Allows to open newer files in older Blender after new distortion model
has been added.
It will behave as if this is a polynomial model with all 0 coefficients
which are then being refined and assigned explicitly after solving the
motion.
The loop of datablocks was using the scene datablock (3D View) only, but all others datablocks were ignored.
Now the loop consider any annotation datablock.
During recent refactoring of the edit weight overlay we moved a
assignment before it was valid. Making everything one frame off what
resulted in a flashing frame during TAA, not drawing the overlay until a
second action happened, making overlays too bright.
The reason whas that the painting overlay wasn't initialized in the
first sample, but the draw passes and groups were filled. Resulting in
rendering the overlay twice or not at all.
This change moves the assignment to where it is valid.
Limit offsets, so each strip contains at least 1 frame of content.
Reviewed By: campbellbarton
Differential Revision: https://developer.blender.org/D7288
Feature can be enabled or disabled in timeline view menu item "Show F-Curves".
Author a.monti
Reviewed By: ISS
Differential Revision: https://developer.blender.org/D7205
Don't set 'special preview' or Solo mode if scrubbing in scrubbing region.
Author: a.monti
Reviewed By: Severin
Differential Revision: https://developer.blender.org/D7234
Following up to b555b8d.
Building Blender with hidden symbols but using libraries with visible symbols
was giving linker warnings, specifically for USD. So revert that for now, as
it was not needed for the bugfix.
Hide USD symbols (some of which are not in the USD namespace) to avoid potential
conflicts. May potentially help with AMD OpenCL issues in T74262.
Pointer search buttons created with `uiItemPointerR()` (which allows
also passing a collection property to search in) did not work with the
split property layout (i.e. `uiLayout.use_property_split`).
For example vertex group search buttons typically use this.
Note that decorators (`uiLayout.use_property_decorate`) are not
supported yet. Although if they are enabled, the decorator column is
still created to keep the layout alignment visually intact. Also re-uses
the existing hack to allow placing multiple items in the row before the
decorator column.
Needed for some in-progress changes to the modifier stack UI.
When updating the active button, the search data wasn't updated, so it
kept using the old pointers. A check in `ui_rna_collection_search_cb()`
relied on this incorrect behavior so I had to fix that too. Note that
the previous commit was also needed for the second fix to work.
Caused by c46dcdf887.
If the `uiButStore` data was freed after the buttons/blocks were updated
from previous instances (see `UI_block_update_from_old()`), e.g. by
delaying that to the "afterfuncs" (`ui_apply_but_funcs_after()`), the
data would get lost. As result, the button pointers that the API is
supposed to keep valid would point to freed memory.
This wasn't an issue so far since the API didn't happen to be used this
way. That changes with the next commit.
Rather than using the last state of the tablet, we now query the history of
pointer events so strokes can follow the pen even if Blender does not handle
events at the same rate.
Differential Revision: https://developer.blender.org/D6675
It was sometimes set before reading preferences, now it's passed to GHOST every
time preferences are read.
Differential Revision: https://developer.blender.org/D5641
Resolves the following issues:
- For the first time you save a .blend file, there was no feedback.
- If the file fails to save (eg "No space left on device") the status
bar message replaces the error with an invalid "Saved" message.
While there is a popup, the user may cancel it with mouse motion
and be left with the status bar message saying the file saved.
D7371 by @XDroid with edits.
Removal of 'Translation' checkbox. Enable translation options when selecting non-English languages.
Differential Revision: https://developer.blender.org/D7210
Reviewed by Brecht Van Lommel
The behavior for subpanels was incorrect, and the ideal behavior isn't even clear.
This disables the "collapse all" ctrl click feature for panels that have parents.
Differential Revision: https://developer.blender.org/D7355
These changes only have an effect when the experimental Undo Speedup preference
is enabled.
* For DEG_id_tag_update, accumulate recalc flags immediately before the undo
push happens instead of afterwards. Otherwise the undo state does not
contain enough flags, and the current state may contain too many flags.
This also means we call DEG_id_tag_update after undo with the accumulated
flags to ensure they are flushed to other datablocks.
* For undo, accumulate recalc flags in id->recalc and clear accumulated flags
immediately. Not clearing would cause circular behavior where accumulated
flags may never end up being cleared.
This matches what happens after an undo push where these are also cleared,
indicating that the undo state and current in-memory state match exactly.
* Don't change id->recalc of identical datablocks, it should not be needed.
There is one exception for armatures where pointers across datablocks
exist which otherwise would cause problems. There may be a better solution
to this but it seems to work in agent 327 production files.
* This contains a change in undofile.c to avoid detecting all datablocks as
changed for the first of the two undo steps, where we restore to the state
of the last undo push before going to the one before.
Without this the whole system is much less efficient. However this is unsafe
in the sense that if an app handler or operators edits a datablock after an
undo push, that change will not be undone.
It can be argued that this is acceptable behavior, since a following undo push
will include that change and this may already have unexpected side effects.
Ref T60695
Differential Revision: https://developer.blender.org/D7339
There is no garuantee 'outliner_find_id()' can find those corresponding
TreeElements, safeguard against failure now.
note: not sure why this was no problem in Release builds? (could only
reproduce crashes in Debug builds...)
Maniphest Tasks: T75471
Differential Revision: https://developer.blender.org/D7365
Since rBcf258b02f449, only wires and edges that are mapped to the
original mesh were drawn if the mesh was modified by modifiers.
Above commit was only meant for showing orig wires for paint mask
overlays [where final wireframe is not desired], so now only use
MR_EXTRACT_MAPPED when we are in a paint mode.
Maniphest Tasks: T75343
Differential Revision: https://developer.blender.org/D7333
Issue was that the noise simulation was trying to bake with a minimized domain size (adaptive domain initializes domain with size (1,1,1)). Similarly to the base resolution bake, there should be no noise baking happening at those domain sizes - a domain at this size is considered empty.
This reverts commit 58e20b432a.
Was calling discard twice, tsk.
Although for some reason it did quiet the leak.
Reverting because this is causing a crash.
The symmetrize operator now tries to make sure that the armature
constraints are correctly mirrored.
Before it would only mirror the subtargets for the constraints (and that
failed too in some cases).
Reviewed By: Sybren
Differential Revision: http://developer.blender.org/D6009
Otherwise file browsers do not recognize the Blender executable. This is
already done for official releases.
We leave it off for non-portable builds, since that's how Linux distribution
packages will typically build Blender and we can continue to follow the OS
default there. Using a file browser to launch executables from e.g. /usr/bin
would be rare as wel.
Differential Revision: https://developer.blender.org/D7363
Other types of datablocks pointing to UI datablocks is unsupported, so
there is no need to store them in fd->libmap. With the experimental undo
speedup enabled preserving such pointers was done. But it didn't work in
2.82 and such pointers are easily lost in cases other than undo.
Differential Revision: https://developer.blender.org/D7329
The issue was that the code tried to use the bones transformation matrix
as a rotation matrix. This works fine as long as the scale is 1.
Now we simply make sure that we only get a pure rotation matrix when
extracting it for the bone's transformation matrix.
The root of the issue was caused by mesh which was a result of to_mesh()
had the same edit_mesh pointer as the input object, causing double-free
error.
This fix makes it so result mesh does not have edit mesh pointer.
Motivation part behind this is to make the result of to_mesh() to be
somewhat independent from the input.
Differential Revision: https://developer.blender.org/D7361
Starting select linked failed when the selected vertex or edge
was it's self delimiting.
Support using these elements for linked select
as long as they're part of an isolated selection.
When the modal operator passes events, free the internal state of
the operator as we can't be sure those events don't cause the mesh data
to be re-allocated or removed.
Longer term it might be best to make this into a tool since
the main purpose of this operator is to run other actions.
The 10g Intel/Win driver doesn't work well with our emulated
intermediate mode. This patch alters the drawing of the drag widget of
the panels to reduce unneeded drawing.
The previous method would draw 16 boxes per widget. This new way would
cache this drawing in a GPU batch and just move the matrix around.
Reviewed By: Clément Foucault
Differential Revision: https://developer.blender.org/D7345
Existing code for ungrouping did not correctly handle user counters:
- counter for the group was not decremented
- counters for containing nodes were not incremented
The latter resulted in losing some nodes after orphan cleaning or several save/reload cycles.
The bug did not have destructive consequences until recently,
because it was compensated by another bug (fixed in rBe993667a46c2).
Maniphest Tasks: T74665, T74682
Differential Revision: https://developer.blender.org/D7332
For a real-world distortion the payload is quite uniformly
distributed across scanlines. Surely, in the corners more
iterations of minimizer is needed, but that happens in threads
without scheduling overhead.
Need to communicate available number of threads to the camera
intrinsics implementation, otherwise default value of 1 is used.
Must have been single-threaded for a very long time.
