This was partially fixed in ef844bad but left some unhandled corner cases with
uninitialized memory.
The core problem is that the `dst_curve_mask` used by the interpolation
functions can contain indices that don't actually exist in either the "from" or
"to" source drawings. Specifically when the "from" drawing has more curves, the
indices are still used but the supplemental `to_curve_indices` array contains
`-1` entries, indicating that only the "from" curve should be used.
The main copy code for positions takes this into account, but the generic copy
of attributes below that does not! It passes the `dst_curve_mask` straight into
the `array_utils` functions and that causes crashes. The original fix in
ef844bad was to clamp the size of the index mask, but the way this was done
could lead to empty index masks, leaving attribute values uninitialized.
The correct solution is to use index masks that exclude invalid entries for the
respective source curves. The new masks are non-overlapping, so the full set of
destination curves can be filled by combining these masks.
Pull Request: https://projects.blender.org/blender/blender/pulls/134865
Was reported for alt-editing, thats the same code though.
Previously ( da1038c768 ), the check for "matching" nodes was based on
the `legacy_type`.
This does not uniquely identify a node type, only the `idname` does.
By "matching", I mean having the same properties (so later setting
values can happen on all selected nodes).
If we allow "non-matching" nodes into our list in
`UI_context_copy_to_selected_list`, setting values on non-existent
properties later can act up (throwing `AttributeError` , also
`RNA_property_boolean_set` would actually try to create an equally named
IDP).
Better to exclude these as soon as possible, now remove nodes from
"selected_nodes" based on the `idname` (rather than `legacy_type`), this
way we can be sure properties match.
Pull Request: https://projects.blender.org/blender/blender/pulls/134930
with #133413 the intent was that VSE Text strips would not use the
fallback font stack if using a custom (non-default) font. However this
determination was done by comparing the font id. This was very weak as
the id can vary quite a bit within the first few fonts. This PR instead
adds a BLF function (BLF_is_builtin) that uses BLF_DEFAULT font flag
instead.
Pull Request: https://projects.blender.org/blender/blender/pulls/135014
Text characters in the CO Controls range, 0-31, are not meant to have
any displayable representation. For most Blender code this hasn't
mattered since we usually specifically exclude this range in text
processing. But for times when we don't we need to avoid rendering
glyphs in this range as some fonts contain blank (but not empty) items
here.
Pull Request: https://projects.blender.org/blender/blender/pulls/135013
Technically a regression in [0] since switching the pivot between
the "Median Point" and "Active Element" would previously use the
same orientation with a single face selected.
However the same issue existed when multiple faces were selected so
changing the pivot type can still change the orientation in some
situations.
Restore the previous behavior by removing logic to pick the "best"
tangent when a single face is selected.
[0]: c7104f5392
Crash in above operation when no drawing exists at current frame. This
is due to accessing value from a nullopt optional. Not fixed with extra
condition before calling `value()`. Also moved this logic into a lambda to
avoid duplicate code.
Pull Request: https://projects.blender.org/blender/blender/pulls/134910
Mistake in a0e6e16da5
Testing of the previous commit was inadequate, it works when a brush
tool is selected and invoked via the operator search, as the
`active_vert` property will be set by the raycast. When the Mask by
Color tool is actually active, the brush cursor, and thus the function
that performs a racyast is inactive, preventing the operator from
running.
Pull Request: https://projects.blender.org/blender/blender/pulls/134960
This shared state between original data and depsgraphs was added in
98a0bcd425. Other physics systems also share
the pointcache, but not the simulation state to this extent, which leads
to this kind of crash.
The mutex lock is not a great solution, you don't really want both render and viewport to be filling the same cache in parallel. However
this kind of problem also exists in other physics systems, and solving
that is certainly beyond the scope of 4.4, and probably needs to wait
for a bigger physics rewrite. In general the recommendation is to bake
everything before rendering.
Pull Request: https://projects.blender.org/blender/blender/pulls/134779
The issue was that in the macro used, the wrong ID was passed into
the rename function. Since nodetrees are a separate embedded ID,
the rename function couldn't resolve the RNA path and thus failed.
