribution.
This makes vertex colors available in geometry nodes similar to
how uvs are available. They can be used using the attribute system.
Vertex colors are stored per corner (as are uvs, but not like vertex weights).
Ref T84297.
Differential Revision: https://developer.blender.org/D10454
Previously, when a Geometry Nodes modifier outputs a point cloud
(e.g. generated using the Point Distribute node), other modifiers
could not use that data. Now, the point cloud data is converted to
mesh vertices for such modifiers.
Ref T85281.
Differential Revision: https://developer.blender.org/D10451
All headers in the svn libraries should be
building with /W0 however MSVC 16.8/16.9
has broken this functionality making OpenVDB
emit lots of warnings.
The breakage was reported [1] in august to MS
but they still have not gotten around to
addressing the issue.
This change explicitly suppresses C4251 in the
module that emitted these warnings. As the warning
is useful in other parts of blender a localized
approach is taken rather than a global suppression.
[1] https://developercommunity.visualstudio.com/content/problem/1167590/bug.html
The integer mode still needs a random value in a 0 to 1 range, just like
floats. So slightly refactor to let the integer randomization use the
float implementation and then round to an int afterwards.
From the the opengl wiki:
> Buffer objects are associated with a program's uniform block similarly to the way that texture objects are associated with sampler uniforms.
* WITH_CPU_SSE was renamed to WITH_CPU_SIMD, and now covers both SSE and Neon.
* For macOS sse2neon.h is included as part of the precompiled libraries.
* For Linux it is enabled if the sse2neon.h header file is detected. However
this library does not have official releases and is not shipped with any Linux
distribution, so manual installation and configuration is required to get this
working.
Ref D8237, T78710
Was reported for Geometry Nodes Editor "New" button, but was true for
any modifier added via modifiers.new().
Now just add appropriate nofifier.
Maniphest Tasks: T85722
Differential Revision: https://developer.blender.org/D10450
The `material_index` attribute can adjust which material in the list
will be applied to each face of the mesh. There are two new things
about this attribute that haven't been exposed by the attribute API yet.
Each comes with limitations:
1. Integer data type: Most attribute nodes are currently written to use
float data types. This means that they can't write to this attribute
because they can't change the type of a built-in attribute.
2. Polygon domain: This is our first attribute using the polygon domain,
meaning until some of the interpolations are implemented, some
operations may not work as expected.
Currently the two nodes that work with this attribute are Attribute Fill
and Attribute Randomize.
Differential Revision: https://developer.blender.org/D10444
When primitive offsets change we need to rebuild or refit BVHs, however this
was also tagging other data as modified too late in the geometry update process.
Now ensure only the BVHs are updated.
Ref D10441
To avoid anti-aliasing artifacts on GPencil strokes that have a size smaller than 1 the thickness
is clamped and the opacity reduced. This was done in vertex shader but this had the side effect
of causing strokes that go from large to small to fade across their lengths.
**Solution**
The opacity modulation has now been moved to the fragment shader as advised by Clément Foucault.
The strokeThickness that was passed to the shader was clamped to prevent it from being too small so an additional
unclamped thickness has been passed to the fragment to calculate the opacity modulation. Alternatively I could have chosen
strokeThickness and clampedThickness but I decided against renaming the variables in the current code.
Reviewed By: fclem
Differential Revision: https://developer.blender.org/D10438
Previously, when the output of a Geometry Nodes modifier would
contain instances, those could not be accessed by other existing
modifiers (e.g. the Array modifier). That is because those modifiers
don't know about instances.
Upcoming commits will improve an this in two ways:
* Also realize instances before deform modifiers.
* Convert a point cloud in the geometry to mesh vertices so that
they can be accessed as well.
Note, making instances real can result in loosing some information
that we do not support in Geometry Nodes yet. That includes some
special builtin attributes like bevel weights.
Ref T85281.
Differential Revision: https://developer.blender.org/D10432
This makes custom mesh attributes available in Cycles. Typically,
these attributes are generated by Geometry Nodes, but they can also
be created with a Python script.
* The `subdivision` code path is not yet supported.
* This does not make vertex weights and some other builtin attributes
available in Cycles, even though they are accesible in Geometry Nodes.
All attributes generated in Geometry Nodes should be accessible though.
* In some cases memory consumption could be removed by not storing all
attributes in floats. E.g. booleans and integer attributes for which
all values are within a certain range, could be stored in less than
4 bytes per element.
Differential Revision: https://developer.blender.org/D10210
This patch has originally been written by Kévin Dietrich, thanks!
It is part of D10210.
As Brecht noted in D10210, this might not handle all cases yet.
I better solution should come soonish.
Since "Applied Modifier" was always added last, it obscured more important
messages when using the shortcut to delete modifiers. The purpose of the
report when using the shortcut was to make it clear that something
happened. Since another report does that anyway, only display the
"Applied Modifier" report if the report list length hasn't changed.
Issue introduced by my commit: rB40b7929cc040
**User-level Problem**:
The issue resulted in a full-replace upper strip with only a Z location
channel full-replacing the XY location channels of the lower stack.
replaced to default. The expected behavior is that only the Z location
channel is affected.
**Technical-level Problem**:
Before the problematic commit, fcurves were blended as they were read.
So only existing animated channels would blend. My recent commit
changed the process to read all fcurve values into an isolated
upper_snapshot then blend with the lower stack. There is no data stored
to know whether the upper snapshot channel values were sampled from
fcurves or were default values. Only those sampled from fcurves should
be blended.
**Solution**:
Added a `blend_domain` bitmask member to NlaEvalChannelSnapshot.
The blending function only blends values within the `blend_domain`.
Sampled fcurve values are now marked as within the `blend_domain`.
We also now always copy the lower snapshot to the result snapshot which
only matters when they aren't the same. Currently, it's always the same
so the change is more for future unseen cases.
Reviewed By: sybren, #animation_rigging
Differential Revision: https://developer.blender.org/D10339
Windows Ghost cursor movement was previously changed to use SendInput
because SetCursorPos sporadically allows the cursor to escape the
window. This is now reverted because SendInput causes mouse history via
GetMouseMovePointsEx to contain invalid movement history, likely due to
interaction with mouse acceleration. This resulted in popups closing
when the cursor appeared to leave their range.
This patch adds icons to the right side of nodes when they encounter a
a problem. When hovered, a tooltip displays describing the encountered
while evaluating the node.
Some examples are: attribute doesn't exist, mesh has no faces,
incorrect attribute type, etc. Exposing more messages to the system
will be an ongoing process. Multiple warnings per node are supported.
The system is implemented somewhat generically so that the basic
structure can also be used to store more information from evaluation
for the interface, like a list of available attributes.
Currently the messages are just button tooltips. They could be styled
differently in the future. Another limitation is that every instance of
a node group in a parent node tree will have the same error messages,
the "evaluation context" used to decide when to display the tooltips
must be extended to support node tree paths.
Differential Revision: https://developer.blender.org/D10290
Since the derived node tree is already build for the evaluation system,
it's simpler to pass a derived node to the params struct. This will also
allow context lookups in nested node groups for node error messages,
since the derived node has that information readily accessible.
This will allow using C++ data structures to draw node error messages.
This required removing a forward declared enum for grease pencil undo.
Compiles with clang tidy.
This commit adds support for `bl_description` and python docstrings for panels.
This is useful for pop-over panel types so they can have a label and description.
This commit also includes an example use case.
Reviewed By: campbellbarton
Differential Revision: https://developer.blender.org/D10429
When dropping an asset somewhere, it is appended and then a drop operation is
called to actually add it to the scene based on current context. If this drop
operation fails, the appended data-block is now still in the .blend. The user
may not notice and not expect this.
Instead idea is to rollback any changes done by dropping code if the operation
fails, namely removing the appended data-block again.
Adds a new `cancel()` callback which is called if the drop operator returns
`OPERATOR_CANCELLED` to drop-boxes and a generic function to deal with assets
on drop failure.
Also removes the `free_id_on_error` property of the `NODE_OT_add_group`
operator, which was used as ad-hoc solution to get this same behavior.
Fix for T84928 .
Considering the changes , issue is resolved ( Ignoring readability issues) .
**Changes **:
- `Change in value assignment of fu/v/w :` Observing previous code , I noticed ,value assigned to them is equivalent to -0.5 ( i.e. co-ordinate of left most vertex of lattice size =1 where centre of lattice is origin ) .
- `Change in value assignment of du/v/w :` Margin ( distance ) between each division of surface along any axis is equivalent to **( (length of surface along axis ) / (no of division line - 1) )** . that's why is changed it to (default_size/unew -1) .
- ` New variable declared "default_size" :` As far as I gone through the code , I noticed values 1 < du ,fu < 1 , which indicates these values were calculated with respect to default lattice of size 1 .