When setting the current playback time in BKE_sound_play_scene we didn't
account for the frame length. So the current frame/time would be wrong
when we asked the audio playback what time it was.
This would lead to playback being offset when using time remapping and
AV sync.
Reviewed By: Richard Antalik and Sybren A. Stüvel
Differential Revision: http://developer.blender.org/D7248
Before this commit, baking an action would only insert keys that are
necessary (i.e. using `INSERTKEY_NEEDED`). When baking to the current
Action, if there are no constraints that influence the final animation,
there are no additional keys necessary. This makes it appear as if
nothing happened. However, when baking to a new Action every additional
frame is necessary and thus a key is added for every frame.
@mont29 and I agreed that this behaviour is confusing, so this commit
changes the behaviour such that baking to the current action and to a
new action result in the same baked animation (that is, keyed on every
frame).
This was already fixed in rB985f33719ce9 once.
But resurfaced after rB7070e4c15e6c [which just reverted a bit too much
;)].
Spotted while looking into T75263.
Differential Revision: https://developer.blender.org/D7308
Splits up wm_xr.c into multiple files in their own folder:
source/blender/windowmanager/xr. So this matches how the message bus and
gizmo code have their own folder and files.
This allows better structuring and should make the code scale better.
I rather do this early on than to wait until we end up with a single,
huge file.
Also improves a bit how data is prepared and updated for drawing.
- Adds select menu
- Removes undo/redo controls
- Adds delete menu
- Refactor
- Combines font and text menu
The goal is to match other edit menus better and match the text editor.
Note this only changes cases where the variable was declared inside
the for loop. To handle it outside as well is a different challenge.
Differential Revision: https://developer.blender.org/D7320
The 'name' parameter of `BKE_idtype_idcode_from_name()`, and the `str`
parameter of `idtype_get_info_from_name()`, are expected to be the
'user visible name' of an `IDTypeInfo` struct. This is made clearer in
the code by renaming those parameters to `idtype_name` and mentioning
it in the documentation of the `BKE_idtype_idcode_from_name()`
function.
Differential Revision: https://developer.blender.org/D7317
The `BKE_animsys.h` and `anim_sys.c` files already had a an "AnimData
API" section. The code in that section has now been split off, and
placed into `BKE_anim_data.h` and `anim_data.c`.
All files that used to include `BKE_animsys.h` have been adjusted to
only include the animation headers they need (sometimes none).
No functional changes.
The island `axismtx` is only necessary in some transform modes.
In the case of `Shrink/Fatten`, the calculated `axismtx` brings an
undesirable result.
This commit rearrange the struct `TransIslandData` in order to
calculate and reference only the arrays that will be used for each
transform mode.
Differential Revision: https://developer.blender.org/D7305
There is no guarantee that depsgraph is ran between two undo steps, so
when re-using an existing data-block we should never wipe completly its
recalc flags, but instead complement them with new ones from accumulated
'storage' as needed.
It should be possible to `#include` any header without having to worry
about its dependencies.
I didn't go and check all include files for this, just the ones that caused
me errors while I was refactoring the `anim_sys.c` file.
No functional changes.
Follow up of b2ee1770d4 and 10c2254d41, part of T74432.
Now the area and region naming conventions should be less confusing.
Mostly a careful batch rename but had to do few smaller fixes.
Also ran clang-format on affected files.
The files are now split up into the following sections:
- `BKE_anim_path.h` and `anim_path.c` for path/curve functions.
- `BKE_anim_visualization.h` and `anim_visualizationanim_path.c` for
animation visualization (mostly motion paths).
- `BKE_duplilist.h` for DupliList function declarations. These were
already implemented in `object_dupli.c`, so they were rather out of
place being declared in `BKE_anim.h` in the first place.
No functional changes.
My compiler (GCC 7.5.0) was warning about these variables potentially not
being initialised. Since the function is highly complex, instead of
analysing it I just trust my compiler and added initial values.
This should be no functional change.
This was only reported for the 'Change Active Layer' operator [which was
not setting the channel as selected in the dopesheet], but this is also
the case elsewhere [where BKE_gpencil_layer_active_set is used], namely:
- gp_layer_remove_exec
- gp_layer_copy_exec
- gp_merge_layer_exec
- gp_layer_change_exec
- gp_layer_active_exec
- gp_stroke_separate_exec
We could set GP_LAYER_SELECT "by hand" in
BKE_gpencil_layer_active_set(), but there is already
animchan_sync_gplayer() that does that. For this, we need the
NA_SELECTED notifier though.
Maniphest Tasks: T75250
Differential Revision: https://developer.blender.org/D7311
There was one function to access both pose/edit bones,
which returned a void pointer type.
Split these into 3 functions which return EditBone, bPoseChannel or Bone
types.
Internally the logic is still shared, this just makes it clearer to
callers which type is expected.
Also use more conventional prefix for picking API:
- ED_armature_pick_(ebone/pchan/bone)
- ED_armature_pick_(ebone/pchan/bone)_from_selectbuffer
In Blender 2.81 we update and draw all nodes inside the view planes.
When navigating with a pen tablet after an operation that tags the whole
mesh to update (like undo or inverting the mask), this introduces some
lag as nodes are updating when they enter the view. The viewport is not
fully responsive again until all nodes have entered the view after the
operation.
This commit delays nodes updates until the view navigation stops, so the
viewport navigation is always fully responsive. This introduces some
artifacts while navigating, so it can be disabled if you don't want to
see them.
I'm storing the update planes in the PBVH. This way I can add support
for some tools to update in real-time only the nodes inside this plane
while running the operator, like the mesh filter.
Reviewed By: jbakker
Differential Revision: https://developer.blender.org/D6269
All DirectX management happens on Ghost level now, higher level code can
just assume everything is OpenGL (except of the upside-down drawing that
still needs to be done for DirectX). This is similar to how the
metal-layer is hidden outside of Ghost.
The Ghost-XR graphics binding for DirectX is responsible for managing
the DirectX compatibility now.
This adds the automasking options to the Sculpt Tool options in a way
that they affect all brushes. This is more convenient when working with
some of these options while switching brushes as they don't need to be
enabled/disabled per brush.
An automasking option is enabled if it is enabled in the brush or in the
sculpt options.
Reviewed By: jbakker
Differential Revision: https://developer.blender.org/D7304
The root cause of the issue reported in T74983 is that an `IDNode` would
not be marked as user-modified. This marking happened while looping over
outgoing relations of one of its operation nodes. Since rBff60dd8b18ed
unused relations are removed, and as a result the `IDNode` would not be
marked.
The solution was to move the responsible code outside the loop; this is
probably a good idea anyway, as the code did not actually use the
looped-over relations at all, and was thus repeated unnecessarily.
With this brush option it is possible to mask the boundary vertices of
all face sets. This is especially useful in the cloth brush, where face
sets can be used to simulate seams between different patches of cloth
and produce different patterns and effects.
Reviewed By: jbakker
Differential Revision: https://developer.blender.org/D7230
Allows to know what level sculpting has been done after the value has
been changed in the MultiresModifierData.
No functional changes, just preparing code to have everything needed
for propagation undo.
Differential Revision: https://developer.blender.org/D7307
Before this change it was not possible to have base geometry
and grid coordinates to be stored in the same undo step.
Differential Revision: https://developer.blender.org/D7298
This fixes the following issues:
- Previously, the face set from the active vertex was used directly. Vertices always return the most recently created face set, so in some cases there may be some face sets that were not possible to select as active. Now the active face set is set in the ray intersection, so it always matches the face under the cursor.
- When drawing face sets they were set per vertex, so it was not possible to paint one face at a time. Now face sets are painted per poly when using the brush on meshes, testing the distance to the center of each poly.
- The code for the active vertex on PBVH_GRIDS was not correct, so I also fixed that to test if everything was working correctly.
{F8441699}
Reviewed By: jbakker
Differential Revision: https://developer.blender.org/D7303
The brush was allocating new memory for storing the displacemnets at the
beginning of each stroke step and not freeing them.
Reviewed By: jbakker
Maniphest Tasks: T75121
Differential Revision: https://developer.blender.org/D7254
As the main use case of this feature is to work with cloth, using this
curve makes more sense than a smoothstep to simulate cloth tension near
the edges.
Reviewed By: jbakker
Differential Revision: https://developer.blender.org/D7262
Due to legacy this overlay was implemented twice (Edit Mesh and Weight
Painting) with different results. This patch consolidates both drawing
and uses only the Weight Painting drawing.
Due to legacy this overlay was implemented twice (Edit Mesh and Weight
Painting) with different results. This patch consolidates both drawing
and uses only the Weight Painting drawing.
Reviewed By: Clément Foucault
Differential Revision: https://developer.blender.org/D7289
Should make it a bit more clear overview of what is going on in this
module. While some of the details might still be missing, having some
sort of top-level overview is better than nothing.
