In some cases, sculpt code currently creates undo steps that are stored
in `step_init` in the undo stack, but then skips actually pushing them.
That can happen when an operator is cancelled (like the transform
operator in #123172) or because pushes in "nested operator calls"
are currently explicitly disabled. That happens when calling the brush
operator from a script.
It turns out the undo code never freed the `step_init`, probably because
it assumed it would be pushed to be part of the stack afterwards.
Personally I'm not convinced that separating undo step creation into
two stages with `step_encode_init` and `step_encode` is a great design
or a necessary one, but I'm trying not to get into that deeper right now.
Fixes#123172
Pull Request: https://projects.blender.org/blender/blender/pulls/123331
Calling `MEM_freeN` on data allocated with `MEM_new` is bad, since it
will not call a destructor matching the one invoked as part of
`MEM_new`.
While in practice cases fixed below were 'not a problem' currently, as
they are trivial Cpp types (and therefore their destructor is doing
nothing), `MEM_freeN` has no way to ensure it is dealing with such a
trivial data type, so allowing such mismatch is dirty and dangerous.
Note that almost all fixed cases look more like unintentional mistakes
(mis-usages of `MEM_new` instead of `MEM_cnew`).
NOTE: There is one more (known!) case in the asset code, which fix is
slightly less trivial, and will go through a separate PR.
NOTE: This is a by-product of some work to detect such invalid usages of
`MEM_freeN` on memory chunks allocated with `MEM_new`.
Pull Request: https://projects.blender.org/blender/blender/pulls/123691
If a non-instanced collection is linked, any collection exporters on the
linked collection would be active and invokable. This is probably not
desired as it could inadvertently overwrite files from the original.
Disable the operators in this case.
See PR for how each of the various append/link scenarios behave.
Pull Request: https://projects.blender.org/blender/blender/pulls/123149
This is an alternative fix to #123524.
This is necessary, because `sculpt_update_object` is run after
the mesh is evaluated, but before the geometry depsgraph operation
is done. Only after this depsgraph node is done, `DEG_object_geometry_is_evaluated`
will return true.
This approach of `unchecked` methods has been preferred for now
over moving the call to `BKE_sculpt_update_object_after_eval`
to a separate depsgraph node or after depsgraph evaluation.
Blender would crash if it tried to convert a grease pencil
data-block with strokes that used the editcurve.
This was because of a wrong assertion that assumed
to work with a poly curve.
The fix moves the assertion to the correct place where
we handle the poly curves.
While the evaluated result is not well defined, we expect Blender to not crash
when there are dependency cycles.
The evaluation of one object often takes the evaluated geometry of another
object into account. This works fine if the other object is already fully
evaluated. However, if there is a dependency cycle, the other object may not be
evaluated already. Currently, we have no way to check for this and were mostly
just relying on luck that the other objects geometry is in some valid state
(even if it's not the fully evaluated geometry).
This patch adds the ability to explicitly check if an objects geometry is fully
evaluated already, so that it can be accessed by other objects. If there are not
dependency cycles, this should always be true. If not, it may be false
sometimes, and in this case the other objects geometry should be ignored. The
same also applies to the object transforms and the geometry of a collection.
For that, new functions are added in `DEG_depsgraph_query.hh`. Those should be
used whenever accessing another objects or collections object during depsgraph
evaluation. More similar functions may be added in the future.
```
bool DEG_object_geometry_is_evaluated(const Object &object);
bool DEG_object_transform_is_evaluated(const Object &object);
bool DEG_collection_geometry_is_evaluated(const Collection &collection);
```
To determine if the these components are fully evaluated, a reference to the
corresponding depsgraph is needed. A possible solution to that is to pass the
depsgraph through the call stack to these functions. While possible, there are a
couple of annoyances. For one, the parameter would need to be added in many new
places. I don't have an exact number, but it's like 50 or so. Another
complication is that under some circumstances, multiple depsgraphs may have to
be passed around, for example when evaluating node tools (also see
`GeoNodesOperatorDepsgraphs`).
To simplify the patch and other code in the future, a different route is taken
where the depsgraph pointer is added to `ID_Runtime`, making it readily
accessible similar to the `ID.orig_id`. The depsgraph pointer is set in the same
place where the `orig_id` is set.
As a nice side benefit, this also improves the situation in simple cases like
having two cubes with a boolean modifier and they union each other.
Pull Request: https://projects.blender.org/blender/blender/pulls/123444
Store library filepath so we can find the right datablock.
Tested to work with the brush assets project branch, as this
currently does not affect anything in main.
Pull Request: https://projects.blender.org/blender/blender/pulls/123449
I found that `packedmap` is effectively unused when I worked on #123243.
The only function that wrote to it was `blo_make_packed_pointer_map`
and that is never called. Packed data already used the normal `datamap` as fallback.
Pull Request: https://projects.blender.org/blender/blender/pulls/123244
Avoid logging actions about `action->idroot` not matching, when evaluating
the NLA and visiting a layered Action. Layered Actions are not limited to
a single data-block type, and so the code should ignore `action->idroot`.