- `removed pntsu/v/w != 1 check :` Following changes inside the if block worked properly for pntsu/v/w = 1 .
Reviewed By: lichtwerk, campbellbarton
Differential Revision: https://developer.blender.org/D10353
The vertex weights were actually interpolated correctly. The issue was that
vertex group names were removed from the output geometry set. We have
to keep track of these names separately from the mesh, for legacy reasons.
Vertex group names are not stored on meshes but objects instead.
In an upcoming commit I'll also move the make-instances-real functionality
to this file. This code is not essential to working with geometry sets in general,
so it makes sense to move it to a separate header.
The logic to remove one Python module before installing another
only worked in simple cases where a file replaced a file.
- Installing a single file add-on over a Python package with the same
name caused an error as the directory isn't empty.
- Removing existing module directories from the zip-file did nothing
as the directories from the zip-file that end with a slash were
compared with directories from `os.listdir` that don't.
- `module_filesystem_remove` assumed ZipFile.namelist() was a list of
files in the root of the zip-file when it's a list of all files.
While I couldn't find any bugs caused by this, it performed checks
that don't make sense, comparing files at different depths of the
file-system.
This improves stability and convergence speed of Workbench Temporal AntiAliasing.
This adds a filtering kernel (blackmann-haris, same as EEVEE/Cycles) to the
temporal antialiasing sampling. We also gather neighbor pixels since they might
end up in the pixel footprint.
We use a 1px radius for the filter window which is a bit less than the 1.5 default
of cycles and EEVEE since it does blur quite a bit more than what we have now.
Another improvement is that the filtering is now in log space which improves
AntiAliasing around highlights.
Theses improvement may not be very useful for every day case but it was an
experiment to try to make TAA usable for GPencil.
Test file used :
{F9798807}
|filtered+logspace|filtered|original|
|{F9798847}|{F9798848}|{F9798849}|
Reviewed By: jbakker
Differential Revision: https://developer.blender.org/D10414
This revert removes handling of cursor move and button press events
during Wintab's WT_PACKET event, instead storing pressure and tilt
information to be combined with Window's WM_MOUSEMOVE events.
This also reverts dynamic enabling and disabling of Wintab, dependent
on the chosen Tablet API. If the Tablet API is not explictly Windows
Ink during startup, Wintab is loaded and enabled.
Left in place is a fallback to Windows Ink when the Tablet API is set
to Automatic and no Wintab devices are present. This allows devices
with Wintab installed but not active to fallback to Windows Ink.
Using position provided by Wintab was found to have too many
regressions to include in Blender 2.93. The primary source of
regressions was tablets which mapped coordinates incorrectly on multi-
monitor and scaled displays. This resulted in an offset between what
the driver controlled Win32 cursor position and the Wintab reported
position. A special case of this included tablets set to mouse mode,
where Wintab reported absolute position while the system cursor moved
as a relative mouse with mouse acceleration.
Reloading preferences didn't update Python's `sys.path` to account
for the modified `script_directory`.
This meant the operator to load settings from a previous version
required a restart to initialize Python when this directory was set.
Updates the descriptions and labels for outliner sync in the properties
editor popover to be more clear.
Differential Revision: https://developer.blender.org/D10010
Updates the descriptions and labels for outliner sync in the properties
editor popover to be more clear.
Differential Revision: https://developer.blender.org/D10010
Blender would crash when selecting multiple keyframes while multiframe edit was
active. This was due to the active frame being NULL in some instances.
The fix checks if the active frame is not NULL.
Reviewed By: antoniov
Maniphest Tasks: T85587
Differential Revision: https://developer.blender.org/D10421
This required using a fork of Embree, newer LLVM version, unreleased ISPC
version and sse2neon directly from Git. Hopefully over time all the required
changes end up in official releases. For now we deviate from other platforms.
Based on contributions by Apple and Stefan Werner.
Ref D9527, D8237, T78710
Building NumPy from source with default options of builder
causes it to link against Accelerate framework which is buggy and
raises a warning mentioned in [2].
"RankWarning: Polyfit may be poorly conditioned"
Accelerate is deprecated with NumPy 1.20+.[1]
So either we build OpenBLAS in dependencies also and set appropriate
env variables suggested in [1] while building NumPy for it to find
OpenBLAS. Or download NumPy wheel from pip and never allow pip to
build NumPy from source while installing.
After this change, pip wheels are used for NumPy for macOS with x86_64.
[1] https://numpy.org/doc/stable/user/building.html#lapack
[2] https://github.com/numpy/numpy/issues/15947
Reviewed By: #platform_macos, sebbas
Differential Revision: https://developer.blender.org/D10368
A fix for T83187 (rBf83aa830) assumed in the overlay code of the uv editor that the object was a mesh
when it did not have to be - causing a crash.
The fix makes sure that the object is a mesh.
Reviewed By: jbakker, campbellbarton
Maniphest Tasks: T85499, T85495
Differential Revision: https://developer.blender.org/D10369
Adds `NODE_OT_add_group` operator to add a node group from a given name, and
uses that to register a node editor drop-box.
When dropping a node-group asset, the ID will be appended. This is what we do
for other ID assets too.
Should the node group insertion fail (e.g. the group is not compatible with the
current tree, as checked by the poll), the appended data-block is removed.
Differential Revision: https://developer.blender.org/D10405
Reviewed by: Jacques Lucke
This is because the node editor added a "real user" to the node group
that it displays. It was mainly added as a hack to solve issues with custom
tree types (T36024). Since we can store id references in custom properties
now, this "real user" is not really necessary anymore.
Given that we are close to a release, I'll only disable line for geometry nodes,
for which the bug has been reported.
Discussed this solution with Hans Goudey.
This reverts a part of rBd3960164163c910d5031a8f076c41b39e0a5503d.
It is not a `std::shared_ptr` but a `boost::shared_ptr`.
This could probably be fixed differently, but `NOLINT` is fine now.
Following code from D8627 this patch corrects multi threaded processing
of proxies, where a 60 sec proxy generation drops to 35 sec.
Differential Revision: https://developer.blender.org/D8659
Allow use all system threads for frame encoding/decoding. This is very
straightforward: the value of zero basically disables threading.
Change threading policy to slice when decoding frames. The reason for
this is because decoding happens frame-by-frame, so inter-frame threading
policy will not bring any speedup.
The change for threading policy to slice is less obvious and is based on
benchmark of the demo files from T78986. This gives best performance so
far.
Rendering the following file went down from 190sec down to 160sec.
https://storage.googleapis.com/institute-storage/vse_simplified_example.zip
This change makes both reading and writing faster. The animation render
is just easiest to get actual time metrics.
Differential Revision: https://developer.blender.org/D8627
USD version 21.02 includes two of the changes Blender used to patch in,
which have now been removed from `usd.diff`. Unfortunately 21.02
introduces another issue where LZ4 symbols are accidentally exported,
causing linker errors. Fortunately these symbols are only used for
resting, so I added a patch hunk that simply removes their `extern "C"`
declaration.
The LZ4 linker issue has been reported upstream at
https://github.com/PixarAnimationStudios/USD/issues/1447.
Reviewed By: sebbas, mont29
Differential Revision: https://developer.blender.org/D10367
Straight up version bump
Things of note:
They started using API calls only available in windows 8, however given
the Python 3.9 update will forcibly bump us to 8.1+ anyhow this is not
an issue.
Will require some minor tweaks to platform_win32.cmake after adding the
libs to svn which are not included in this diff so this diff can land
without having to have the libs in place yet.
Reviewed By: sebbas, sybren
Differential Revision: https://developer.blender.org/D10349
Straight forward version bump, some of the variables to detect a static
OpenEXR changed and the folder structure seemingly changed a little
requiring updates to the diff
Reviewed By: sebbas, sybren
Differential Revision: https://developer.blender.org/D10340
Version bump + no longer using Boost.
Building Alembic with Boost gave compiler errors, and having one less
inter-dependency is good as well.
Reviewed By: sebbas, sybren
Differential Revision: https://developer.blender.org/D10329
Straight forward version bump.
2020U2 is significantly louder in the deprecated header usage warning
department, we should probably see if we need to act on this: P1949
Differential Revision: https://developer.blender.org/D10359
Previously `mesh.attributes.new(...)` would return a generic attribute that
one could not do much with. Now it returns refined attributes like `FloatAttribute`.
Add a boundary check, avoiding access past actual data.
Ideally would need to report error to the user somehow,
but it doesn't seem to be easy to do.
This is a minimal safe patch. The proper complete fix is
being worked on by Jesse.
Differential Revision: https://developer.blender.org/D10357
The inset operator would display the inset thickness and depth as
Blender units during transform. This meant that when the scene units
were set to something different than meters, the display value would
give different results than entering the same number using the value
input. (Similar to D10325)
The fix makes sure that the numbers are always displayed in the correct
scene units.