Differential Revision: https://developer.blender.org/D7300
Made it so grids array is properly allocated when first node in the
undo list does not contain grid data.
Differential Revision: https://developer.blender.org/D7299
This patch adds an AVX implementation of Perlin noise in Cycles.
An avxi type was also added as a utility based on the respective
type in Intel Embree.
Only 3D and 4D noise were implemented, there is no benefit for
utilizing AVX in 1D and 2D noise. The SSE trilinear interpolation
function was used in the AVX implementation because there is no
benefit from using AVX in interpolating the last three dimensions.
Differential Revision: https://developer.blender.org/D6680
strips
This uses the new "selected_nla_strips" context member in
UI_context_copy_to_selected_list().
bonus: this also makes the "Copy To Selected" button operator [in the
button context menu] work for anything NLA Strip related.
Maniphest Tasks: T66494
Differential Revision: https://developer.blender.org/D7281
tile
This happened when the UDIM tiled image needed to be colormanaged, so
- when you set up the image as sRGB, then save as EXR/HDR/...
- other way around as well: when you set up the images as Linear then
save as PNG/JPG/...
Reason being that for UDIM tiled images, `image_save_single` is called
multiple times [once for each tile] and everytime `image_save_post` will
fire the `IMA_SIGNAL_COLORMANAGE` signal which clears the cache if any of
the above two is the case. Without the cache, the next tiles cannot be
saved.
Now determine if the colorspace changed from
`image_save_single`/'image_save_post' and only fire
IMA_SIGNAL_COLORMANAGE once from BKE_image_save in the end.
(thx @brecht for suggesting this alternative to the original fix)
Maniphest Tasks: T75234
Differential Revision: https://developer.blender.org/D7296
This basically generalizes what was being done in `write_mesh`,
since we need to clean up ID tags anyway, it's easier to do it for all IDs.
Then ID write funcs themsleves can do whatever they want on the passed
struct, without risking interferring with regular Blender operations.
Note that Text write function is doing a suspicious change on one of its
flags, but this seems to be by-passed anyway by read code currently, so
think it's OK to not do that on orig data-block.
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D7294
Made it so there is a single UNDO node in the list which has
both original and modified mesh state.
Makes it easier to achieve "interleaved" undo nodes stored in
the undo step (as opposite of either storing geometry or other
data).
Should be no functional changes, just preparing for an upcoming
work to support undo of operation like Apply Base.
Differential Revision: https://developer.blender.org/D7290
Actually, begin will do the entire initialization.
Refine will only refine if there is a topology refiner associated
with the Subdiv descriptor.
Allows to refine Subdiv to new coarse positions without touching
displacement evaluation. Will be needed to update SubdivCCG during
sculpt undo.
Change to recent renaming of "Edge Collapse" as it has multiple uses,
as it collapses edge-rings, but isn't limited to collapsing single edges,
it can be used to collapse faces with arbitrary topology.
The name "Collapse Regions" is too vague, users might not think to use
this to collapse edge-rings.
Use a more verbose name "Collapse Edges & Faces", referencing edge-rings
in the tool-tip.
This is useful for collapsing regions of faces & edges,
similar to a 'Merge -> Collapse' which can operate on multiple regions,
merging UV's so they don't need to be manually corrected.
The name & description didn't mention this.
This implements the Sculpt Mode API functions needed for Face Sets and
visibility management for PBVH_GRIDS. No major changes were needed in
the operators and the sculpt mode code. This implementation stores the
face sets in the base mesh, so faces created in higher subdivision
levels can't be modified individually. Also, we are not checking for
multiple face sets per vertex (that can be added in the future), so
relax tools don't work yet. The rest of the features (paint, undo,
visibility operators..) work as expected.
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D7168
This adds the Voxel Mode to the current remesh modifier. It works
exactly the same way as the voxel remesh operator and uses the same
properties to control the remeshing. We can exand this with more options
in the future (fix poles, reprojection...)
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D7292
The original code has two logics, extrude the end points or duplicate points (making new splines).
Now all the logic has been redone by extruding contiguous selected segments.
Fix T47169
Reviewed By: campbellbarton
Differential Revision: https://developer.blender.org/D6982
This operator lets the user control the voxel/detail size of the voxel remesher directly from the 3D view in a similar way the Brush radius and strength are controlled. The shorcut from sculpt mode is Shift + R (similar to Shift + F for brush strength).
It shows a grid that represents the real voxel size of the object. The grid and the text are automatically aligned to the view to avoid rendering all voxels with thousands of lines.
It also has a slow mode when pressing shift that works like the slow mode of the brush radius control.
This operator controls the value changes sensitivity automatically to avoid jumping to extremelly high resolutions and run out of memory.
This way, adjusments done in lower voxel sizes are more precise. Pressing Ctrl disables this functionality and allows changing the voxel size directly in a linear way.
Reviewed By: jbakker, #user_interface, billreynish
Differential Revision: https://developer.blender.org/D6449
The text and icon were supposed to change but didn't reliably, which was
a race condition I think. It depended on how fast the OpenXR runtime
would transition the session state.
This also makes sure the correct notifier is sent on session exit.
Draw batch extraction wrongly assumed that when mapped extraction
happened that all original data could be found. This is not the case as
mapped extraction is also enabled when part of the data is present.
This fix does additional nullptr checks.
EEVEE and Workbench both had the same issue that they continue with the
last sample when leaving navigating. This is ok for regular meshes as
they are all the same. For multiresolution it ain't as a low res version
of the mesh is used during navigation.
This patch also resets the AA samples when the user leaves navigation.
As the 'M' key is free, it's convenient to use for the merge menu,
especially since this contains "Merge by Distance",
a frequently used action.
Use 'Alt-M' for a new split menu, following our convention of Alt being
used for opposite functionality.
Also move merge/split menu's into the "Mesh" menu as neither operate
solely on a single mesh element type.
The edge split operator can now split faces & edges
from selected vertices.
This has the same functionality as manually ripping all
faces and edges away from a vertex.
This means symbols from Blender itself and most external libraries. We can't
just hide all because that breaks some libraries. The better solution would
be to rebuild all library dependencies with hidden visibility.
Fixes T75223: Luxrender add-on failing to load on macOS
Mesh writes a modified copy, which meant recalc_undo_accumulated was never
cleared on the actual datablock. Also clear mesh->runtime on write to avoid
detecting changes, since it's cleared on read anyway.
Differential Revision: https://developer.blender.org/D7274
Clear is_identical_future before adding a next undo step, to avoid wrong
values for cases where we undo and then add a step with different changes
than what was previously the next step.
Ref D7274
When using generative modifiers too many center dots were rendered in
the normal overlay. This patch only renders the normals of original
center dots.
Known issue: decoding the `norAndFlag` has issues on Intel GPU.
The loop normals were always drawn. We used to only draw the normals if
it was mapped to an original loop of the mesh. Due to recent changes we
can not find the correct loop and decide if we need to draw them.
Note still need to check the face dots normals. This is more complicated
as facedot normals needs to be encoded in a different way
This was caused by the removal of some `BONE_TRANSFORM_MIRROR` flag
handling in rBde530a95dc7b482dc22c933b9b8b2a98c79b5663. I simply
restored those lines that caused this issue.
This makes the code a bit simpler to follow, by replacing
`if (x) { all the code here }` with `if (!x) { continue; }` and un-indenting
the remaining code, and by returning early.
No functional changes.
The paint mask overlay showed the wires and edges of the final mesh.
This change will only draw wires and edges that are mapped to the
original mesh.
This change enables mapping data in regular Mesh extraction. This
can also be used for better drawing of the normal overlay.
Reviewed By: Clément Foucault
Differential Revision: https://developer.blender.org/D7277
Selecting linked would only select a single arbitrary chain.
Now select linked follows all child-chains of the bone.
Also add support for following all links, similar to how this would work
if it were a mesh with connected edges instead of only child chains.
Leave this off by default to match pose mode.
This was crashing, when looking into a fix I noticed that it gave
hap-hazard results dissolving past forks in the parent/child hierarchy
arbitrarily following one chain.
This functionality is almost identical to "dissolve" which delimits
forks in the chain predictably.
So remove this in favor of dissolve (available from the delete menu).
This works for all selection modes,
include last below a separator since it's a specialized function.
Note that the previous commit was raising an exception as operator
properties don't support 'or'.
Edges were hard to see in some cases in edit-mesh vertex/face modes.
Since 804e90b42d alpha is handled differently,
update edge alpha to visually match 2.82.
Required for collisions with moving rigid bodies. Otherwise the static optimization mode will be kept and the obstacles would be calculated only once at the beginning.
The button seems to behave more as I'd expect without these
additional checks. Previously, the button was often grayed out when
it was actually working.