Add layered Action support to `BKE_action_frame_range_calc()`, by
looping over all F-Curves of all Bindings in the Action.
Introduce `animrig::fcurves_all(action)`, which returns a vector of all
F-Curves in the Action, both for legacy and layered Actions.
No functional changes for legacy Actions.
For base mesh sculpting, we already draw the original data instead of the
evaluated data, so the evaluated mesh doesn't have to be updated to
contain the new attribute. This is different from multires sculpting
because sculpt brushes and drawing only deal with the evaluated
SubdivCCG, not the original multires modifier data. Removing the
unnecessary update removes a noticeable pause when clearing and
adding a mask.
This adds additional logic to `GreasePencil::rename_node` to rename
the strings in the modifier influence data.
This is similar to how `ED_armature_bone_rename` handles renaming of
strings.
Resolves#123321.
Pull Request: https://projects.blender.org/blender/blender/pulls/123365
This fixes#121695. `float4x4` matrices are generally expected to be 16 byte aligned.
Currently, there is no mechanism (afaik) that allows allocating these overaligned types
when loading files from disk. This patch adds an array with alignment information for
each type in `SDNA`. Currently, the alignment is just `__STDCPP_DEFAULT_NEW_ALIGNMENT__`
for all types and is manually set for the `mat4x4f` DNA type. The .blend file format is
not changed at all. The alignment information is purely runtime data.
In the future it would probably be good to generalize this a bit more instead of
hardcoding the alignment for `mat4x4f`, but would make it unnecessarily complex for
now because this is intended for the release branch.
Pull Request: https://projects.blender.org/blender/blender/pulls/123271
This adds a "Legacy Behavior" option to the Limit Rotation constraint that makes
it behave how Limit Rotation constraints did prior to
ed2408400d. Newly created constraints have this
option disabled, but versioning code enables the option on constraints from
older files to ensure that the behavior of e.g. existing rigs is not altered.
This is one part of a two-part fix for #123105. The other part is in PR
extensions/rigify#4.
Pull Request: https://projects.blender.org/blender/blender/pulls/123361
The current temperature unit adjusts to Celsius or Fahrenheit based on
unit system, but specifically for color temperatures the convention is
to display them in Kelvin, and it'd be strange to e.g. see 11240°F when
opening the white balance panel.
Therefore, this adds a dedicated Color Temperature unit, and uses it
for the two existing blackbody temperature inputs in shader nodes.
Pull Request: https://projects.blender.org/blender/blender/pulls/123337
When copying a local ID into a library, or a linked ID into a different
library, the 'linked' tags would not be properly preserved or set, and
the relative file paths would not be properly remapped.
These issues were not currently exposed (no existing code could trigger
them), but some of these fixes are needed for upcoming refactor of the
partial write code.
NOTE: not very happy with the split in library handling between
`BKE_id_copy_in_lib` and `BKE_libblock_copy_in_lib`, this will have to
be solved at some point.
Pull Request: https://projects.blender.org/blender/blender/pulls/123247
Currently the corner_vert and corner_edge attributes go through the
generic custom data interpolation, only to be overwritten directly
afterwards. In a profile of the execution of the subdivision surface
modifier, 5% of the time was spent in `layerInterp_propInt`. To fix
this, first build a `CustomData` layout to be used just for face corner
attribute interpolation. Then allocate the topology attributes
separately and add them to the result mesh later. This requires a fair
amount of boilerplate right now, but it should be improved with various
changes to mesh attribute storage in the future.
In a file with many object with a subsurf modifier, this change
improved the overall execution time by between 5 and 10% in my tests.
Pull Request: https://projects.blender.org/blender/blender/pulls/123254
The issue was that we used `from_orthonormal_axes` which obviously expects that
the axes are orthonormal. However, with custom curve normals this is not always the
case. Math wise, an additional normalization is necessary because the cross product
is not automatically normalized anymore.
This change also means that `point_matrix` may have a shearing component. But I
think that's fine here because the matrix is only immediately applied on vertex positions.
This shouldn't affect any case where the normal and tangent are orthonormal.
Pull Request: https://projects.blender.org/blender/blender/pulls/123238
The issue was a floating point precision issue
between the number of the cycle and the time
within the cycle (`cycle` and `cyct`).
Due to that the `cycle` would advance to the
next whole number while the `cycle time`
is still slightly under that.
This is fixed by calculating the `cycle` with double precision.
This has a measurable performance
impact (for `fcm_cycles_time`)
| Before | After |
| - | - |
| 39ns | 44ns |
For fcurves with a cycle modifier,
this code is run at every evaluate call.
The only time when that would be an issue is
in the graph editor, where there is an evaluate
call for roughly every pixel for curves that have a modifier.
However even in that scenario the performance
is the same within run to run variance (for `graph_draw_curves`)
| Before | After |
| - | - |
| 91565ns | 91430ns |
Pull Request: https://projects.blender.org/blender/blender/pulls/123222