Reviewed By: campbellbarton
Maniphest Tasks: T85488
Differential Revision: https://developer.blender.org/D10366
Cycles has supported path-traced subsurface scattering for a while, but while it's
more accurate than other approaches, the increase in noise makes it an expensive option.
To improve this, this patch implements Dwivedi guiding, a technique that is based on
zero-variance random walk theory from particle physics and helps to produce shorter
random walks with more consistent throughput.
The idea behind this is that in non-white materials, each scattering event inside the
medium reduces the path throughput. Therefore, the darker the material is, the lower the
contribution of paths that travel far from the origin is.
In order to reduce variance, Dwivedi guiding uses modified direction and distance sampling
functions that favor paths which go back towards the medium interface.
By carefully selecting these sampling distributions, variance can be greatly reduced, and
as a neat side effect shorter paths are produced, which speeds up the process.
One limitation of just blindly applying this is that the guiding is derived from the
assumption of a medium that covers an infinite half-space. Therefore, at corners or thin
geometry where this does not hold, the algorithm might lead to fireflies.
To avoid this, the implementation here uses MIS to combine the classic and guided sampling.
Since each of those works on one of the three color channels, the final estimator combines
six sampling techniques. This results in some unintuitive math, but I tried to structure
it in a way that makes some sense.
Another improvement is that in areas where the other side of the mesh is close (e.g. ears),
the algorithm has a chance to switch to guiding towards the other side. This chance is based
on how deep the random walk is inside the object, and once again MIS is applied to the
decision, giving a total of nine techniques.
Combining all this, the noise of path-traced subsurface scattering is reduced significantly.
In my testing with the Rain character model and a simple lighting setup, the path-traced
SSS is now actually less noisy than the Christensen-Burley approximation at same render time
while of course still being significantly more realistic.
Differential Revision: https://developer.blender.org/D9932
This resolves build error with Py3.9x, use meson build system
since autoconf/automake are no longer supported.
This adds ninja & meson as dependencies for Linux.
Reviewed By: brecht, sybren
Ref D10282
Getting an "ouput" attribute is equivalent to creating an attribute and
then getting a write attribute. Replace the latter with the former for
consistency with other code, and to decrease the used surface area
of the attribute API to hopefully facilitate future cleanup.
Corrects approximately 36 spelling errors in source variable names.
Differential Revision: https://developer.blender.org/D10347
Reviewed by Hans Goudey
This changes the roughness mapping to better utilize the mip chain resolution.
This improves glossy reflections with small roughness.
Lightcache version bumped because old data does not have the same roughness
mapping and cannot be used.
This modifies the principled BSDF and the Glass BSDF which now
have better fit to multiscatter GGX.
Code to generate the LUT have been updated and can run at runtime.
The refraction LUT has been changed to have the critical angle always
centered around one pixel so that interpolation can be mitigated.
Offline LUT data will be updated in another commit
This simplify the BTDF retreival removing the manual clean cut at
low roughness. This maximize the precision of the LUT by scalling
the sides by the critical angle.
I also touched the ior > 1.0 approximation to be smoother.
Also incluse some cleanup of bsdf_sampling.glsl
This refactor was needed for some reasons:
- closure_lit_lib.glsl was unreadable and could not be easily extended to use new features.
- It was generating ~5K LOC for any shader. Slowing down compilation.
- Some calculations were incorrect and BSDF/Closure code had lots of workaround/hacks.
What this refactor does:
- Add some macros to define the light object loops / eval.
- Clear separation between each closures which now have separate files. Each closure implements the eval functions.
- Make principled BSDF a bit more correct in some cases (specular coloring, mix between glass and opaque).
- The BSDF term are applied outside of the eval function and on the whole lighting (was separated for lights before).
- Make light iteration last to avoid carrying more data than needed.
- Makes sure that all inputs are within correct ranges before evaluating the closures (use `safe_normalize` on normals).
- Making each BSDF isolated means that we might carry duplicated data (normals for instance) but this should be optimized by compilers.
- Makes Translucent BSDF its own closure type to avoid having to disable raytraced shadows using hacks.
- Separate transmission roughness is now working on Principled BSDF.
- Makes principled shader variations using constants. Removing a lot of duplicated code. This needed `const` keyword detection in `gpu_material_library.c`.
- SSR/SSS masking and data loading is a bit more consistent and defined outside of closure eval. The loading functions will act as accumulator if the lighting is not to be separated.
- SSR pass now do a full deferred lighting evaluation, including lights, in order to avoid interference with the closure eval code. However, it seems that the cost of having a global SSR toggle uniform is making the surface shader more expensive (which is already the case, by the way).
- Principle fully black specular tint now returns black instead of white.
- This fixed some artifact issue on my AMD computer on normal surfaces (which might have been some uninitialized variables).
- This touched the Ambient Occlusion because it needs to be evaluated for each closure. But to avoid the cost of this, we use another approach to just pass the result of the occlusion on interpolated normals and modify it using the bent normal for each Closure. This tends to reduce shadowing. I'm still looking into improving this but this is out of the scope of this patch.
- Performance might be a bit worse with this patch since it is more oriented towards code modularity. But not by a lot.
Render tests needs to be updated after this.
Reviewed By: jbakker
Differential Revision: https://developer.blender.org/D10390
# Conflicts:
# source/blender/draw/engines/eevee/eevee_shaders.c
# source/blender/draw/engines/eevee/shaders/common_utiltex_lib.glsl
# source/blender/draw/intern/shaders/common_math_lib.glsl
Removes values from various Outliner context menu enums that were unused.
Many of these had been commented out for years. Also deletes some
unused code related to the removed enums. No functional changes.
A few related improvements to the three functions:
- Reduce variable scope
- Use for loops instead of while loops
- Use const, bool instead of int
- Generally make logic easier to read
This is a complete refactor over the old system. The goal was to increase quality
first and then have something more flexible and optimised.
|{F9603145} | {F9603142}|{F9603147}|
This fixes issues we had with the old system which were:
- Too much overdraw (low performance).
- Not enough precision in render targets (hugly color banding/drifting).
- Poor resolution near in-focus regions.
- Wrong support of orthographic views.
- Missing alpha support in viewport.
- Missing bokeh shape inversion on foreground field.
- Issues on some GPUs. (see T72489) (But I'm sure this one will have other issues as well heh...)
- Fix T81092
I chose Unreal's Diaphragm DOF as a reference / goal implementation.
It is well described in the presentation "A Life of a Bokeh" by Guillaume Abadie.
You can check about it here https://epicgames.ent.box.com/s/s86j70iamxvsuu6j35pilypficznec04
Along side the main implementation we provide a way to increase the quality by jittering the
camera position for each sample (the ones specified under the Sampling tab).
The jittering is dividing the actual post processing dof radius so that it fills the undersampling.
The user can still add more overblur to have a noiseless image, but reducing bokeh shape sharpness.
Effect of overblur (left without, right with):
| {F9603122} | {F9603123}|
The actual implementation differs a bit:
- Foreground gather implementation uses the same "ring binning" accumulator as background
but uses a custom occlusion method. This gives the problem of inflating the foreground elements
when they are over background or in-focus regions.
This is was a hard decision but this was preferable to the other method that was giving poor
opacity masks for foreground and had other more noticeable issues. Do note it is possible
to improve this part in the future if a better alternative is found.
- Use occlusion texture for foreground. Presentation says it wasn't really needed for them.
- The TAA stabilisation pass is replace by a simple neighborhood clamping at the reduce copy
stage for simplicity.
- We don't do a brute-force in-focus separate gather pass. Instead we just do the brute force
pass during resolve. Using the separate pass could be a future optimization if needed but
might give less precise results.
- We don't use compute shaders at all so shader branching might not be optimal. But performance
is still way better than our previous implementation.
- We mainly rely on density change to fix all undersampling issues even for foreground (which
is something the reference implementation is not doing strangely).
Remaining issues (not considered blocking for me):
- Slight defocus stability: Due to slight defocus bruteforce gather using the bare scene color,
highlights are dilated and make convergence quite slow or imposible when using jittered DOF
(or gives )
- ~~Slight defocus inflating: There seems to be a 1px inflation discontinuity of the slight focus
convolution compared to the half resolution. This is not really noticeable if using jittered
camera.~~ Fixed
- Foreground occlusion approximation is a bit glitchy and gives incorrect result if the
a defocus foreground element overlaps a farther foreground element. Note that this is easily
mitigated using the jittered camera position.
|{F9603114}|{F9603115}|{F9603116}|
- Foreground is inflating, not revealing background. However this avoids some other bugs too
as discussed previously. Also mitigated with jittered camera position.
|{F9603130}|{F9603129}|
- Sensor vertical fit is still broken (does not match cycles).