Reviewers: ISS
Differential Revision: https://developer.blender.org/D7252
The Screw Modifier had a lower limit for the steps value, which not only
was inconsistent between render and viewport steps, but also was capped
to 2 in UI and also in the code internally.
Removing the GHOST_kKeyUnknown check from processKeyEvent() produces
epeated unknown key events whenever a modifier key is held down, due
to the way ghost uses GHOST_kKeyUnknown as a filter value for modifier
key events.
Differential Revision: https://developer.blender.org/D7257
This commit introduces a new mode for calculating the positions and
weights of the IK segments in the Pose Brush based on the Face Sets.
The first segment of the chain will always include all face sets inside
the brush radius and it will propagate until the boundary of the last
face sets added in the flood fill. Then consecutive connected face sets
are added to the chain until the chain length limit is reached or all
face sets of the mesh are already part of the chain.
This feature enables complete control over the pose brush origins in
case that is needed. Also, with this mode, the user can have a library
of base meshes with face sets already configured to get to the initial
pose as fast as possible.
Reviewed By: jbakker
Differential Revision: https://developer.blender.org/D7235
This makes more clear what this automasking operation does and helps to differenciate it from the future face sets boundary automasking.
Reviewed By: jbakker
Differential Revision: https://developer.blender.org/D7185
Previously, all vertices inside the brush radius were taken into account
equally when calculating the sculpt normal and area. This was causing
artifacts and unpredictable results with large brushes or meshes with
curvatures, as the strongest deformation point of all brushes is usually
in the center. By weighting the vertex normal and position towards the
center when sampling, all brushes should now behave in a more
predictable way in non-uniform surfaces.
Reviewed By: jbakker
Differential Revision: https://developer.blender.org/D6989
Last time I checked Face Sets were preserved in a more or less
predictable way when modifying the mesh with dyntopo. As it looks that
in some problems this may cause bugs and you can't see or use face sets
when modifying the topology of the mesh whith dyntopo active, it is
probably better to reset them when going from dyntopo to mesh. This way
you know that you are always going to get a predictable face sets state.
Reviewed By: jbakker
Maniphest Tasks: T74637
Differential Revision: https://developer.blender.org/D7099
This is not dependent on having an armature as mentioned in T75111.
The collider simply has to be animated.
Reviewers: sebbas
Differential Revision: https://developer.blender.org/D7251
This commit aims to add functionality to the surface deform modifier that
gives more control and allows it to work better with the modifier stack.
* Maintains compatibility with older files. The default settings keep it
so that the whole object is bound and vertex coordinates get overwritten
as the modifier currently does.
* Turns the deformations from an absolute vertex coordinate overwrite into
an additive offset from the vertex location before the modifier to the
resulting bound deformation. This gives the ability to control the
strength of the deformation and mix the deformation of the modifier
with the modifier stack that comes before it.
* Also adds in a vertex group with the invert option. This is applied after
the bind deformation is added. So the whole object is still bound to target,
and the vertex group filters afterwards what parts get affected.
I experimented with a version to only binds the geometry weighted to the
vertex group, but that would break compatibility with old files.
I may bring it in later as a separate option/mode for the surface deform.
With several fixes from @mont29.
Reviewed By: mont29
Differencial Revision: https://developer.blender.org/D6894
Might have happened when Apply Base is used in sculpt mode.
In practice this probably was fine, since the operator tags object
for update, so the pointers will be restored back to what they should
be.
This commit gives the solidify modifier the ability to assign the newly created shell
and rim geometries to selected vertex groups. This expands the procedural control over
the modifier stack by letting users apply modifiers to the shell geometry without affecting
the original geometry.
This will be especially helpful for NPR users that use solidify to create backface
culling lines on their characters giving them the ability to add displace noise
and other effects.
Differential Revision: https://developer.blender.org/D6903
This commit adds the option to use armature bones for the From and To targets
when using armature objects.
The changes are based on the UV Warp modifier.
Reviewed By: mont29
Differential Revision: https://developer.blender.org/D6820
Problem is that the RenderEngines will change the RenderData cfra when
rendering (when time remapping is used -- at least workbench/eevee/
gpencil do a combination of BKE_scene_frame_get() plus
RE_GetCameraWindow() which alters the RenderData cfra).
Later on in the pipeline, the Compositor will use this RenderData cfra
to determine the output file name for the FileOutput node. (In contrast
to this, the 'regular' Output will use the Scene's RenderData -- not the
Render's -- cfra [which hasnt been altered])
It is not entirely clear why RE_GetCameraWindow was setting the cfra on
the Render, but it appears to be legacy OGL rendering related and is not
needed anymore.
Removing this will keep the cfra as needed for the Compositor FileOutput
node.
Caused by rBedb3b7a323a1.
Using evaluate_fcurve_only_curve actually causes quite a bit of slowdown
[6x] compared to bezier forward differencing [which was used prior to
rBedb3b7a323a1]. But full fcurve evaluation is desired with Dynamic
Interpolation Effects [Back/Elastic] since their min/max will not be
captured with forward differencing.
So now gain back speed [using bezier forward differencing] and only do
the full fcurve evaluation for dynamic interpolation effects.
Maniphest Tasks: T74927
Differential Revision: https://developer.blender.org/D7196
operator cancelled.
Lots of operators return OPERATOR_CANCELLED when no data really changed.
Reports from those operators do not show immediately in the Info Editor
[they only do if the operator returns OPERATOR_FINISHED].
Now also notify the Info Editor in case of OPERATOR_CANCELLED.
Maniphest Tasks: T58439
Differential Revision: https://developer.blender.org/D7238
- Use addon keyconfig for registered tools so reloading the keymap
doesn't clear them.
- Ensure there is a default keymap, needed for addon keymaps
to be available in the user keyconfig.
On some drivers, the default values is not respected correctly.
To workaround this we create a small VBO that contains only 1 vec4 worth of
data and just bind it using glBindVertexBuffer to ensure 0 stride.
This fixes T75069 Instances not rendered correctly by workbench.
This adds support for macOS aliases in addition to symlinks. It also adds
support for hidden, readonly and system file attributes.
Contributed by Ankit (ankitm) with modifications by me.
Differential Revision: https://developer.blender.org/D6679
Modify the view layer add operator (and underlying `BKE_view_layer_add`)
to allow for copying the current view layer, as well as adding a new one
but with all LayerCollections disabled by default (this is important for
heavy scenes where currently adding view layers can take a long time due
to enabling every collection by default).
Reviewed By: mont29
Differential Revision: https://developer.blender.org/D6862
This works similarly to the CYCLES_OPENCL_TEST
environment variable to allow testing on unsupported
hardware.
Note: like the OPENCL test override, this is
for *testing* only and bug reports on unsupported
hardware will *not* be accepted at this point in
time.
This implements a new mode in the Face Sets Create operator to create a
new face sets from the faces selection in edit mode. This can be used
when the user considers that the edit mode tools are more convenient for
a more precise control or a certain type of selection, like creating a
face set from a face loop.
Reviewed By: jbakker
Differential Revision: https://developer.blender.org/D7211
This operator initializes all face sets in the sculpt at once using
different mesh properties. It can create face sets by mesh connectivity,
material slots, face normals, UV seams, creases, sharp edges, bevel
weights and face maps.
For properties that are already in the faces, this is implemented as a
loop. Properties that depend on edge attributes use a similar operation
to sculpt flood fill, but using face adjacency instead of edge vertex
connectivity.
As Multires also stores the face sets in the base mesh, this should work
in the face sets Multires implementation without any changes.
This is implemented as a separate operator as this resets the visibility
and creates all face sets at once, while the create face set operator
creates a single face sets, leaving the rest of the face sets in the
mesh as they are.
Reviewed By: jbakker
Differential Revision: https://developer.blender.org/D7209
This implements the Surface Smooth Brush as a mode inside the Smooth tool,
which uses the HC algorithm from "Improved Laplacian Smoothing of Noisy Surface Meshes".
Comparted to the regular smooth brush with laplacian smooth, this brush removes
the surface while preserving the volume of the object.
The smooth result can be controlled by tweaing the original shape preservation,
displacement and iteration count.
The same surface smooth operation is also available as a mesh filter.
Reviewed By: jbakker
Differential Revision: https://developer.blender.org/D7057
Using a float to store and render the mask seems like a waste of memory
without any noticeable difference in the viewport for its use case.
After this commit, the mask and the face sets combined should take the
same amount of GPU memory than only the mask in previous versions.
Reviewed By: jbakker
Differential Revision: https://developer.blender.org/D7148
This new method is only enabled if Overlay Smooth Wire is enabled.
This method gives really nice results but has some downside:
- Require a depth copy or loose the ability to write wire depth to the
depth buffer and have correct depth ordering of wires. This patch use the former, with its associated cost.
- Require some depth sampling and prevent early depth test (i.e: has
some performance impact).
