In the case of weight, tilt or radius (these dont don't change [handle]
positions), dont change the handle types.
See the doc for `BKE_nurb_handles_test` ("Use when something has changed
handle positions")
NOTE: a siilar check is done in `createTransCurveVerts`
Pull Request: https://projects.blender.org/blender/blender/pulls/129056
The root cause is that the editor functions
(`retrieve_editable_and_selected_points` etc.) use the evaluated object,
while the operator poll which checks for the active layer uses original
data. The geonodes modifier removes all GP layers. The operator runs
anyway because the original data has an active layer, but then crashes
layer when trying to access the layer in evaluated data.
The fix here is to use original data throughout, which is usually the
data we want to look at for edit/sculpt/paint tools. There is a caveat
that the internal functions like `calculate_view_positions` still should
be using evaluated data for some things, like finding intersections with
other visible strokes. This isn't distinguished clearly atm and might
cause further bugs down the line.
Pull Request: https://projects.blender.org/blender/blender/pulls/129109
In Pose copy/paste case, (almost) all dependencies of the armature
object are cleared during the copy phase (only its obdata Armature is
kept). This would also clear ID pointers from the liboverride data if
the source object is a liboverride.
While manual handling of such cases is possible, it's simpler here to
simply make the data in the pastebuffer fully local, since only the Pose
part of it is used anyway.
Runtime UID should be generated after other basic ID info is set, in
this case the `tag` flags were still the one read from file, which could
lead to an assert in `BKE_lib_libblock_session_uid_ensure` in case the
blendfile was generated from a temp main data-base.
This makes it so the stroke selection is taken into account by the
interpolation tool.
This does NOT use the selection _order_, just limits interpolation by
index to the selection. The result will be the subset of selected
strokes, excluding non-selected strokes, which is closer to the GPv2
behavior.
The `sample_curve_attribute` function was using the target curve index
for both the src and dst curves. This worked when all the curves are
interpolated, but with selections the dst curve is generally not the
same index as the src curve. The `from_curve_indices` array must be
passed along as well to retrieve the correct source index.
Pull Request: https://projects.blender.org/blender/blender/pulls/129096
Fix the failing rendering of volumes on Windows with HIP SDK 6.1
by reducing the optimization level.
There should be no functional or performance difference for the average
user as the Blender foundation currently does not use HIP SDK 6.1
on Windows. This change is primarily to fix issues for community members
building Blender locally.
Pull Request: https://projects.blender.org/blender/blender/pulls/128836
Line art uses `DEG_OBJECT_ITER_BEGIN` to load all objects that has a
geometry in the scene, which is kind of a hack since the beginning. This
left potential problem where the iterator could go through some objects
that line art didn't have a dependency on (e.g. other grease pencil
objects):
- Since those "other" objects often evaluates fast enough, so
line art always end up having valid data from everywhere after lengthy
mesh evaluation prior to itself, so this problem was not prominent.
- However, in rare cases, there are other objects that takes a lot of
time to evaluate, this causes line art to potentially iterate to objects
that are still invalid.
This fix will build a `Vector<Object *>` during `update_depsgraph`, and
use such list inside `DEGObjectIterSettings` to filter out objects that
the modifier isn't dependent on, thus remove the possibility of reading
objects that hasn't been evaluated.
Since Line art will not add grease pencil objects as potential geometry
inputs, all mesh/curve types of geometries generated by geometry nodes
modifier directly inside other grease pencil objects won't be loaded.
This can be mitigated by having a third mesh object that reads the
result from the grease pencil object that generates geometries, then
directly output them for line art to read.
This also fixes#128888.
Pull Request: https://projects.blender.org/blender/blender/pulls/128890
Blender would crash in certain cases when using the `Set Curve Resolution`
operator to increase the curve resolution.
This was because the `evaluated_offsets_cache` was not tagged
to be updated. The fix makes sure to use `tag_topology_changed`
to tag the offsets cache as dirty.