- Scattred bokeh shapes can be a bit strange at polygon vertices. This is due to the distance field
stored in the Bokeh LUT which is not rounded at the edges. This is barely noticeable if the
shape does not rotate.
- ~~Sampling pattern of the jittered camera position is suboptimal. Could try something like hammersley
or poisson disc distribution.~~Used hexaweb sampling pattern which is not random but has better
stability and overall coverage.
- Very large bokeh (> 300 px) can exhibit undersampling artifact in gather pass and quite a bit of
bleeding. But at this size it is preferable to use jittered camera position.
Codewise the changes are pretty much self contained and each pass are well documented.
However the whole pipeline is quite complex to understand from bird's-eye view.
Notes:
- There is the possibility of using arbitrary bokeh texture with this implementation.
However implementation is a bit involved.
- Gathering max sample count is hardcoded to avoid to deal with shader variations. The actual
max sample count is already quite high but samples are not evenly distributed due to the
ring binning method.
- While this implementation does not need 32bit/channel textures to render correctly it does use
many other textures so actual VRAM usage is higher than previous method for viewport but less
for render. Textures are reused to avoid many allocations.
- Bokeh LUT computation is fast and done for each redraw because it can be animated. Also the
texture can be shared with other viewport with different camera settings.
This commit adds a drop-down to the attribute randomize node to support
a few operations on the values of existing attributes: "Replace/Create"
(the existing behavior), "Add", "Subtract", and "Multiply".
At this point, the operations are limited by what is simple to implement.
More could be added in the future, but there isn't a strong use case
for more complex operations anyway, and a second math node can be used.
Differential Revision: https://developer.blender.org/D10269
Currently every attribute node assumes that the attribute exists on the
"points" domain, so it generally isn't possible to work with attributes
on other domains like edges, polygons, and corners.
This commit adds a heuristic to each attribute node to determine the
correct domain for the result attribute. In general, it works like this:
- If the output attribute already exists, use that domain.
- Otherwise, use the highest priority domain of the input attributes.
- If none of the inputs are attributes, use the default domain (points).
For the implementation I abstracted the check a bit, but in each
node has a slightly different situation, so we end up with slightly
different `get_result_domain` functions in each node. I think this makes
sense, it keeps the code flexible and more easily understandable.
Note that we might eventually want to expose a domain drop-down to some
of the nodes. But that will be a separate discussion; this commit focuses
on making a more useful choice automatically.
Differential Revision: https://developer.blender.org/D10389
Removal of Win32 code that allows background windows to receive raw input.
Differential Revision: https://developer.blender.org/D10408
Reviewed by Brecht Van Lommel
Ref T84819
Build System
============
This is an API breaking new version, and the updated code only builds with
OpenColorIO 2.0 and later. Adding backwards compatibility was too complicated.
* Tinyxml was replaced with Expat, adding a new dependency.
* Yaml-cpp is now built as a dependency on Unix, as was already done on Windows.
* Removed currently unused LCMS code.
* Pystring remains built as part of OCIO itself, since it has no good build system.
* Linux and macOS check for the OpenColorIO verison, and disable it if too old.
Ref D10270
Processors and Transforms
=========================
CPU processors now need to be created to do CPU processing. These are cached
internally, but the cache lookup is not fast enough to execute per pixel or
texture sample, so for performance these are now also exposed in the C API.
The C API for transforms will no longer be needed afer all changes, so remove
it to simplify the API and fallback implementation.
Ref D10271
Display Transforms
==================
Needs a bit more manual work constructing the transform. LegacyViewingPipeline
could also have been used, but isn't really any simpler and since it's legacy
we better not rely on it.
We moved more logic into the opencolorio module, to simplify the API. There is
no need to wrap a dozen functions just to be able to do this in C rather than C++.
It's also tightly coupled to the GPU shader logic, and so should be in the same
module.
Ref D10271
GPU Display Shader
==================
To avoid baking exposure and gamma into the GLSL shader and requiring slow
recompiles when tweaking, we manually apply them in the shader. This leads
to some logic duplicaton between the CPU and GPU display processor, but it
seems unavoidable.
Caching was also changed. Previously this was done both on the imbuf and
opencolorio module levels. Now it's all done in the opencolorio module by
simply matching color space names. We no longer use cacheIDs from OpenColorIO
since computing them is expensive, and they are unlikely to match now that
more is baked into the shader code.
Shaders can now use multiple 2D textures, 3D textures and uniforms, rather
than a single 3D texture. So allocating and binding those adds some code.
Color space conversions for blending with overlays is now hardcoded in the
shader. This was using harcoded numbers anyway, if this every becomes a
general OpenColorIO transform it can be changed, but for now there is no
point to add code complexity.
Ref D10273
CIE XYZ
=======
We need standard CIE XYZ values for rendering effects like blackbody emission.
The relation to the scene linear role is based on OpenColorIO configuration.
In OpenColorIO 2.0 configs roles can no longer have the same name as color
spaces, which means our XYZ role and colorspace in the configuration give an
error.
Instead use the new standard aces_interchange role, which relates scene linear
to a known scene referred color space. Compatibility with the old XYZ role is
preserved, if the configuration file has no conflicting names.
Also includes a non-functional change to the configuraton file to use an
XYZ-to-ACES matrix instead of REC709-to-ACES, makes debugging a little easier
since the matrix is the same one we have in the code now and that is also
found easily in the ACES specs.
Ref D10274
Currently moving or changing an object references in a node modifier's
node group does not trigger re-evaluation. Because there is no collection
relation in the dependency graph, we must add the relation to all objects
in the collection individually.
This commit makes the geometry output of the collection info usable.
The output is the geometry of a collection instance, but this commit
adds a utility to convert the instances to real geometry, used in the
background whenever it is needed, like copy on write.
The recursive nature of the "realize instances" code is essential,
because collection instances in the `InstancesComponent`, might have no
geometry sets of their own containing even more collection instances,
which might then contain object instances, etc.
Another consideration is that currently, every single instance contains
a reference to its data. This is inefficient since most of the time
there are many locations and only a few sets of unique data. So this
commit adds a `GeometryInstanceGroup` to support this future optimization.
The API for instances returns a vector of `GeometryInstanceGroup`.
This may be less efficient when there are many instances, but it makes
more complicated operations like point distribution that need to iterate
over input geometry multiple times much simpler.
Any code that needs to change data, like most of the attribute nodes,
can simply call `geometry_set_realize_instances(geometry_set)`,
which will move any geometry in the `InstancesComponent` to new "real"
geometry components.
Many nodes can support read-only access to instances in order to avoid
making them real, this will be addressed where needed in the near future.
Instances from the existing "dupli" system are not supported yet.
Differential Revision: https://developer.blender.org/D10327
Ids stored in the `id` attribute cannot be assumed to be unique. While they
might be unique in some cases, this is not something that can be guaranteed
in general. For some use cases (e.g. generating "stable randomness" on points)
uniqueness is not important. To support features like motion blur, unique ids
are important though.
This patch implements a simple algorithm that turns non-unique ids into
unique ones. It might fail to do so under very unlikely circumstances, in
which it returns non-unique ids instead of possibly going into an endless
loop.
Here are some requirements I set for the algorithm:
* Ids that are unique already, must not be changed.
* The same input should generate the same output.
* Handle cases when all ids are different and when all ids are the same
equally well (in expected linear time).
* Small changes in the input id array should ideally only have a small
impact on the output id array.
The reported bug happened because cycles found multiple objects with
the same id and thought that it was a single object that moved on every
check.
Differential Revision: https://developer.blender.org/D10402
The problem was the stroke was reproject flat to view if the axis was View. Now, if the operation is using depth, the stroke is not reprojected.
Related to T85082
Baking vertex colors per-corner leads to unwanted discontinuities when there is
sampling noise, for example in ambient occlusion or with a bevel shader node for
normals. For this reason the code used to always average results per-vertex.
However when using split normals, multiple materials or UV islands, we do want to
preserve discontinuities. So now bake per corner, but make sure the sampling seed
is shared for vertices.
Fix T85550: vertex color baking crash with split normals, Ref D10399
Fix T84663: vertex color baking blending at UV seams
Create zero-byte manual page files `nasm.1` and `ndisasm.1` such that
nasm's `make install` step succeeds.
Installing nasm requires that its manual pages are built. This requires
local packages `asciidoc` and `xmlto` to be installed.
Not only does `asciidoc` pull in 110 MB of packages (itself +
dependencies), there is also no need for these manual pages. Nasm is
just used for building other dependencies, and not even part of our
precompiled libraries in SVN.
Reviewed By: sebbas
Differential Revision: https://developer.blender.org/D10396
The nodes update function geo_node_volume_to_mesh_update would not run
if it is not the very first node in the tree.
