Turns out we were not building OSL with OptiX enabled anymore.
Also check now if the OSL builds has OptiX support and if not
disable it in Cycles.
Building OSL with support for this (still) does not require
either the OptiX SDK or CUDA, it only needs LLVM.
Pull Request: https://projects.blender.org/blender/blender/pulls/118234
Images do not update in the compositor and other places when their file
path changes or their generated options change. This is because the
compositor relies on depsgraph updates, which weren't tagged in this
case. Fix this by tagging appropriately in RNA.
Pull Request: https://projects.blender.org/blender/blender/pulls/118187
`BKE_library_id_can_use_filter_id` was missing the ShapeKey type for
geometry IDs (meshes, legacy curves, and lattices), leading to the
remapping code to fail to do its job when deleting the ShapeKey only.
The new keyframing functions introduced in #113504
didn't call the functions to decompose the NLA stack.
In practice this meant that when inserting keys into strip
that is under an additive strip, it would take the result of the additive Strip and
bake it back into the base. This would double the transform.
The fix is to call `BKE_animsys_nla_remap_keyframe_values`.
Unfortunately to do so, I had to pass through a few more
arguments to the keyframing functions.
Also adds unit tests to cover the caused bug.
Pull Request: https://projects.blender.org/blender/blender/pulls/118053
Convex hull calculation would "cap" line segments with the first point.
Besides causing a buffer overflow when passing in two points,
it's not needed as matching first/last ends aren't expected
to be set by this function.
Fixed duplicate nodes when converting texture scale/bias and
channel names on material import. This required extending
the node caching to handle cases where a USD shader is converted
to multiple Blender nodes.
Pull Request: https://projects.blender.org/blender/blender/pulls/118002
The GPU compositor Displace node produces NaN pixels if connected to the
Image node. That's because the Displace node access the MIP levels of
the texture produces by the Image node, but those levels were never
initialized, so fix this by completing the levels.
The result of cross_poly_v2 was flipped compared with cross_tri_v2 &
cross_poly_v3 (with the Z values zeroed).
Ensure cross_poly_v2/3, cross_tri_v2/3 return compatible results and
updating the doc-strings noting that a negative Z is for clock-wise
polygons.
When calculating the width of strings for the tooltip window size, enum
labels were measured using the string from the text_suffix field but
length from the text field. Introduced with std::string changes.
Pull Request: https://projects.blender.org/blender/blender/pulls/118064
An allocated char pointer was reassigned without being freed.
Fix by replacing the error-prone combination of C strings and
std::string with a complete conversion to std::string. Also change
the logic a bit to allow moving the path string in some cases
instead of copying it.
Pull Request: https://projects.blender.org/blender/blender/pulls/118046
After bb080140c0, `SCULPT_pbvh_calc_area_normal` returned
std::nullopt when it didn't sample any vertices, instead of returning
an ignorable boolean return value. There was a false negative failure
though, because it didn't keep track of when it sampled BMesh "orco"
vertices.
When reordering vertex groups on a mesh, some cloth properties would
get affected by another group, because they were based on group index
instead of name.
This issue was already handled by
`BKE_object_defgroup_remap_update_users()` for some properties, but
not all.
This commit extends the function for missing properties: Shrinking,
Shear, Internal Springs, Pressure, Self Colllisions and Object
Collisions.
Pull Request: https://projects.blender.org/blender/blender/pulls/114032
Baked armature properties are now placed in a single group "Armature
Custom Properties", instead of creating a new group "Group.nnn" for each
custom property.
This bug was not reported.
Pull Request: https://projects.blender.org/blender/blender/pulls/117993
When baking custom properties, avoid keeping references to custom
property values that are known to be impossible to animate anyway.
The crash was caused by custom properties containing collections of ID
properties. Keeping Python references around for too long and then
accessing them caused Blender to crash.
My solution is to only keep track of custom property values that might
be keyable. For some this is certain: floats, ints, bools are keyable,
whereas lists, dicts, etc. are not. However, strings can be the RNA
value for an enum property, which is keyed via its integer
representation. So, the new function `can_be_keyed()` can return `True`,
`False`, or `None`. By skipping those values where it returns `False`
the crash is already resolved, making it good enough for now.
Pull Request: https://projects.blender.org/blender/blender/pulls/117993
The Inpaint node crashes for large images. That's because the output
buffer was only allocated for a single chunk. So allocate the entire
output buffer in the tile initialization step.
Caused by #113504
While basic cycle aware keying works, there is a special case
when inserting the first key. In that case, after the key has been inserted,
it is duplicated and moved around so the FCurve range from
first to last key is exactly the range of the action.
It also auto-creates the Cycle modifier on the FCurve .
Fix the issue by calling the function that does the key duplication
and cover with unit tests.
Pull Request: https://projects.blender.org/blender/blender/pulls/116943
This means the array can be shared between geometries when unchanged,
reducing memory usage and increasing performance. It also means that
handling the data can be done more generically using the attribute
system. In the future the transforms can become attributes too.
Two other changes are related here:
- The "almost unique ids" are cached with a shared cache so they
are shared too.
- The reference indices are baked as an attribute now, making the process
more generic and potentially easier to optimize in the future.
Pull Request: https://projects.blender.org/blender/blender/pulls/117951
Currently, the geometry nodes modifier always creates a time dependency
if there is a bake animation node, even if there is no bake. Changing this is
not entirely trivial, because it is only known during evaluation if there is baked
data. Making `Still` the default works around that issue in many cases where
the Bake node is in a node group asset that does not depend on time.
Furthermore, it's nice to have the "cheaper" version of the node by default,
which is baking a single frame, instead of baking everything.
Pull Request: https://projects.blender.org/blender/blender/pulls/117999
This PR applies the UI range for the Dyntopo `detail_size` and `detail_percent` to the actual property range. Previously, 0 and negative values were allowed if manually entered into the field, the former leading to blender hanging upon the next stroke, and the latter leading to potentially odd or unwanted behavior.
Pull Request: https://projects.blender.org/blender/blender/pulls/117976
Campbell Barton