Pull Request: https://projects.blender.org/blender/blender/pulls/129098
Instead of assigning `any_point_selected` to selection span elements,
pass `select` value if any point of that curve exists in vgroup.
Also renamed `any_point_selected`->`any_point_in_group`.
Pull Request: https://projects.blender.org/blender/blender/pulls/129035
The problem was that the screen space size of the brush was not updated.
This was not fully noticeable because the radius control operator was using
a not working scale correction.
The solution remove the scale correction and just update the brush size.
Note: This update currently happens when the cursor is drawn.
Pull Request: https://projects.blender.org/blender/blender/pulls/126773
The updated autopep8 was ignoring the "exclude" values because
they started with a "./" instead of a "/".
The documentation I found doesn't go into details, even if this is a
regression in autopep8, update the configuration as it causes
"make format" to format files that aren't meant to be modified.
When baking an Action on an object which has custom properties that come
from an addon or python script that isn't loaded, the baking would fail with:
```
TypeError: Cannot assign a 'dict' value to the existing 'hops' Group IDProperty
```
This comes from an addon that isn't present or loaded on a users machine.
The fix is to check for `IDProperty` and skip those.
Pull Request: https://projects.blender.org/blender/blender/pulls/129057
The issue was that the shader `gpu_shader_gpencil_stroke_vert_no_geom.glsl`
assumed a wrong format of the color attribute (`uchar4` instead of `float4`).
The fix uses `vertex_fetch_attribute` with `float4`.
Pull Request: https://projects.blender.org/blender/blender/pulls/129072
GPv3 stores the vertex group inside `CurvesGeometry`.
When the object level vertex group name is changed, we have to loop over
the vertex group list of each drawing. If a matching name is
found, copy the new name to the vertex group in `CurvesGeometry`. A separate
function is created to handle this: `BKE_grease_pencil_vgroup_name_update`.
Pull Request: https://projects.blender.org/blender/blender/pulls/129038
When symmetrizing a bone using "Armature->Symmetrize" in Edit Mode,
Settings on the pose bone were not symmetrized. This patch symmetrizes
ONLY the setting mentioned in the bug report. (limit rotation Z)
IMO other settings should be reviewed on a 1 by 1 basis.
Do note that this assumes that the armature is symmetrized across the YZ-plane
Pull Request: https://projects.blender.org/blender/blender/pulls/129004
The issue was that the `merge_layers` function assumed that the
`src_grease_pencil` always returns a drawing when calling
`get_eval_drawing` which can return `nullptr`.
The fix makes sure to check that the `src_drawing` exists.
Pull Request: https://projects.blender.org/blender/blender/pulls/129074
The problem was that when removing FCurve modifiers via the python API
the Depsgraph wasn't notified.
Fixed by tagging the owner ID of the fcurve modifier for an update.
Pull Request: https://projects.blender.org/blender/blender/pulls/128984
When duplicating a bone (has to be done in edit mode)
the pose bone colors were not copied.
This adds the code to do just that
This also fixes it for symmetrising because that uses the
same code path
Pull Request: https://projects.blender.org/blender/blender/pulls/129007
Regression in [0] although prior to this the `deformed_only` was not
being set, caused by a separate regression.
[0]: 80fa49f24ae768c5869065363e116b51175fb2df
Regression in [0] caused the outliner not to draw at the clamped scroll
position. This meant searching in the outliner could show no results
even when results were found. It also meant deleting objects could show
nothing until the user manually forced the outliner to redraw.
Workaround the problem by detecting when outliner drawing changes the
scroll position and draw again. While inelegant this doesn't happen
often in practice.
Fixing without a second draw likely involves resolving the order of
initialization problem by separating bounds calculating from drawing.
[0]: a63ac425d0
The `u_scale` should be given a default value of `1.0f` so that newly drawn
strokes with the drawing already having `u_scale` attributes will have
correct texture mapping size instead of a default of `0.0f`.
Pull Request: https://projects.blender.org/blender/blender/pulls/129060
The build modifier crashes when getting lengths of cyclic curves because `evaluated_points_by_curve`
was previously always called with `cyclic` parameter set to `false`.