If the update function is not run, there are sockets not cleared from
the SOCK_UNAVAIL flag (but this needs to be done -- these get available
depending on the chosen mode).
Havent tracked down why this was actually updating when it was the first
node in the tree, but now make sure we always get an update by specifing
an appropriate RNA update callback for the property.
Maniphest Tasks: T85558
Differential Revision: https://developer.blender.org/D10403
This patch will share the AOV settings between Cycles and Eevee.
It enable using the AOV name conflict detection of Blender. This
means that unlike how Cycles used to work it isn't possible to add an
AOV with a similar name. Conflicts with internal render pass names will
be indicated with an Warning icon.
Reviewed By: Brecht van Lommel
Differential Revision: https://developer.blender.org/D9774
This makes it so that normals are tagged dirty whenever the position
attribute is requested for writing. This seems like a good default. If the
calling code is aware of normals, it could untag normals when they are
not changed by the operation.
Differential Revision: https://developer.blender.org/D10397
In 2.81 there was a change to increase the performance of viewport
animation rendering. This change would perform the color management on the
GPU if the only 8bit was needed. This saved CPU cycles and data
transfer.
The issue is that in the image editor or when saving the image the CM
will be reapplied. Although the speed is desired, exporting the actual
colors has more priority.
In the ticket there is an analysis that shows that shows that this fix
is the correct short term step to take. It would be better that the
render result is aware of the color space of its buffers so the applying
color management could be skipped when saving to disk or drawing in the
image editor.
The issue with this change is the performance penalty it has.
Reviewed By: Brecht van Lommel
Maniphest Tasks: T79999
Differential Revision: https://developer.blender.org/D10371
In 2.81 there was a change to increase the performance of viewport
animation rendering. This change would perform the color management on the
GPU if the only 8bit was needed. This saved CPU cycles and data
transfer.
The issue is that in the image editor or when saving the image the CM
will be reapplied. Although the speed is desired, exporting the actual
colors has more priority.
In the ticket there is an analysis that shows that shows that this fix
is the correct short term step to take. It would be better that the
render result is aware of the color space of its buffers so the applying
color management could be skipped when saving to disk or drawing in the
image editor.
The issue with this change is the performance penalty it has.
Reviewed By: Brecht van Lommel
Maniphest Tasks: T79999
Differential Revision: https://developer.blender.org/D10371
The previous behavior meant that changing an objects visibility
effectively changed the current mode - which missed necessary
updates for the tool-system (for example).
There was already a check for edit-mode, now expected to all modes.
This makes the test-case described in T83013 work as expected.
Adds a new property called "Confirm On Release" that does what it says,
confirm the action without having to do left-click or to press any other
extra-key which concludes in a more dynamic and efficient way of changing
brush size or strength for example, especially for tablet users.
Reviewed By: campbellbarton
Ref D10233
- Move non-blender build targets into their own section.
- Expand 'make help' text, noting a local 'make deps' overrides.
- Note where the spell checkers word-list is maintained.
- Note on why sqlite is built without 'tcl'.
Use Python 3.8's API for setting the initial configuration.
This replaces a mix of our logic and direct calls to the Python API
and has no user visible changes.
Using the Python API makes the logic easier to follow and provides
utilities such as `PyConfig_SetBytesArgv`
that wasn't available in previous releases.
Note that this uses Python's utf8/wchar_t conversions,
which used to cause problems (see T31506).
Since `Py_UTF8Mode` was set, the systems locale isn't used for decoding,
allowing us to use Python's utility functions that call
`Py_DecodeLocale` internally.
Ref D10382
MeshTest now compares selection between evaluated mesh and expected mesh. This way, we can test more operators
such as `faces_select_linked_flat`
Note: selection comparison intentionally does not happen in BKE_mesh_cmp() on C side but rather on Python side, because
selection is independent of mesh generation.
Reviewed By: calra, mont29
Differential Revision: https://developer.blender.org/D10279
Paths and names of system packages-installed libhpdf on Debian (and
probably its derived distributions?) are slightly different than what
was given to CMake finding script.
Print number of total tests with each test to show how many tests have been executed and how many are left.
Example: `Running test 27/36: PlaneFaceSplitByEdges...`
Reviewed By: calra, mont29
Differential Revision: https://developer.blender.org/D10278
Tests are not identified with indexes, so no need to maintain comments with indexes anymore
Reviewed By: calra, mont29
Differential Revision: https://developer.blender.org/D10277
Having a vertex group in a mesh slowed down unrelated operations
such as selection.
De-duplicating custom-data arrays for layers that contain pointers
can become slow without any benefit as the content never matches.
Use full copies when storing custom-data for edit-mesh undo.
Caused by rB0f95f51361d7.
Similar to T85515, T84717 and their related fixes.
In this case, add missing shading notifier as in rB9274bd457a25.
Maniphest Tasks: T85543
Differential Revision: https://developer.blender.org/D10395
This crash is caused by accessing object data in the kernel at an out of bound index from a deleted instance.
Cycles represents instances as Object nodes sharing the same Geometry node, so we need to tag the GeometryManager for an update if some objects are added or removed as no geometry might have been added or removed in order to properly update the BVH and its associated data arrays.
Regression caused by rBbbe6d4492823.
It's not necessary to check if the link has to be removed,
if it was removed already.
This regression was caused by rB8f707a72e81833bb835324ddc635b29dfbe87a9f.
This commit makes links connected to multi-input sockets spread verticaly
along the socket. Sockets grow if more links are connected and the node
layout updates accordingly. Links are sorted by their incoming angle
to avoid crossing links. Also, link picking is updated to work with
spread links and bezier links.
Currently the multi-input sockets are used in the join geometry node.
The mutli-input sockets look like a vertical rounded rectangle.
Currently they do not support the other custom socket shapes.
Reviewed By Hans Goudey, with cleanup and additional edits
Differential Revision: https://developer.blender.org/D10181
This updates the join node to use the new multi-input socket from D10067.
The change just requires slightly changing the arguments of the exec
function, and changing the socket input list.
Note that this commit does not contain the UI changes, it only allows for
more input links, and combines the two sockets into one. The UI changes
will come next.
Reviewed By: Hans Goudey
Differential Revision: https://developer.blender.org/D10069
The triangle overlap test failed for exactly overlapping triangles.
When none of the segments intersect, testing a if a single corner
is inside the other triangle fails when the triangles share UV
coordinates.
Resolve by comparing the triangle centers.
* USD and OpenVDB headers use deprecated TBB headers, suppress all deprecation
warnings there since we have no control over them.
* For our own TBB includes, use the individual headers rather than the tbb.h that
includes everything to avoid warnings, rather than suppressing all.
This is in anticipation of the TBB 2020 upgrade in D10359. Ref D10361.
Currently, `createTransData` is doing much more than the name implies.
This commit makes it clearer through smaller and more specific functions
what the real purpose of that function.
For the debug version of cython pip was trying to link
against the release version of python for some strange
reason. Passing some flags to explicitly target the
debug version fixes the issue.
Given other platforms do not have different builds for
debug/release this is not an issue there.
This removes indentations from if statements by converting them to early
returns and continue.
Most of the code of brushes and tools has loops with a full indented
body inside of an if, which was also copied into some of the new tools.
Reviewed By: JacquesLucke
Differential Revision: https://developer.blender.org/D10333
This potentially could fix some missed cases in dependency tagging (when
dealing with overrides hierarchies), since relying on tag in ID itself
is not a good idea to check whether an ID has been propcessed or not
(exterior code may have forced that tag on some IDs e.g., which would
prevent them from ever being processed properly).
The local orientation chosen was that of the active object, but as
confirmed in other parts of the code, the orientation of the selected
Bone has priority.
Caused by rB5b34d11b55e0.
Above commit restored the view matrices in ED_view3d_draw_offscreen
_before_ they can be stored in the ImBuff (see ED_view3d_draw_offscreen
/ texture_paint_image_from_view_exec).
Now make restoring the view matrices optional and dont do this in case
of reprojection, so the used matrices can still be saved in the ImBuff
later.
Reviewed By: campbellbarton
Maniphest Tasks: T85395
Differential Revision: https://developer.blender.org/D10331
Fixes an issue in a node setup with the point separate node, where muting
a node that does nothing breaks the operation, resulting in the point
separate not copying the position attrbute to either result.
The fix is straightfoward, it looks just like a typo.
Differential Revision: https://developer.blender.org/D10379
After {rBa0867f05a48e2017a}, it is no longer necessary to pad
structs with 1 and 2 byte sizes. Most of these are geometry node
structs, where we've been avoiding using `node.custom1`.
I only found two others, the freestyle flags in custom data DNA.
Differential Revision: https://developer.blender.org/D10335
This fixes two issues that were preveting normal orientation for working:
- The translation of component of the object matrix should not be
considered when converting a normal to world space.