This broke keys and drivers when renaming nodes.
Not only for the compositor, but all node trees.
This was caused by 1c7c1829b6
Pull Request: https://projects.blender.org/blender/blender/pulls/134936
This happend for meshes without UVs.
The call to `texpaint_request_active_uv` in
`DRW_mesh_batch_cache_get_surface_edges` is from 9c010c44f4.
Looks like this is not needed (assumed to be copy paste error).
Pull Request: https://projects.blender.org/blender/blender/pulls/134856
In the Vertex Weights panel of the 3D Viewport sidebar, allow for
normalization of vertex groups when any number of vertex groups are
locked.
As with other normalization operators that handle locked vertex groups,
locked weights are summed, and whatever space is left between that sum
and 1.0 is used to normalize the unlocked weights.
Pull Request: https://projects.blender.org/blender/blender/pulls/134535
Make `action_channelbag_ensure()` and `action_fcurve_ensure()` return a
reference. Earlier it returned pointer that should never be `nullptr`,
but in certain cases (which would indicate a bug, as the function's
preconditions aren't met) could technically still be a `nullptr`.
Instead, the preconditions are checked with `BLI_assert()`, and a
release build will simply assume that they are met. This simplifies the
code of the functions themselves, as well as the callers.
For reference: this was discussed in !134866.
No functional changes.
Pull Request: https://projects.blender.org/blender/blender/pulls/134918
Over time the `FCurveDescriptor` class has grown, and some concern was
raised (!134866) about the performance impact of passing it by value.
It's now passed by const reference instead.
No functional changes. I did find some non-`const` uses that are now
also `const`.
Pull Request: https://projects.blender.org/blender/blender/pulls/134913
This change brings the following improvements on the user level
- Support of GPUs with gfx12 architecture
- New HIP-RT library which in addition to the gfx12 support brings
various bug-fixes.
The known limitation of gfx12 is that OpenImageDenoiser does not yet
support this GPU architecture. This means that while Cycles will use the
full advantage of the gfx12 (including hardware accelerated ray-tracing),
denoising will only be possible on CPU, or secondary gfx11 or below GPU.
This is something that requires a change in OIDN and it is to late to do
it for Blender 4.4, but it is something to look forward for Blender 4.5.
The gfx12 changes for the pre-compiled kernels is rather trivial,
so it comes together (in the same PR) as the bigger HIP-RT change.
On the development side this change brings the following improvements:
- One step compile and link (much simpler CMake rules)
- Embedding BVH binaries in hiprt dll (which makes it easier to package
and load, without relying on special path configuration)
Co-authored-by: Sahar Kashi <sahar.kashi@amd.com>
Co-authored-by: Sergey Sharybin <sergey@blender.org>
Co-authored-by: Brecht Van Lommel <brecht@blender.org>
Pull Request: https://projects.blender.org/blender/blender/pulls/133129
This adds support for Overlay tests.
There are some differences with how we handle tests for other engines:
- The renders are captured using `bpy.ops.render.opengl()`, but this
won't work on our GPU build bots.
- A single blend file can run multiple tests by outputting a txt list
with the test names.
- Each overlay test blend file requires a matching script file with
the same name inside `tests/python/overlay/`.
- To reproduce a specific test state you can run
`blender "(...)/tests/data/overlay/<test>.blend" -P "(...)/tests/python/overlay/<test>.py" -- --test <test-number>`.
Note:
The current test permutations are WIP, so reference images are not
committed to the data repo for now.
Pull Request: https://projects.blender.org/blender/blender/pulls/133879
The `BKE_fcurve_bezt_shrink()` function had an artificial limitation to
only allow shrinking the `bezt` array. That limitation is now removed,
and therefore the function renamed to `BKE_fcurve_bezt_resize()`.
A note was added to the documentation that newly added array elements
should be initialized by the caller.