- Has some relatively minor instability with geometry that are perpendicular
to the view and intersecting with other geometry.
Pros:
- Compared to a fullpass approach this is surely going to have less
performance impact and much higher quality.
- Removes the additional vertex offset. (see T74961)
- Fixes all half edges z-fighting.
{F8428014}
{F8428015}
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D7233
SCULPT_FACE_SET_NONE default value is 0 and it is rendered hidden, so
the invert sign operation to show it was not working. Now the show all
function sets this face set to ID 1 before setting its sign.
I also refactored this check in gpu_buffers.
Not related to the reported issue, but the mesh in attached contains non
manifold geometry with hidden loose vertices, so the visibility state
was not syncing correctly to those vertices. Now the toggle operators
checks the current visibility only on the face sets, so no manifold
vertices are ignored (as they are in the rest of operations in sculpt
mode).
Reviewed By: jbakker
Maniphest Tasks: T74780
Differential Revision: https://developer.blender.org/D7188
This fixes multiple issues:
- Adds tag to update shading when changing vertex visibiliyt. This makes the mesh visibility update when the operator ends.
- Sync vertex to face sets no longer requires the pmap, so it does not crash. (Maybe we can initialize the pmap on undo to avoid these problems in the future).
- Sync vertex to face sets now works in a coherent way with the rest of visibility operations. Hide Box and Hide mask now sync the visibility changes to the face sets, so the all the operations are now getting a correct visibility state.
Reviewed By: brecht
Maniphest Tasks: T74761
Differential Revision: https://developer.blender.org/D7187
Brushes are created automatically when the tools is enabled, but this
way it gets correct defaults and it is accesible from scripts.
Reviewed By: jbakker
Maniphest Tasks: T74899
Differential Revision: https://developer.blender.org/D7199
The default face set color is white, so we can skip drawing the default
face set. This allows to enable again the optimization of not drawing
overlays in nodes where the mask is empty.
This will still slow down the viewport when a new face set is created
for the whole mesh or when inverting the mask, like in previous
versions.
I also renamed the function to make more clear that now it is checking
for both mask and face sets.
Reviewed By: brecht
Maniphest Tasks: T74692
Differential Revision: https://developer.blender.org/D7207
This checks that the distance of the current positions of two connected
vertices is not 0 before calculating the correction vectors for those
vertices.
Reviewed By: jbakker
Maniphest Tasks: T74808
Differential Revision: https://developer.blender.org/D7184
When using Face Sets to mask the mesh filter the pmap needs to be
initialized to check the face sets of each vertex, otherwise it will
crash because it is null.
Probably now we should just initalize the pmap when building the PBVH as
almost all tools need it, so we can avoid these crashes in the future.
Reviewed By: jbakker
Maniphest Tasks: T75089
Differential Revision: https://developer.blender.org/D7236
Show control edges stores the control edges in the mesh which is
picked up by the draw manager. When applyng a subsurf (or multires) we
don't want that data present in the base mesh. Any rebuilding of the mesh
would overwrite the data anyway.
This patch introduces a new flag for applying modifiers
that can be checked to ignore storing display specific data in
the base mesh.
Reviewed By: Brecht van Lommel
Differential Revision: https://developer.blender.org/D7163
The normals of loose edges can be non uniform as they aren't normalized.
Checked with what happens with edit loose edges and synchronized the
implementation.
Reviewed By: Brecht van Lommel
Differential Revision: https://developer.blender.org/D7127
Weight paint overlay was not working when XRay was turned on.
The Weight Paint overlay is rendered directly into the default
framebuffer with a depth equal test. This test fails as the depth won't match.
This patch will update the depth buffer in these cases.
Reviewed By: fclem
Differential Revision: https://developer.blender.org/D7176
Events for keys specific to certain keyboard layouts unknown to Blender
were ignored. Now pass them along as unknown key events for which we
can still handle text input, like we already do for Linux and macOS.
Differential Revision: https://developer.blender.org/D7229
Random Walk subsurface scattering did look different with OptiX because transmittance is
calculated based on the hit distance, but the OptiX implementation of `scene_intersect_local`
would return the distance in world space, while the Cycles BVH version returns it in object
space. This fixes the problem by simply skipping the object->world transforms in all the
places using the result of `scene_intersect_local` with OptiX.
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D7232
Happens on macOS 10.15.4 and Xcode 11.4.
The reason of failure is caused by following factors:
- xcodebuild reports full semantic macOS SDK version 10.15.4
- The actual SDK file path will only include major and minor part
of the version (10.15, MacOSX10.15.sdk)
- Previous CMake code of ours expected direct match between SDK
version and file path.
The solution is to make our detection code a bit more flexible and
additionally check for major.minor macOS SDK version in the path.
The specific goal of this change is to get rid of separate code paths
for older and newer Xcode versions.
The version 8.2 is picked since it's the latest version which runs on
macOS 10.11 (which is our current deployment target). If that turns
out too new for some reason the alternative would be to require Xcode
version 5.
The selection in Draw mode works as a quick eraser and must erase only the points selected in that operation and not any previous selected point.
Now, before erase, unselect any previous selected point.
Note: It's planned to split select & erase operators for Draw mode.
Disable functionality reported in T60766 & only partially worked.
This could be used if the key-map was added after Blender started
as a way to customize modal key-maps, however it didn't work with
the add-on enabled on startup.
Add-on key-maps are intended to extend existing key-maps
so they can call the add-on, not as a way to change modal key-maps
for Blender's built-in functionality.
Disable this since it's not needed as add-ons
can't yet define modal key-maps.
The main CMakeLists already requires CMake 3.5, so there is no point of
requiring "newer" CMake on macOS.
This was a code from a while back where CMake 3 was not required on all
platforms.
This issue became visible after fixing other TAA issues recently.
The sample count of the first frame wasn't reset resulting that the
incorrect resolve took place. This issue was already there beforehand,
it is just much clearer during the recent changes.
Now the `taa_sample will be reset when performing an animation playback
in the 3d viewport.
This simplifies freestyle render pipeline integration so we don't have to do
much manual ID user management at all. The complexity here was legacy from
Blender Internal.
Based on fix provided by Sybren A. Stüvl.
This makes it work again at least for the non-UDIM case. For UDIM it's not
great still but I'll consider that a known limitation. A proper solution is
probably to find the closest tile at the start of the stroke and then only
paint in that one tile for the rest of the stroke.
There is no need to have another font embedded in the Blender executable, we
can assume the bundled font exists. In the future we may provide a fallback
if the font specified by the user in the preferences is missing a character,
but that can use our bundled international font.
Differential Revision: https://developer.blender.org/D6854
This means Blender can display more text correctly without having to enable
user interface translation. Previously the quality of the font was lower,
but that has been fixed now.
The font files have now been ungzipped, which results in faster file loading
as Freetype can read only the parts of the file that it needs. Blender download
size should not increase since the release package is compressed.
This includes improvements for Cyrillic characters from the latest DejaVu
Sans fonts from D6960, contributed by Harley Acheson. Fixes T74097.
Differential Revision: https://developer.blender.org/D6854
The numbers here can probably be tweaked to be better, but it's hard to
predict and this should at least avoid excessive memory swapping.
Fixes T57064.
The current viewer pose position as determined by the OpenXR runtime
would be applied as offset. This offset should however only be set when
toggling the positional tracking while the session already runs.
This patch adds support for alpha hash for hair rendering in EEvee. Here's a comparison of with alpha hashing:
{F7588610}
And no alpha hashing:
{F7588615}
Note that this needs "soft shadows" enabled, otherwise shadows will be noisy; here's a render with soft shadows disabled:
{F7588621}
Reviewed By: fclem
Differential Revision: https://developer.blender.org/D5221
Caused by 18b693bdbd, due to lack of thread safety.
Beteween calling BKE_sequencer_cache_get_num_items and BKE_sequencer_cache_iterate
New items could be inserted in the cache.
BKE_sequencer_cache_iterate() now use 2 callbcack functions for initial setup
during which buffers with correct length can be initialized and finally iterating.
Additionally drawing of unselected items was fixed again introduced in 18b693bdbd.
T74662 is reporting quite different symptoms, than I get on my machine, so I am not
entirely sure this is complete fix.
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D7220
Fix T74038, the logic didn't handle the case where there was not any button with focus.
Reviewed By: Julian Eisel
Maniphest Tasks: T74038
Differential Revision: https://developer.blender.org/D7208
Tested with AMD Radeon Pro WX 9100, where it brings performance back to 2.80
level, and combined with recent changes is about 2-15% faster than 2.80 in
our benchmark scenes.
This somehow appears to specifically address the issue where adding more shader
nodes leads to slower runtime. I found no additional speedup by applying this
to change to 2.80 or removing the new shader node code.
Ref T71479
Patch by Jeroen Bakker.