- Whe using cursor for depth, the depth for the shape should be taken
directly from the cursor (which is already unprojected and updated)
instead of from the brush, which may have not been updated.
Reviewed By: dbystedt, JacquesLucke
Differential Revision: https://developer.blender.org/D10231
Simplification of window creation code to allow greater flexibility.
Differential Revision: https://developer.blender.org/D10311
Reviewed by Brecht Van Lommel
The texture paint mode was accidentally removed from the list
Reviewed By: JacquesLucke
Maniphest Tasks: T83047
Differential Revision: https://developer.blender.org/D10216
The thumb brush was updating the area normal per brush sample, which was
making unstable sampled normals for setting the displacement direction
when the vertices are moved too much from their original positions.
Now it always uses the original normal except when using anchored
strokes. In those cases, the normal always needs to be updated.
Reviewed By: JacquesLucke
Maniphest Tasks: T85079
Differential Revision: https://developer.blender.org/D10214
The following packages also have received updates:
- IDNA 2.10
- CHARDET 4.0.0
- URLLIB3 1.26.3
- CERTIFI 2020.12.5
- REQUESTS 2.25.1
- NUMPY 1.19.5
numpy has gained a hard dependency on cython:
- CYTHON 0.29.21
Notes:
- This only updates the build environment files,
once these are built, Blender can default to Python 3.9.
- The 'm' suffix for Python binaries/libs has been removed.
- The macOS patch in Python 3.7 is has been removed.
Reviewed By: sybren, campbellbarton, sebbas
Ref D10257
The span fill was in multithreaded code, so calculated values were
sometimes reset. The fix is to move FLT_MAX fill outside of parallel_for.
Differential Revision: https://developer.blender.org/D10378
This potentially could fix some missed cases in dependency tagging (when
dealing with overrides hierarchies), since relying on tag in ID itself
is not a good idea to check whether an ID has been propcessed or not
(exterior code may have forced that tag on some IDs e.g., which would
prevent them from ever being processed properly).
The render visibility column in the outliner should be enabled by
default. This change makes it more obvious which objects will be
rendered which can be misleading when only viewport visibility toggles
are shown by default. This commit enables the render visibility column
in all existing and new Outliner editors.
Differential Revision: https://developer.blender.org/D10365
Invalid override properties ( i.e. invalid RNA paths) are not strictly
speaking errors, many things can lead to that situation. Just ignore and
skip those cases.
These are similar to the regular "Combine XYZ" and "Separate XYZ" nodes,
but they work on attributes. They will make it easier to switch between
vector attributes and float attributes.
Differential Revision: https://developer.blender.org/D10308
The repository information for cloning the OpenXR repository, rather than just
downloading the source archive, was outdated.
* Use the OpenXR-SDK repository link, not the OpenXR-SDK-Source one. Only the
former is needed, it contains pregenerated files with minimum dependencies.
The latter contains additional tools and the un-generated source files.
* Update the commit hash to the version we currently use, 1.0.14 (was 1.0.6).
Differential Revision: https://developer.blender.org/D10373
Reviewed by: Bastien Montagne
Properly fully extract
`lib_override_hierarchy_dependencies_recursive_tag` from
`lib_override_linked_group_tag`.
Was supposed to be that way already (see comment of
`lib_override_linked_group_tag`), for some reasons this case was missed.
As shown on the T85383, attempts are made to edit the precision mode key.
But that key was hardcoded.
That key now appears among the custom modal keymap items.
Add a bar between keyframes to indicate that keyframe is still used. This is part of a change to make the rec button be used by gpencil.
Example:
Before:
{F9592704}
After:
{F9592702}
Reviewed By: #user_interface, #grease_pencil, pepeland, Severin
Maniphest Tasks: T85463
Differential Revision: https://developer.blender.org/D10179
This commit is a refactor of the fill tool to solve several problems we had since the first version of the tool.
Changes:
* The filling speed has been improved for each step of the process with the optimization of each algorithm/function.
* New `AutoFit` option to fill areas outside of the viewport. When enable, the total size of the frame is calculated to fit the filling area.
* New support multiframe filling. Now it is possible to fill multiple similar frames in one go.
* New `Stroke Extension` option to create temporary closing strokes. These strokes can be displayed and adjusted dynamically using wheel mouse or PageUp/Down keys.
* Parameter `Resolution` now is named `Precision` and has been moved to topbar.
* `Resolution` now has decimals and can be lower than 1 to allow quick filling in storyboarding workflows. Maximum value has been set as 5.
* Parameter `Simplify` has been moved to Advanced panel.
* Improved fill outline detection. In some cases, the outline penetrated the area to be filled with unexpected results.
* Fixes some corner case bugs with infinite loops.
As a result of this refactor, also these new functionalities has been added.
* New support for multiframe in `Draw` mode. Any drawing in active frame is duplicated to all selected frame.
* New multiframe display mode. Keyframes before or after of the active frame are displayed using onion colors. This can be disable using Onion overlay options.
Since GPencil changes depending on libharu may be committed to master
before SVN libraries for all platforms are in place, avoid build issues.
Extension of D9928.
Reviewed By: sybren
Differential Revision: https://developer.blender.org/D10280
This patch adds support for accessing corner attributes on the point domain.
The immediate benefit of this is that now (interpolated) uv coordinates are
available on points without having to use the Point Distribute node.
This is also very useful for parts of T84297, because once we have vertex
colors, those will also be available on points, even though they are stored
per corner.
Differential Revision: https://developer.blender.org/D10305
Goals:
* Clarify the distinction between builtin and other attributes at the code level.
* Reduce number of places that need to be modified to add more builtin attributes.
* Reduce number of virtual methods that need to be implemented by e.g. `MeshComponent`.
To achieve these goals, this patch implements the concept of "attribute providers".
An attribute provider knows how to give access to attributes on a geometry component.
Each geometry component can have multiple attribute providers, whereby each provider
manages an different set of attributes.
The separation of builtin and other attributes is now done at the attribute provider level.
There are two types of attribute providers. One for builtin attributes and one for all others.
This refactor also helps with T84297.
Differential Revision: https://developer.blender.org/D10341
This patch will show textures in the image editor with the maximum
available resolution determined by the GPU Hardware/Driver.
Currently the size is limited by the user preference texture size limit.
An image user can set the `IMA_SHOW_MAX_RESOLUTION` flag to request
gpu textures in the max supported resolution. When this flag isn't
set the gpu texture is limited by the user preference setting.
When the gl resolution limit is disabled the GPU texture is always
created for the max supported resolution.
Reviewed By: Clément Foucault
Maniphest Tasks: T81206
Differential Revision: https://developer.blender.org/D9160
This avoids adding DNA_ID.h into other headers, recently changed in
cfa48c84d0
Note that other enums could be moved too, this is a smaller change
to avoid indirectly including DNA_ID.h in many places.
This commit moves the property layout callbacks for node types to their
implementation files from `drawnode.c`. This was proposed a while ago in
T75724.
**Benefits**
- Fewer files need to be changed when adding a new node.
- Makes it possible to reuse functions from the node's implementation
in the layout code.
- Except for RNA, all of the node "inputs" are in the same place.
- Code gets shorter overall, avoids the large switch statements.
**Downsides**
- Requires including two UI headers.
- Requires adding an editors dependency to the nodes folder.
This commit only changes function nodes and geometry nodes, more can be
moved later.
Differential Revision: https://developer.blender.org/D10352
The header files in freestyle utilize the using-directive at the global
file scope. This is a bad practice as it pollutes the global name space
causing possible ambiguous reference compilation errors. In particular,
the DNA files that are included by freestyle will cause those ambiguous
reference errors when the developers adds a DNA member with a type name
that also exist in the Freestyle name space, such as Curve and possibly
others.
This patch does the minimal work needed to resolve that by moving the
using-directives from the headers into the corresponding translation
units.
Reviewed By: Brecht
Differential Revision: https://developer.blender.org/D10351
The logic to ensure a valid region state was too aggressive in setting the
region hiding. It would just always update it based on the operator's
`hide_props_region` option, not only when file browser was newly opened.
It's more selective now.
Essentially, `lib_link_all` would perform some post-processing over data
in given `bmain` that **may** fail when not all data from all libraries
has been properly loaded yet.
This happens when `lib_link_all` is called from `read_libraries`, where
the bmains are split by libraries.
Now those post-processing is put into its own utils function, which
asserts that it is only called on a merged bmain.
Bonus point, this will avoid re-runing those not-so-cheap operations
more than once on the same data.
Reproducible in r1442 of Sprite repository when opening
`pro/animation_test/rex/performance/rex_crowdcamping/rex_crowdcamping.lighting.blend`
NOTE: Not so sure why we have to call `lib_link_all` several times (once
for each library, and then once on the whole merged bmain, including
local IDs then). So that it can get called for libs while we still have
that specific .blend file handle around?