The `cyclic` parameter necessary for `curves::segments_num` to return correct number of segments.
Pull Request: https://projects.blender.org/blender/blender/pulls/128955
Caused by 364d62e59b.
In draw mode, we want to make sure to show vertex colors by default
because users can draw with them. Otherwise this can lead to unexpected
behavior in solid shading mode when vertex colors used to not be
rendered in draw mode.
The previous fix tried to do this by enforcing the `V3D_SHADING_VERTEX_COLOR`
mode in draw mode. But this has the side effect that the material
stroke and fill color are overriden with pure white.
To fix the issue correctly, simply ensure that `vert_col_opacity` in
`grease_pencil_layer_cache_add` is set to 1.0f. This means that
vertex colors and material colors are shown just like in the render.
Also resolves#128680.
Pull Request: https://projects.blender.org/blender/blender/pulls/129054
Intel Windows drivers for 10th gen and lower has some strange behavior
when using dynamic rendering. It requires pipeline conditions to be met,
when beginning a new rendering scope. This is strange as specs + VVL
notes that these conditions should be met during vkCmdDraw* commands.
Pull Request: https://projects.blender.org/blender/blender/pulls/129055
This change makes unused attachments extension optional.
This extension is fairly new and not all drivers have support for it.
The workaround will create additional pipelines when attachments are
not set.
Pull Request: https://projects.blender.org/blender/blender/pulls/129046
When writing to a property that doesn't exist e.g. `frame.drawing.strokes.test = 42`
no exception would be raised and it would silently fail.
The fix defines the `__slots__` on the classes explicitly which then raises an exception
if the user tries to write something that wasn't previously defined.
Pull Request: https://projects.blender.org/blender/blender/pulls/129047
When e.g. executing `drawing.strokes[0].softness = 3`, the API would
always create a new attribute `softness` even if that attribute existed
already.
The issue was that the code was using the `.get(value, fallback)` syntax
but the `fallback` expression is always evaluated by python.
The fix removes the use of the `fallback` and uses a simple `if/else` to
check if the attribute doesn't exist yet and only then create it.
Pull Request: https://projects.blender.org/blender/blender/pulls/129044
It looks like this specific case never really worked. This wasn't found before,
because in the large majority of cases, execution uses a more optimized code
path instead of this general one.
Pull Request: https://projects.blender.org/blender/blender/pulls/128993
Previously, in case of a failure during BVH transfer, when running out
of memory for example, we could get an error such as "BVH failed to
migrate to the GPU due to Embree library error (no error)", because
embree error status was actually reset before being queried.
This commit fixes its propagation.
Pull Request: https://projects.blender.org/blender/blender/pulls/129022
During development of Collection Export, it was noticed that our various
IO formats deal with errors in vastly different ways [1]. The crashes
were all fixed but now the motivating scenario is as follows:
If you setup a Collection Exporter for OBJ, STL, or PLY, and if it
runs into an error during processing, the C++ operator will return
"FINISHED" to the caller but the operator will also RPT_ERROR. This
causes the caller, Collection Export, to indicate "success" to the user
but the RPT_ERROR then causes a UI report that indicates failure.
This PR chooses to use the presence of the RPT_ERROR as indication of
"CANCELLED" operator status and should be safe enough to also apply
to 4.2 LTS.
This aligns with what Python does [2] and with what the new GSoC import
nodes have chosen to do [3]. Though doing this on import isn't quite
correct so this PR does not add that in to the base import operators.
Alembic and USD are better behaved in this scenario already. However,
they have their own quirks to follow-up afterwards.
[1] https://projects.blender.org/blender/blender/issues/117881
[2] https://projects.blender.org/blender/blender/src/branch/main/source/blender/python/intern/bpy_capi_utils.cc#L25
[3] https://projects.blender.org/blender/blender/src/branch/main/source/blender/nodes/geometry/nodes/node_geo_import_obj.cc#L49
Pull Request: https://projects.blender.org/blender/blender/pulls/127525