Differential Revision: https://developer.blender.org/D6252
This commit adds the option to invert the resulting weights of the
falloff curve.
There is a workflow used by some to convert a texture mask into
vertex weights by using a custom curve and inverting the points.
This allows the same effect with a single click, and gives the modifier
more procedural functionality.
With minor UI tweaks by @mont29.
Differential Revision: https://developer.blender.org/D6899
We appear to be hitting some limit where adding any amount of code causes a
significant performance regression, no matter what it does. To work around
that a new node level was added.
Ref T71479
```
...\gpu_texture.c(466,7): warning C4555: result of expression not used
...\gpu_texture.c(559,7): warning C4555: result of expression not used
...\gpu_texture.c:1205:72: warning: pointer targets in passing argument 4 of ‘glGetTexLevelParameteriv’ differ in signedness [-Wpointer-sign]
```
When use the dopesheet add layer operator a new blank frame is created. It's very strange to create a layer and don't see anything in dopesheet.
If the layer is added in properties, the frame is not created.
Reviewed by: @pepeland @mendio
This commit adds a `mipmaps` member to the `GPUTexture` struct and also
computes to the memory used by these mipmaps and the memory used for
textures that are created from an external bindcode.
So it solves the following inconsistencies:
- The memory value for mipmaps was not being computed.
- As `GPU_texture_from_bindcode` didn't call
`gpu_texture_memory_footprint_add`, it brought inconsistencies to the
value of the used memory, especially when the texture is freed.
Differential Revision: https://developer.blender.org/D3554
When the TAA is finished the screen can still be redrawn by other
operations without the TAA resets.
If that happened the TAA did add a blank sample to the result as the
scene wasn't drawn, but the was processed.
Reviewed By: Clément Foucault
Differential Revision: https://developer.blender.org/D7226
The theme color is used in code, but not exposed in RNA and therefore
there's no button in the theme editor for it.
Reviewed by: Julian Eisel
Differential Revision: https://developer.blender.org/D7219
Give pointcaches a proper path function which e.g. also resolves
ALT+click (assign to all selected) not working for anything relating to
pointcaches.
This also cleans up the usage of the 'eModifierTypeFlag_UsesPointCache'
flag (removed from the boolean modifier, added to the softbody modifier).
Maniphest Tasks: T64573
Differential Revision: https://developer.blender.org/D7115
the settings
Caused by rBc476c36e4008.
This hooks into the existing FIXME (workaround for a missing update
tagging), reactivates the NS_VIEW3D_GPU notifier (introduced in
rB2ad3d8f158d2 -- but not going anywhere atm.) to check changes to rv3d
rflag which indicated UserStudioLight has changed. To not have updates
all the time, the rflag also needs to be cleared again (see original
rB2ad3d8f158d2).
Maniphest Tasks: T74744
Differential Revision: https://developer.blender.org/D7194
Caused by rBc476c36e4008.
This hooks into the existing FIXME (workaround for a missing update
tagging), needs to also check the clip_state (to detect
changes in DRW_STATE_CLIP_PLANES).
Maniphest Tasks: T74872
Differential Revision: https://developer.blender.org/D7193
Caused by rBc476c36e4008.
This hooks into the existing FIXME (workaround for a missing update
tagging from operators), needs to also check the shading.flag (to detect
changes in V3D_SHADING_MATCAP_FLIP_X).
Differential Revision: https://developer.blender.org/D7192
Caused by rB9516921c05bd.
Dont really see a reason to draw overlays here if overlays are disabled.
Looks like this only affects Face/Vertex mask selection drawing [which
should indeed be hidden when overlays are disabled] next to two
exceptions:
- OVERLAY_paint_vertex_cache_populate draws weights as well [D7176 /
T70807 might be related here, but to me it looks like drawing weights
here is actually not needed at all]
- OVERLAY_paint_texture_cache_populate calls
DRW_cache_mesh_surface_texpaint_get [not sure about this one, this is
also called from workbench_cache_texpaint_populate, looks like this is
not needed when overlays are hidden]
Maniphest Tasks: T74096
Differential Revision: https://developer.blender.org/D7179
The removal of duplicate faces that are created during the handling of
degenerate cases was implemented already but didn't work.
This patch should fix some crashes with the solidify complex mode
related to that.
See D7221 for details.
This has some advantages over operator search:
- Some operators need options set to be usefully accessed.
- Shows key bindings to access menus
(for actions that don't have key bindings themselves).
- Non operator actions such as check-boxes are also shown.
- Menu items can control execution context, using invoke or execute
where appropriate so we can control how the operator runs.
Part of the design task T74157.
This can be tested using the 'Experimental' preferences section
or selected in the key-map editor.
More detailed symptoms: there was no curve cache created for an object
which was used by draw manager.
A bit tricky situation, which involves collection instances and their
proxies.
The root of the problem in the dependency graph was that instanced
collections visibility was not updated when object is requested with
different visibility. So what was happening is that one of the objects
was pulled as an indirect dependency of something invisible, so it
built instanced collections as if the instancer is invisible. After
that the same object was built as visible. Before this fix this was
only update object flags, the instanced collections still believed they
are invisible. Since there is no path via relations which would connect
visible object with instanced objects the visibility flush which is
happening during graph finalization did not "fix" the visibility flags.
This change makes it so instanced collections are updating their
visibility when their instancer's visibility is changing to truth.
This is similar to how collections will accumulate their visibility
when same collection is used from multiple ones with different
visibility.
However, this alone wasn't enough to get crash fixed. This marked
collections as visible, but the geometry component of the curve object
was still considering self as invisible.
This is something tricky, since the code which is responsible for this
issue was added as an optimization in afb4da6650. This looks like like
an oversight in that commit since it's rather weird that ID node's
flag would depend on construction order (in "normal" object builder the
ID node's directly_visible flag is initialized to object's visibility).
So it seems logical to get this part of code in sync between "regular"
and "accumulative" object builder.
And last but not least the naming is_directly_visible is old and does
not really represent what it actually mans now: a more correct name
would be "will be used by the draw manager".
Differential Revision: https://developer.blender.org/D7217
When the TAA is finished the screen can still be redrawn by other
operations without the TAA needs to be reset.
If that happened the TAA did add a blank sample to the result.
This patch will add an early exit in the case TAA was finished. Note
that there are still some cases still not working. The overlay engine
can in certain circumstances draw directly into the default_fb what can
lead to render artifacts.
This patch adds ability to set up colors and size of background
(transparency) checkerboard pattern in viewport and 2d editors. No new
backgrounds, only changing colors in existing ones.
This is not the background of the viewport, it is a transparency
checkerboard that is turned on only in render mode, when the
transparency mode is on. And also in 2D-editors, (image, sequencer,
etc).
Reviewed By: Pablo Vazquez, Julian Eisel
Differential Revision: https://developer.blender.org/D6791
this patch will change the behaviour of the material offset in complex mode to fit simple mode output.
Previously in complex mode this would offset the material of the enire shell,
because when you read the tooltip it says material for new generated geometry.
In complex mode everything is new generated geometry though.
In simple mode on the other hand, this would give you a way to only change the inside faces
material. There may be cases in large modifier stacks where material offset like it is implemented
currently in complex mode may be useful, but it is much more useful in the way it is implemented
by simple mode.
Reviewed By: mont29
Differential Revision: https://developer.blender.org/D7215
The `build_driver_relations()` function in the depsgraph relations builder
adds relations between drivers that potentially write to the same memory
location. This of course is only useful when there are two or more drivers.
For In Front Objects we need to use the in front depth buffer.
This patch will use the in front depth buffer and also makes sure that
it is filled with the center pixel depth.
Reviewed By: fclem
Differential Revision: https://developer.blender.org/D7216
When using the image editor the display channels attribute can become
invalid when selecting another image/buffer. This patch will check what
display channels are valid and when an invalid channel is selected it
will fall back to the color channel.
To de-duplicate the code it also introduces a
`ED_space_image_get_display_channel_mask` function that will determine
the valid bitflags for the display channel of a given `ImBuf`.
Started to happen after recent fix for T72632.
Was caused by runtime fields backup doing an early exit in the case the
given ID was never expanded by the Copy-on-Write mechanism, but it was
not done int the backup restore function (since it was not possible to
know "locally").
Now both init() and restore() will do an early exit when the ID had
nothing to be backed up.
On some platforms the stereo viewport rendering was not working. The
issue was that the fragment shader and vertex shaded didn't match. Some
platforms will remove the non-matching in/out parameters and blender
needs to provide only the optimal set of parameters. Other platform
still want to receive data for the parameters that aren't used.
This fix uses the correct vertex shader that matches the fragment shader
making both platforms render the same result.
Fixes an issue with secondary particles being out of sync with the main simulation. Cleaned up the secondary particle code in general too (making sure that all solver attributes - timestep, framelength, etc. - are set correctly).