In any case, the overhead here is minimal since we do ensure a data-block is
never lib-linked more than once, so this is not a serious concern right now.
Differential Revision: https://developer.blender.org/D10307
This patch will show textures in the image editor with the maximum
available resolution determined by the GPU Hardware/Driver.
Currently the size is limited by the user preference texture size limit.
An image user can set the `IMA_SHOW_MAX_RESOLUTION` flag to request
gpu textures in the max supported resolution. When this flag isn't
set the gpu texture is limited by the user preference setting.
When the gl resolution limit is disabled the GPU texture is always
created for the max supported resolution.
Reviewed By: Clément Foucault
Maniphest Tasks: T81206
Differential Revision: https://developer.blender.org/D9160
The changes are:
- Split conversion of the texture space data to its own file.
- Skip adding keyframes with AutoKeyframes.
- Skip recalculation of the trasform dependencies between objects.
- Skip `special_aftertrans_update_...`.
No real user functional changes
The code takes many turns to get a suitable "target" for the snap to grid.
Perhaps there were other reasons awaited for `transformCenter_from_type`
and `TransCenterData center_cache[5]`.
But since nothing is defined, it is better to simplify the code.
No user functional changes
Simple change that shows all the enums used in the transform code and helps
to better understand the role of the members of the struct TransInfo.
It also allows the IDE to show the name of the value represented by the
member instead of a number that needs to be consulted.
Differential Revision: https://developer.blender.org/D10312
When clamp is enabled, it should clamp between the output min and max
and not between 0 and 1.
Differential Revision: https://developer.blender.org/D10324
`seq_effect_speed_get_strip_content_length()` checked only for number
of inputs of target strip, even if it's not an effect.
Only effects are treated in different way, so check for type as well.
Broken by 93c10797dc
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D10356
This is something what comes after an experiment, which makes behavior
more desirable. Basically, for Lock-to-Selection functionality always
use control point position if any of control point itself or handles
are selected.
Initial patch from Sebastian, modification from Sergey.
Differential Revision: https://developer.blender.org/D10265
These weren't used since ages. Comment out the DNA define (value shouldn't be
reused to avoid compatibility breakage) and remove icon code for the types.
We generally prefer using enums over precompiler defines. This adds an actual
type for the values, which the compiler can use to throw warnings, give better
messages, etc.
* Attempt to improve readability by using focused, coherent helper classes.
* Replace ListBase with blender::Vector, which is more efficient and has a
better API.
* Split user reporting from error checking.
* Use namespace (as we usually do for C++ code).
* Remove unused headers
While Boolean is not guaranteed to work if the operands are not
volume-enclosing (technically: PWN - piecewise constant winding number),
it needs to do something in those cases. This change makes
more cases meet user expectations in T84493, T64544, T83403,
T82642 (though very slow on that one).
The original new boolean code used "generalized winding number"
for this fallback; replaced this with code that uses raycasting.
Raycasting would have been faster, but for unfortunately also
switchd to per-triangle tests rather than per-patch tests since
it is possible (e.g., with Suzanne) to have patches that are
both inside and outside the other shape. That can make it much
slower in some cases, sadly.
The old property never worked as expected because it was impossible expose the data as props.
Now, there are two methods to handle this: weight_get and weight_set
Example use:
```import bpy
ob = bpy.context.active_object
gpd = ob.data
gps = gpd.layers[0].frames[0].strokes[0]
i = 0
print("Weights\n================================")
for pt in gps.points:
gps.points.weight_set(vertex_group_index=0, point_index=i, weight=0.5)
i +=1
i = 0
for pt in gps.points:
weight = gps.points.weight_get(vertex_group_index=0, point_index=i)
print(weight)
i +=1
```
Reviewed By: brecht
Maniphest Tasks: T84967
Differential Revision: https://developer.blender.org/D10177
b3f989
It seems generally preferred to have new files be created with C++.
The only reason I didn't do that when I initially created the files is that I
was unsure about some C-API aspect.
Also use nullptr instead of NULL.
It seems generally preferred to have new files be created with C++.
The only reason I didn't do that when I initially created the files is that I
was unsure about some C-API aspect.
Also includes some minor C++ related cleanup (nullptr instead of NULL, remove
redundant `struct` keyword).
Currently the displayed distance when using the shrink/fatten transform
operator does not respect the scene units (they would always be in
blender units).
This changes makes sure the number is displayed in the correct unit.
Reviewed By: mano-wii
Maniphest Tasks: T85378
Differential Revision: https://developer.blender.org/D10325
Introduced by my commit: rB09709a7e64ff
This bug effectively enabled the NLA when it should be disabled, when
all tracks muted or no strips exist. Specifically, the NLA Action Track
settings would be used (extrapolation, blendmode, influence).
The function `nlastrip_evaluate_transition()` has been slightly
modified to use `nlasnapshot_blend()` instead of it's own special
blending function `nlaeval_snapshot_mix_and_free()`.
No user functional changes
Reviewed By: sybren, #animation_rigging
Differential Revision: https://developer.blender.org/D10221
This implements an operator to flip the contents of an image buffer. It
supports flipping the image horizontally and vertically.
Reviewed By: JacquesLucke
Differential Revision: https://developer.blender.org/D10310
In some python scripts, the coordinates of the points are manipulated, but there are no way to force the recalculation and need hack to force the refresh.
The new api allows to call to the refresh function directly.
example: `gp_stroke.points.update()`
Unfortunately one issue raising another one, this ended up being a
fairly big update, with:
*Fixes:
** Dependencies between libraries are better handled (would fail
previously in some cases when the lib was also available as package).
** Dependency of Boost over Python is now taken into account.
** Fix CMake flags to use local own built boost.
** Fix building numpy.
*Changes:
** We now use pip for all python modules (including numpy).
** Added all missing python modules (requests etc.).
*Updates:
**USD: 20.08, Due to build failing with 20.05 (missing include).
Make the SHADING_PARAMETERS operation on node tree a real no-op.
The function used as a callback was only doing a debug print.
The issue with this function was that an original ID was passed
as an argument. This is a violation of design, and if anything it
should be left up to the implementation to check whether depsgraph
is active.
There are no functional changes for users in the interface. The
only difference is that the debug print will no longer happen.
Resolves modernize-raw-string-literal Clang-Tidy warning
The way warning works is it suggests to use raw literal when
overhead of having escape characters is higher than the overhead
of having raw literal syntax (talking about code size overhead).
This means that the warning will not trigger for "foo\"bar".
Differential Revision: https://developer.blender.org/D10322
This updates fixes the following issues (critical for 2.92):
- Issue that prevented dense 'int' grids from being exported (incorrect clip value)
- Issue with particles outside out of domain bounds (position between -1 and 0) not being deleted
Using this persistent field in `select_more_less_seq__internal` was
inapropriate in this case.
Split select more/less and linked code, because it was mixed in
`select_more_less_seq__internal()`
These operators work in completely different way.
`select_linked_internal()` doesn't use seq->tmp at all.
`seq->tmp` was used to mask strips selected by operator while iterating.
Use GSet to store temporary data instead of `seq->tmp`.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D10326
This node takes a volume and generates a mesh on it's "surface".
The surface is defined by a threshold value.
Currently, the node only works on volumes generated by the
Points to Volume node. This limitation will be resolved soonish.
Ref T84605.
Differential Revision: https://developer.blender.org/D10243
When the pivot axis is not set to auto, the node will try to align the rotation
to vector as best as possible, given the selected rotation axis.
Ref T85211.
Differential Revision: https://developer.blender.org/D10292
This node calculates a distance from each point to the closest position
on a target geometry, similar to the vertex weight proximity modifier.
Mapping the output distance to a different range can be done with an
attribute math node after this node.
A drop-down changes whether to calculate distances from points,
edges, or faces. In points mode, the node also calculates distances
from point cloud points.
Design task and use cases: T84842
Differential Revision: https://developer.blender.org/D10154
Regression introduced by {rBed809866b172} where overlays weren't drawn
when overlays are turned off. The background in wireframe, solid
(always) and material preview and render preview (when not using scene
world) are also part of the overlays.
Caused by rBf83aa830cd00.
Since above commit, only meshes in editmode were considered for drawing
(because
BKE_view_layer_array_from_objects_in_edit_mode_unique_data_with_uvs was
used), but the option needs to work for texture paint mode as well, so
use BKE_view_layer_array_from_objects_in_mode_unique_data instead on
pass the draw_ctx->object_mode.
note: there is no good filter_fn to check if we have UVs if mesh is not
in editmode, this shouldnt cause much of a performance hit though.