Pull Request: https://projects.blender.org/blender/blender/pulls/134864
Split the majority of `animrig::action_fcurve_ensure()` into a new
function `action_channelbag_ensure()`. This ensures that the Action has
a layer, keyframe strip, action slot, and channelbag for the given ID.
`animrig::action_fcurve_ensure()` now just calls into that function, and
then ensures that there is an F-Curve in that channelbag.
No functional changes.
Pull Request: https://projects.blender.org/blender/blender/pulls/134864
Hydra's viewport and final engines conflict on whether the rendering loop
is internal or external. The old approach works for Hydra Delegates that
converge with a single sample (hdStorm) but causes engines relying on sample
accumulation (hdEmrbee) to become stuck.
This minimal change retains compatibility with delegates like Storm while
replicating the behavior of the viewport engine in the final engine.
Pull Request: https://projects.blender.org/blender/blender/pulls/134804
The Grease Pencil extrusion operator didn't transfer the vertex group names
from the old `CurvesGeometry` to the new one, resulting in the `gather_attributes`
function not transferring the vertex group weights correctly. This PR adds
`BKE_defgroup_copy_list` to the operator to remedy this.
Pull Request: https://projects.blender.org/blender/blender/pulls/134695
The interpolation tool is using stroke length to map points between strokes with
uneven number of points. This fails when a source stroke has zero length (all points
in the same location).
Added a special case check to ensure uniform mapping by index is used in this case.
Segment mapping has been moved to a separate function for clarity.
Pull Request: https://projects.blender.org/blender/blender/pulls/134849
Fix an issue where the versioning of Action & slot assignments did not
use RNA properties to do the slot assignment. This caused certain
on-update callbacks to be missed, which in turn meant that an Action
constraint could remain disabled even though its action slot assignment
had been corrected.
This is now resolved by actually using RNA to set the assigned slot in
the versioning code.
Unfortunately that does mean that any reporting done will be by the
generic RNA code as well, and won't be specific to versioning. This
shouldn't be much of an issue in practice, as any warning was only shown
in the rare case of mis-matched `action.idroot` properties.
Pull Request: https://projects.blender.org/blender/blender/pulls/134759
This switches clang to be the default compiler on Windows ARM64, allowing for an override to MSVC.
Turns out MSVC builds have been broken for months, but nobody checked, so I'm just switching them off for now and setting clang as the default.
These updated scripts allow for the msbuild generator to use an external (ie, non-MSVC) clang installation properly, otherwise they failed.
They also allow for users to specify their own desired clang compiler via an environment variable.
An update to the docs will come seperately.
Pull Request: https://projects.blender.org/blender/blender/pulls/134566
EEVEE crashes when it is not able to allocate buffers. Previously we had a
message showing to the user that it tries to allocate a texture larger than
supported by the GPU. But was not implemented for EEVEE-next. This fix will
add back this error message.
Pull Request: https://projects.blender.org/blender/blender/pulls/134725
Use the new `action.fcurve_ensure_for_datablock()` function to ensure
the opacity F-Curve exists.
This function also ensures that the Action itself is ready for keying
(it has a layer and a keyframe strip), and ensures the slot for the
Scene exists and is assigned.
Co-authored-by: Pratik Borhade <pratikborhade302@gmail.com>
Pull Request: https://projects.blender.org/blender/blender/pulls/134753
The File Output node crashes when saving a 16-bit vector image in an
RGBA image. That's because the OIIO writer assumes 4-channel buffer
while the buffer provided by the node is only 3-channel. To fix this,
the OIIO writer is extended to support all possible combination of
source and target channels.
Pull Request: https://projects.blender.org/blender/blender/pulls/134789
When a text string is longer than the available space it can be at a
scrolled offset, so that you can edit text that is wider than its
container. When selected text is deleted we don't update this offset,
so it is possible to have newly pasted text scrolled out of view. This
PR decreases the offset by the amount of the selected string that is
currently out of view.
Pull Request: https://projects.blender.org/blender/blender/pulls/134815
These were reading in "widths" and not adjusting the values when setting
Blender's "radius" properties.