Actually, the fade objects always fade to Black color, but this is not a good solution.
This patch fade the object to the viewport color.
Differential Revision: https://developer.blender.org/D7206
Back in 2.79 you could either use the debug panel or an
environment variable to override using OpenCL for unsupported
hardware. Which was rather useful for developers when testing
on NVidia just to be sure the CL kernels at-least build properly.
This broke in rB949ab753bb2
This diff restores testing though the CYCLES_OPENCL_TEST
environment variable.
Differential Revision: https://developer.blender.org/D7202
Reviewers: brecht
Add new option to change the stroke hardeness. This option works at stroke level, not at point level.
Also replaced the "Both" name mode by "Stroke and Fill".
Differential Revision: https://developer.blender.org/D7195
Parenting in the outliner via drang and drop would always happen without
the 'Keep Transforms' option. Since this is often desired, this adds the
ability to hold Alt for doing this to the drop action.
Adding the hint to hold Alt to the operator name is not nice, but since
the operator name is used for the UI, there doesnt seem to be a nicer
way of doing this.
If modifier keys are needed back for other actions, spawning a menu
instead could be an alternative for the future.
Maniphest Tasks: T74649
Differential Revision: https://developer.blender.org/D7120
It is not guaranteed that with Multires modifier existing there
will be CD_MDISPS and CD_GRID_PAINT_MASK custom data layers.
Fixes assert in the following scenario:
- With default cube, go to edit mode
- Add Multires modifier
- Remove the Multires modifier
The Lens metadata stamped on rendered images was wrong when the camera lens
is animated. This was caused by the render pipeline passing the original
camera to the metadata system, and not the evaluated camera.
When using grease pencil in a stereo rendering the grease pencil objects are
only visible in the left eye. In the viewport it renders both.
Issue is related that `DRW_render_gpencil` only renders a single view. But
`DRW_render_to_image` renders all views. This patch puts this in a loop to
render both eyes.
Reviewed By: fclem
Differential Revision: https://developer.blender.org/D7154
When in editmode faces can be hidden, but in object mode these faces are
still visible. The flag if a face was hidden in edit mode is stored in
object mode, but should not be used.
The edge detection gpu batch did detect hidden faces and didn't add them
to the draw batch. The edge detection gpu batch is used for workbench
shadows, custom bone shapes and object outlines.
This patch adds all faces to the edge detection batch.
Reviewed By: fclem
Differential Revision: https://developer.blender.org/D7157
Mpving utils from idcode to idtype proved to be somewhat painful for
some reasons, but now all looks good.
Had to add a fake/empty shell for the special snowflake too,
`ID_LINK_PLACEHOLDER/INDEX_ID_NULL`...
This is a vector that does not own the memory its elements are stored in.
Instead, a linear allocator is passed to every call to `append`, that will
be used if the vector has to grow. Since a linear allocator does not support
deallocation, no memory is dealloced until the linear allocator is destructed.
This structure can be used when many small vectors of initially unknown
size are required.
The internal implementation is working, but not very efficient yet (every string
is allocated separately without any cache). I'm still thinking about a better
solution, but for now having any working text socket is more important.
This allocator is necessary because sometimes we don't want to use
Blenders guarded allocator. Especially when an object lives longer
than the guarded allocator.
All tests are runnings still, but I need to do more testing in the upcoming days.
This implementation uses a different approach for sharing code between
the different structures. I tried different abstractions in the last weeks
and this is the best one I found so far.
Types can be expected to stay alive until Blender exists.
This also works well with the fact that they are identified
by their pointer.
Furthermore, the number of types is fairly low, but they
are passed around a lot. Maintaining a thread-safe reference
count seems to be unnecessary overhead.
Instead of using a templated list, it is easier to use a "generic" list, that can
handle every type with a CPPTypeInfo extension. This simplifies the code quite
a bit, because templates can be avoided in many places.
This also makes it easier to add new types.
The list will still be refcounted and still have the shared immutability semantic.
So when a list has more than one user, it is immutable.
This simplifies the code and speeds up accessing function
bodies and type extensions. Now, a lock is only necessary
when create new bodies or extensions.
This is achieved at the cost of having a slighly less generic
system, because we have to give e.g. different function body
types different IDs. Since the number of different types
is expected to be fairly low, this trade off is worth it.
The fact that the data structure has small object optimization
is not important enough to justify a more complex name.
Following commits will rename other data structures.
This reverts commit eddb4d0af4.
Sometimes the evaluated node trees are not updated when the real
node tree changes. Use original trees until this is fixed.
Depending the use case, unlinked node inputs should be handled differently.
Sometimes, it is okay to load them from RNA on every call, but
sometimes they should also be hardcoded into the generated assembly.
While this might be less efficient, it is much more comfortable to work with.
Also it might be possible to make it a bit more efficient again in the future.
This is achieved by storing strings in a single buffer instead
of using `std::string`, which does small object optimization
for every string individually. This optimization can be done
because all strings are known when the function is build and
don't change later on.
The main goal is to be able to traverse the graph very efficiently
once it is build. This is hard when the structure has to be dynamic
(i.e. it can be changed). The solution is to use a builder that
can dynamically construct the graph. Once everything is set up,
the graph is converted into the compact form for further use.
This implements a tree wide synching procedure, that corrects
all invalid things it can find. Also, trees are updated in
the correct order, when they depend on each other.
Function input/output use the same structure to declare
sockets as other nodes now.
The type inferencer can be invoked in the panel on the
right in the node editor.
The list nodes work with floats, vectors and ints.
echo Visual Studio is detected but the "Desktop development with C++" workload has not been instlled
echo.
echo Visual Studio is detected but no suitable installation was found.
echo.
echo Check the "Desktop development with C++" workload has been installed.
echo.
echo If you are attempting to use either Visual Studio Preview version or the Visual C++ Build tools, Please see 'make help' on how to opt in to those toolsets.
The C and Python binding for audaspace were designed with simplicity in mind. This means however that to use the full capabilities of audaspace, there is no way around the C++ library.
The C and Python binding for audaspace were designed with simplicity in mind.
This means however that to use the full capabilities of audaspace,
there is no way around the C++ library.
Simple Demo
-----------
The **simple.py** example program contains all the basic building blocks for an application using audaspace. These building blocks are basically the classes :class:`aud.Device`, :class:`aud.Sound` and :class:`aud.Handle`.
The **simple.py** example program contains all the basic
building blocks for an application using audaspace.
These building blocks are basically the classes :class:`aud.Device`,
:class:`aud.Sound` and :class:`aud.Handle`.
We start with importing :mod:`aud` and :mod:`time` as the modules we need for our simple example.
We start with importing :mod:`aud` and :mod:`time`
as the modules we need for our simple example.
..code-block::python
#!/usr/bin/python
importaud,time
The first step now is to open an output device and this can simply be done by allocating a :class:`aud.Device` object.
The first step now is to open an output device and this
can simply be done by allocating a :class:`aud.Device` object.
..code-block::python
device=aud.Device()
To create a sound we can choose to load one from a :func:`aud.Sound.file`, or we use one of our signal generators. We decide to do the latter and create a :func:`aud.Sound.sine` signal with a frequency of 440 Hz.
To create a sound we can choose to load one from a :func:`aud.Sound.file`,
or we use one of our signal generators. We decide to do the latter
and create a :func:`aud.Sound.sine` signal with a frequency of 440 Hz.
..code-block::python
sine=aud.Sound.sine(440)
..note:: At this point nothing is playing back yet,:class:`aud.Sound` objects are just descriptions of sounds.
..note:: At this point nothing is playing back yet,
:class:`aud.Sound` objects are just descriptions of sounds.
However instead of a sine wave, we would like to have a square wave to produce a more retro gaming sound. We could of course use the :func:`aud.Sound.square` generator instead of sine, but we want to show how to apply effects, so we apply a :func:`aud.Sound.threshold` which makes a square wave out of our sine too, even if less efficient than directly generating the square wave.
However instead of a sine wave, we would like to have a square wave
to produce a more retro gaming sound. We could of course use the
:func:`aud.Sound.square` generator instead of sine,
but we want to show how to apply effects,
so we apply a :func:`aud.Sound.threshold`
which makes a square wave out of our sine too,
even if less efficient than directly generating the square wave.
..code-block::python
@@ -40,13 +56,19 @@ However instead of a sine wave, we would like to have a square wave to produce a
..note:: The :class:`aud.Sound` class offers generator and effect functions.
The we can play our sound by calling the:func:`aud.Device.play` method of our device. This method returns a :class:`aud.Handle` which is used to control the playback of the sound.
The we can play our sound by calling the
:func:`aud.Device.play` method of our device.
This method returns a :class:`aud.Handle`
which is used to control the playback of the sound.