Maniphest Tasks: T85396
Differential Revision: https://developer.blender.org/D10319
When transforming multiple strips to limits of sequencer timeline they get
squashed into one channel.
Store selection minimum and maximum channel in TransSeq and limit
transformation so no strip can be transformed beyond timeline boundary.
Reviewed By: Sergey, mano-wii
Differential Revision: https://developer.blender.org/D10013
Replace `long long` with an explicit `int64_t`. This is also what is
used in the FFmpeg headers.
Fixes clang diagnostics warning about wrong format used in the log.
Should be no functional changes.
Move `eUndoStepDir` to `BKE_undo_system.h` and use its values
everywhere.
Note that this also introduce the `STEP_INVALID` value in that enum.
Finally, kept the matching struct members in some lower-level readfile
code as an `int` to avoid having to include `BKE_undo_system.h` in a lot
of unrelated files.
When we have a local override, its linked reference may not be detected
as needing to be overridden anymore.
In that case, if there is no actual override defined by the user, assume
that we can get rid of this local override as part of the resync
process, and use the linked data directly instead.
Make it possible to attach a preview image to an Action.
In the #asset_browser_pose_libraries project, poses will be stored as
individual Action datablocks. Having a thumbnail for each pose is of
course essential.
Reviewed By: mont29
Differential Revision: https://developer.blender.org/D10306
This node outputs true when geometry nodes is currently evaluated
for the viewport and false for final renders.
Ref T85277.
Differential Revision: https://developer.blender.org/D10302
No need to show certain operators when they have no functionality.
For example, when adding a library override from the outliner menu, there should only be an option to add one (i.e. reset, delete, resync operators not relevant for objects without library override).
Reviewed By: mont29
Differential Revision: https://developer.blender.org/D10301
The rotation direction was flipped recently.
This meant that rotation on a single axis gave negative results compared
with previous Blender releases.
This partially reverts a9b53daf23.
Differential Revision: https://developer.blender.org/D10299
Adds an exception so that the key-binding `control + Tab` does not get
duplicated as `command + Tab` on macOS because this is a global system
shortcut for the app switcher.
The key-binding `command + Tab` was shown for the operators “Toggle
Graph Editor“ and “Toggle Dope Sheet” in Dope Sheet and Graph Editor,
but only `control + Tab` actually worked.
Differential Revision: https://developer.blender.org/D10304
Win32 child windows on top of parents. Short-term solution of forcing is_dialog when owned.
Differential Revision: https://developer.blender.org/D9971
Reviewed by Brecht Van Lommel
Simplification of Win32 GHOST_WindowWin32 with improved support for owned windows.
Differential Revision: https://developer.blender.org/D9971
Reviewed by Brecht Van Lommel
When adding a new preset the name would be converted to lower case and
then displayed in the interface in title case. This was confusing
because the name didn't reflect what was typed, and there are many cases
when the name shouldn't be forced into title case (like 8K UHDTV for
example).
This commit leaves the custom preset names in the original casing, and
removes the conversion of filenames to title case for preset lists.
Differential Revision: https://developer.blender.org/D10224
Before committing D10224, bundled preset names that are lowercase need
updating to title case for display in the UI. This changes several of
the preset file names to title case, and fixes a couple of other naming
issues in the preset names. A few strings in the code are also changed
to reflect the new filename of the Blender keymap.
The driver variable `splines[0]` (on a curve)
behaved as if the path was `splines`.
Resolve pointers was well as properties as it's supported by the caller.
Error in original commit that added property support
82b0a9e369.
Refactor
//nlastrip_evaluate_actionclip()// and //nlaeval_blend_value()// into
//nlasnapshot_blend()//, //nlastrip_evaluate_actionclip()//,
//nlasnapshot_from_action()//.
**Motivations**:
* {T83615} Requires reading all pose bone fcurves before being able to
apply pre-blend transforms. The function //nlasnapshot_from_action()//
achieves this. This effectively removed the need to specially handle
Quaternion blend queuing so that code has been removed.
* {D8296} Adds support for keyframe remapping through an upper stack of
strips. Instead of introducing a variant of the form:
//nlastrip_evaluate_actionclip_inverted_get_lower()//,
//nlastrip_evaluate_actionclip()// will later be extended to take an
`evaluation_mode` as input to avoid duplicating the recursion functions
related to //nlastrip_evaluate()//.
* //nlasnapshot_blend()// will eventually have variants of
//nlasnapshot_blend_get_inverted_lower_snapshot()// and
//nlasnapshot_blend_get_inverted_upper_snapshot()// which are all
independent of NlaStrips and NlaTracks, further simplifying the
blending implementation. Ideally, //nlastrip_evaluate()// would get
renamed to //nlasnapshot_blend_strip()// but that'll be a later patch
to avoid unnecessary patches slowing the review of more important
patches.
No User-side Functional changes
Reviewed By: sybren, #animation_rigging
Differential Revision: https://developer.blender.org/D10220
**Problem**:
Translating a strip will truncate it when next to a transition.
**Solution**:
The code only accounted for the prev/next strip for whether it exceeds.
When it was a transition, the exceed logic fails. Now, we use the
nearest non-transition strip instead.
Reviewed By: sybren, #animation_rigging
Differential Revision: https://developer.blender.org/D10083
Dyntopo does not have Face Sets implemented yet, so the entire mesh is
considered a single Face Set. In that case, the check unique face set
function should always return true.
Reviewed By: JacquesLucke
Maniphest Tasks: T85230
Differential Revision: https://developer.blender.org/D10259
This ensured that the sculpt mode deformation is flushed to the mesh
datablock after leaving sculpt mode, so it can be duplicated and
extracted correctly.
Reviewed By: JacquesLucke
Maniphest Tasks: T85136
Differential Revision: https://developer.blender.org/D10254
A missing flush of the deformed PBVH coordinates to the shape key.
Reviewed By: JacquesLucke
Maniphest Tasks: T84370
Differential Revision: https://developer.blender.org/D10174
The edge preview data of the boundary needs to be freed with the
boundary data created for the brush. Also, all data for the
ScultpSession preview that is used to render the cursor needs to be
freed with the ##SculptSession##.
Found by Jacques Lucke
Reviewed By: JacquesLucke
Differential Revision: https://developer.blender.org/D10173
The white rotation Gizmo should always use the View orientation.
Before, if you set an orientation for the gizmo other than the default,
that gizmo would behave like the contraint Z Gizmo.
This is an old bug (strange not to have been reported before).
To define the direction of rotation, the View Vector is used.
However, only the projection matrix was being considered, but in
perspective mode, the position of the pivot also matters.
Good news is, this version support LLVM 11.
But it was a serious pain to get working, even ended having to add an
ugly hack to manually specify the OIIO include dir...
Ref: T83631.
Normally sockets only have one input link. This commit adds the back-end
changes needed to use multiple input links per socket.
Multi-input sockets can be defined with a new flag in `bNodeSocketType`.
The changes necessary to make the sockets work in the geometry nodes
evaluator are generalizing input socket values as a vector of values,
and supporting this in the derived node tree structure.
This patch should contain no functional changes. Two upcoming patches
will use this system for the "Join Geometry" node and expose link picking
and updated display in the UI: D10069 and D10181.
Reviewed By: Jacques Lucke, Hans Goudey
Differential Revision: https://developer.blender.org/D10067
Implements a node to get collection objects.
These objects are then passed along as instances in the node tree.
Follow up tasks:
Multiple nodes does not support instancing yet: T85159
Changing collection offset does not trigger a refresh: T85274
Reviewed By: Jacques, Dalai, Hans
Differential Revision: http://developer.blender.org/D10151
Adds Document Type Identifier configuration to use a system generated
document icon for Blender files on macOS. The system composites the
app’s icon and the extension name onto the document sheet icon
(new style on macOS 11 Big Sur and old style on macOS 10.15 Catalina
and earlier). This is in line with Apple’s Human Interface Guidelines.
Removes the previously used file icon because it is no longer needed.
Differential Revision: https://developer.blender.org/D10267
The 'ui_pan_to_scroll' uses 'WM_event_absolute_delta_y'. For scrolling the List View
we need to invert the direction to respect "natural scroll direction" system preferences.
Differential Revision: https://developer.blender.org/D10291
2021-02-03 17:13:27 +01:00
1606 changed files with 54691 additions and 44204 deletions
/* eq. 2 in distribution of visible normals sampling
* pm = Dw = G1o * dot(m, I) * D / dot(N, I); */
* `pm = Dw = G1o * dot(m, I) * D / dot(N, I);` */
/* eq. 38 - but see also:
* eq. 17 in http://www.graphics.cornell.edu/~bjw/wardnotes.pdf
* pdf = pm * 0.25 / dot(m, I); */
* `pdf = pm * 0.25 / dot(m, I);` */
*pdf=G1o*common;
returnout;
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.