Found while cleaning up the radius API usage as part of another change.
Pull Request: https://projects.blender.org/blender/blender/pulls/134709
This commit moves the initialization of the node-sized accumulation
array out of the parallel loop to avoid odd optimization errors when
differing between debug and release builds as well as errors due to
integer overflow in both builds.
Additionally it only accumulates results from affected nodes.
Pull Request: https://projects.blender.org/blender/blender/pulls/134720
The `init_positions` variable that the `ss.sculpt_persistent_co` member
gets set to was not passed into the corresponding calculation function,
resulting in a no-op.
Pull Request: https://projects.blender.org/blender/blender/pulls/134781
With the brush assets project, brushes were moved from being local to
the working blendfile to being linked from asset libraries. This breaks
the Image Paint 'Clone' brush, as it has a brush property that links to
other Image datablocks.
To support this functionality, this commit adds the corresponding
properties into the `ImagePaintSettings` struct so that it is stored
locally with the images that will be used by the tool, inside the main
blendfile.
The source image property is shared with the 3D version of the 'Clone'
brush instead of adding a separate field to preserve old behavior.
Notably, this has the following limitations:
* If clone brush assets have been made and shared with external packs,
they would not work out of the box with linked image assets.
* Despite these settings being stored on the scene, they are populated
inside the tool window under "Brush Settings" which is potentially
misleading. However, this is already the case for the 3D version of
the brush, so further UI refinement will happen outside of this PR.
* Users will be unable to use separate images simultaneously for the
Image editor and the 3D viewport, unlike in pre-4.3 versions. This
can be adjusted in the future if it is a critical workflow.
Because the intended design and functionality of this tool is currently
questionable, this commit opts to make these changes instead of doing
further design to support both accessing data on the brush and on the
scene.
Pull Request: https://projects.blender.org/blender/blender/pulls/134474
The previous attempt to restore this behavior was in
0a2d5d5801
That commit introduced clearing the related `SculptSession` variables
when the paint BVH was freed to avoid bad behaviors seen in 4.2 and
prior. However, this solution is somewhat incorrect, as the prior state
of this data is not necessarily restored when the BVH is recreated,
unlike other temporary mapping data.
This results in the persistent base data being cleared occasionally with
no indication to the user that this is happening, causing the setting to
appear to be inactive.
To fix this, this commit makes a few changes:
* Removes clearing this data in the `BKE_sculptsession_free_pbvh`
function.
* Initializes the displacement arrays to the same size as the position
and normal array.
* Introduce new variables to track the saved multires grid size so
that it is not deleted, only considered invalid when the topology
changes.
Pull Request: https://projects.blender.org/blender/blender/pulls/134780
The playhead is redrawn as an overlay, so instead of a full region
redraw using `ARegionType.draw()`, at least an overlay only redraw using
`ARegion.draw_overlay()` needs to be triggered.
Any redrawing within a window is skipped if neither the screen, nor any
of its areas or regions are tagged for redraw. So since there are no
other areas or regions to be fully redrawn in this window, no redrawing
will happen. The screen needs to be tagged for redraw, which will skip
most drawing in this case, and just draw the overlays as wanted.
Pull Request: https://projects.blender.org/blender/blender/pulls/134579
When changing editors within an area we are currently always setting
the subtype if there are modes. This causes problems when returning to
a previous editor versus starting with the correct one. We have tried
various ways of returning to the old editor, but never quite works.
This PR only sets the subtype if we are creating a new area. If not
just leave it as it was so we naturally return to its old state.
Pull Request: https://projects.blender.org/blender/blender/pulls/134642
The new CPU compositor in v4.4 is much slower than the old CPU
compositor in v4.3 on Windows. This is because MSVC does not inline many
of the core methods in the Result class of the compositor. To fix this,
we force inline those methods, adding a new macro for inlining methods
in the process, since the existing macro has the static keyword, which
only works for functions, not methods.
Pull Request: https://projects.blender.org/blender/blender/pulls/134748