..code-block::python
handle=device.play(square)
Now if we do nothing else anymore the application will quit immediately, so we won't hear much of our square wave, so we decide to wait for three seconds before quitting the application by calling :func:`time.sleep`.
Now if we do nothing else anymore the application will quit immediately,
so we won't hear much of our square wave,
so we decide to wait for three seconds before
quitting the application by calling :func:`time.sleep`.
..code-block::python
@@ -55,29 +77,47 @@ Now if we do nothing else anymore the application will quit immediately, so we w
Audioplayer
-----------
Now that we know the basics of audaspace, we can build our own music player easily by just slightly changing the previous program. The **player.py** example does exactly that, let's have a short look at the differences:
Now that we know the basics of audaspace,
we can build our own music player easily
by just slightly changing the previous program.
The **player.py** example does exactly that,
let's have a short look at the differences:
Instead of creating a sine signal and thresholding it, we in fact use the :func:`aud.Sound.file` function to load a sound from a file. The filename we pass is the first command line argument our application got.
Instead of creating a sine signal and thresholding it,
we in fact use the :func:`aud.Sound.file` function to load a sound from a file.
The filename we pass is the first command line argument our application got.
..code-block::python
sound=aud.Sound.file(sys.argv[1])
When the sound gets played back we now want to wait until the whole file has been played, so we use the :data:`aud.Handle.status` property to determine whether the sound finished playing.
When the sound gets played back we now want to wait until
the whole file has been played, so we use the :data:`aud.Handle.status`
property to determine whether the sound finished playing.
..code-block::python
whilehandle.status:
time.sleep(0.1)
We don't make any error checks if the user actually added a command line argument. As an exercise you could extend this program to play any number of command line supplied files in sequence.
We don't make any error checks if the user actually added a command
line argument. As an exercise you could extend this program to play
any number of command line supplied files in sequence.
Siren
-----
Let's get a little bit more complex. The **siren.py** example plays a generated siren sound that circles around your head. Depending on how many speakers you have and if the output device used supports the speaker setup, you will hear this effect. With stereo speakers you should at least hear some left-right-panning.
Let's get a little bit more complex. The **siren.py** example
plays a generated siren sound that circles around your head.
Depending on how many speakers you have and if the output
device used supports the speaker setup, you will hear this effect.
With stereo speakers you should at least hear some left-right-panning.
We start off again with importing the modules we need and we also define some properties of our siren sound. We want it to consist of two sine sounds with different frequencies. We define a length for the sine sounds and how long a fade in/out should take. We also know already how to open a device.
We start off again with importing the modules we need and
we also define some properties of our siren sound.
We want it to consist of two sine sounds with different frequencies.
We define a length for the sine sounds and how long a fade in/out should take.
We also know already how to open a device.
..code-block::python
@@ -88,27 +128,35 @@ We start off again with importing the modules we need and we also define some pr
device=aud.Device()
The next thing to do is to define our sine waves and apply all the required effects. As each of the effect functions returns the corresponding sound, we can easily chain those calls together.
The next thing to do is to define our sine waves and apply all the required effects.
As each of the effect functions returns the corresponding sound,
The next step is to connect the two sines, which we do using the :func:`aud.Sound.join` function.
The next step is to connect the two sines,
which we do using the :func:`aud.Sound.join` function.
..code-block::python
sound=high.join(low)
The generated siren sound can now be played back and what we also do is to loop it. Therefore we set the :data:`aud.Handle.loop_count` to a negative value to loop forever.
The generated siren sound can now be played back and what we also do is to loop it.
Therefore we set the :data:`aud.Handle.loop_count` to a negative value to loop forever.
..code-block::python
handle=device.play(sound)
handle.loop_count=-1
Now we use some timing code to make sure our demo runs for 10 seconds, but we also use the time to update the location of our playing sound, with the :data:`aud.Handle.location` property, which is a three dimensional vector. The trigonometic calculation based on the running time of the program keeps the sound on the XZ plane letting it follow a circle around us.
Now we use some timing code to make sure our demo runs for 10 seconds,
but we also use the time to update the location of our playing sound,
with the :data:`aud.Handle.location` property, which is a three dimensional vector.
The trigonometic calculation based on the running time of the program keeps
the sound on the XZ plane letting it follow a circle around us.
..code-block::python
@@ -119,33 +167,54 @@ Now we use some timing code to make sure our demo runs for 10 seconds, but we al
As an exercise you could try to let the sound come from the far left and go to the far right and a little bit in front of you within the 10 second runtime of the program. With this change you should be able to hear the volume of the sound change, depending on how far it is away from you. Updating the :data:`aud.Handle.velocity` property properly also enables the doppler effect. Compare your solution to the **siren2.py** demo.
As an exercise you could try to let the sound come from the far left
and go to the far right and a little bit in front of you within the
10 second runtime of the program. With this change you should be able
to hear the volume of the sound change, depending on how far it is away from you.
Updating the :data:`aud.Handle.velocity` property properly also enables the doppler effect.
Compare your solution to the **siren2.py** demo.
Tetris
------
The **tetris.py** demo application shows an even more complex application which generates retro tetris music. Looking at the source code there should be nothing new here, again the functions used from audaspace are the same as in the previous examples. In the :func:`parseNote` function all single notes get joined which leads to a very long chain of sounds. If you think of :func:`aud.Sound.join` as a function that creates a binary tree with the two joined sounds as leaves then the :func:`parseNote` function creates a very unbalanced tree.
The **tetris.py** demo application shows an even more
complex application which generates retro tetris music.
Looking at the source code there should be nothing new here,
again the functions used from audaspace are the same as in the previous examples.
In the :func:`parseNote` function all single notes get joined which leads
to a very long chain of sounds. If you think of :func:`aud.Sound.join`
as a function that creates a binary tree with the two joined sounds as
leaves then the :func:`parseNote` function creates a very unbalanced tree.
Insted we could rewrite the code to use two other classes::class:`aud.Sequence` and :class:`aud.SequenceEntry` to sequence the notes. The **tetris2.py** application does exactly that. Before the while loop we add a variable that stores the current position in the score and create a new :class:`aud.Sequence` object.
Insted we could rewrite the code to use two other classes:
:class:`aud.Sequence` and :class:`aud.SequenceEntry` to sequence the notes.
The **tetris2.py** application does exactly that.
Before the while loop we add a variable that stores the current position
in the score and create a new :class:`aud.Sequence` object.
..code-block::python
position=0
sequence=aud.Sequence()
Then in the loop we can create the note simply by chaining the:func:`aud.Sound.square` generator and :func:`aud.Sound.fadein` and :func:`aud.Sound.fadeout` effects.
Then in the loop we can create the note simply by chaining the
:func:`aud.Sound.square` generator and :func:`aud.Sound.fadein`
The entry returned from the :func:`aud.Sequence.add` function is an object of the :class:`aud.SequenceEntry` class. We can use this entry to mute the note in case it's actually a pause.
The entry returned from the :func:`aud.Sequence.add`
function is an object of the :class:`aud.SequenceEntry` class.
We can use this entry to mute the note in case it's actually a pause.
..code-block::python
@@ -158,9 +227,14 @@ Lastly we have to update our position variable.
position+=length
Now in **tetris2.py** we used the :data:`aud.SequenceEntry.muted` property to show how the :class:`aud.SequenceEntry` class can be used, but it would actually be smarter to not even create a note for pauses and just skip them. You can try to implement this as an exercise and then check out the solution in **tetris3.py**.
Now in **tetris2.py** we used the :data:`aud.SequenceEntry.muted`
property to show how the :class:`aud.SequenceEntry` class can be used,
but it would actually be smarter to not even create a note for pauses and just skip them.
You can try to implement this as an exercise and then check out the solution in **tetris3.py**.
Conclusion
----------
We introduced all five currently available classes in the audaspace Python API. Of course all classes offer a lot more functions than have been used in these demo applications, check out the specific class documentation for more details.
We introduced all five currently available classes in the audaspace Python API.
Of course all classes offer a lot more functions than have been used in these demo applications,
check out the specific class documentation for more details.
@@ -38,17 +38,19 @@ class GHOST_EventCursor : public GHOST_Event {
* \param type The type of this event.
* \param x The x-coordinate of the location the cursor was at the time of the event.
* \param y The y-coordinate of the location the cursor was at the time of the event.
* \param tablet The tablet data associated with this event.
*/
GHOST_EventCursor(GHOST_TUns64msec,
GHOST_TEventTypetype,
GHOST_IWindow*window,
GHOST_TInt32x,
GHOST_TInt32y)
GHOST_TInt32y,
constGHOST_TabletData&tablet)
:GHOST_Event(msec,type,window)
{
m_cursorEventData.x=x;
m_cursorEventData.y=y;
m_cursorEventData.tablet=window->GetTabletData();
m_cursorEventData.tablet=tablet;
m_data=&m_cursorEventData;
}
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.
