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
This implements (most of) the proposal in #122743:
* Add a new `IDP_FLAG_STATIC_TYPE` IDProperty flag.
* Update `BPy_IDProperty_Map_ValidateAndCreate` and related to never
change an existing property type if statically typed.
The biggest change happens in bpy assignement code, since instead of
replacing the old exisitng property by a newly created one, and copying
over a few settings, now the old property is kept if possible, and a new
one is only created if needed.
And in case the existing property is statically typed, if it cannot be
re-used to store the given value, and error is reported and it remains
unchanged.
`IDP_ARRAY` is also supported for basic numeric types, so 'vector'
properties and such work as expected. Lentgh is considered as part of
the static type (i.e. one can only assign a 3 components py sequence to
a 3-len array property, etc.).
Such in-place update is not yet implemented for `IDP_IDPARRAY` and
`IDP_GROUP` types. While important (especially the group one), they are
not that critical for the current issues related to changing IDProperty
types.
Masks are not updated when edited when using the GPU compositor. That's
because the GPU compositor caches static resources and invalidates them
according to the recalculate flags that the depsgraph flushes to IDs.
The issue is that the flags are not flushed to the evaluated IDs of the
compositor depsgraph, but rather to some other evaluated versions of the
IDs.
To fix this, we make the compositor depsgraph persistent and store it in
the scene runtime. This allows us to reliably track changes to resources
used by the compositor and also reduces the overhead of depsgraph
creation in the compositor job.
Patch originally provided by Sergey.
Fixes#121188.
Pull Request: https://projects.blender.org/blender/blender/pulls/123085
This patch removes the Auto Render option from the compositor. This is
done for the following reason:
- The option didn't really work except in the case of transforming an
object. So it wasn't really reliable.
- It made little sense to use since the introduction of the Viewport
Compositor.
- It had a number of UX issues, including the fact that it can't be used
with animation playback, and the fact that rendering can get in the
way of the UI depending on the preferences for temporary editors.
Pull Request: https://projects.blender.org/blender/blender/pulls/123132
The documentation of some of `BKE_animsys_nla_remap_keyframe_values()`'s output
parameters were unclear/misleading, in particular making it unclear how those
output parameters should be used/interpreted by calling code.
This commit attempts to improve that, in particular trying to make it clearer
what actions calling code can/should take based on the returned output
parameters.
Pull Request: https://projects.blender.org/blender/blender/pulls/123081
This type was not necessary because `ptr_` is allowed to
be `nullptr`. This also means that we don't need to handle
the `None` case in switch statements. We assume that at
this point the owner is valid.
Previously these were shown as missing add-ons, since they have been
intentionally removed, remove them as part of versioning instead
of showing them as "Missing Add-ons". This is especially important
for X3D & STL which were enabled by default which meant any user
loading 4.1 preferences would have them shown as missing add-ons.
Add the color for the new keytype 'generated' to the VSE theme settings
as well.
This includes the versioning code that loads the default from the scene.
Without this, the preference would be initialised to black.
Note that the 'Blender Light' theme does not have any colors for any of
the key types, so I also didn't add the 'generated' type there.
Pull Request: https://projects.blender.org/blender/blender/pulls/123005
`GVolumeGrid` and `VolumeGrid<T>` are basically pointer wrappers.
Assigning a nullptr is possible and allowed by code, but node socket
code doesn't currently check if a grid pointer is null before accessing
it.
Disallow null grid pointers in socket values. Asserts are added for this
purpose. The Points-to-SDF node checks the grid value before writing to
output.
Pull Request: https://projects.blender.org/blender/blender/pulls/122996
Fixes#116567
The issue was with a flag that skips a particular post-evaluation step
specially for the Child Of constraint. The flag wasn't getting properly
unset when the constraint target was removed, and therefore that
post-evaluation step was getting erroneously skipped in some cases.
This commit fixes the issue by always setting the flag appropriately in
the Child Of evaluation function itself. This is admittedly rather
hacky, but no more hacky than the existence of the flag in the first
place.
Pull Request: https://projects.blender.org/blender/blender/pulls/122881
This changes fixes a wrong mesh as a result of apply base mesh from the Multires modifier
In the previous code some corner vertices where incorrectly assigned to null and skipped form processing as a safety guard
that was only relevant when the mesh had hidden faces in edit mode. The hidden corner vertices bug is handled now when checking
> if (corner_x_index < 0 || corner_y_index < 0) {
> continue;
> }
Since the patch that caused this error was a fix to a previous infinite loop bug,
I also tested and made sure this updated patch still fixes the original issue it was trying to fix.
Note: This PR replaces #122845 due to a large amount of git rebase issues.
Pull Request: https://projects.blender.org/blender/blender/pulls/122859
The attribute API defined in `attribute.cc` was dependent on
the assumption that `ID`s are always the "direct" owners of attributes.
For Grease Pencil drawings, this is not the case. The Grease Pencil ID
stores the attributes for layers, and the attributes for drawings are stored
in `CurvesGeometry` on the drawings themselves.
In order to make use of `rna_attribute.cc`, we need that API to handle
other types of attribute owners.
This adds an `AttributeOwner` which is basically just a type and a
pointer. We replace the `ID` pointers and pass `AttributeOwner`s instead.
For cases where we have to do a switch based on the type, all the
types are handled and the `default` statment is left out. This ensures
that we get a compiler warning when a new `AttributeOwnerType`
is added.
No functional changes expected.
Pull Request: https://projects.blender.org/blender/blender/pulls/122765
Mainly correct/update some comments (e.g. missing reference to Boolean
type), and add some notes essentially about issues with current IDProp
String code (see also #86960 ).
No functional change.
The C++ classes that wrap DNA structs should never store any data.
Runtime data needs to be stored on the runtime classes/structs
and anything else needs to be put into DNA.
This adds static asserts to make sure (at compile time) that these
classes have the same size as the DNA structs (i.e. they don't store
additional data).
This fixes a bug that happens when two drawings in the same grease pencil
data-block reference the same array. In this case, the same attribute array
pointer is written to .blend files for both drawings. However, when loading the
.blend file, it was not detected that the data is shared. Instead each drawing
would think that it was the single owner of the array even though it was shared.
The same problem can technically occur for all kinds of shared data, and not
just for grease pencil. However, only grease pencil shows this issue currently,
because it can easily happen that the same attribute is shared between different
`CustomData` within the same data-block. This very rarely happens with meshes or
curves, because different `CustomData` usually have different sizes and don't
share the same arrays.
A previous fix (cc891aa699) solved pretty much the same issues for the case
when two layers in the same `CustomData` share data. This original fix is now
generalized and is integrated with `BLO_read_shared` which makes it less error
prone. For each data-block, the `BlendDataReader` now remembers which shared
data it has loaded before so that it is not loaded again.
In the future this could be extended to support data that is shared between
data-blocks, but that's not as straight forward as one would hope currently. For
various reasons, different data-blocks could store independent data at the same
pointer. I specifically noticed that with a regression test file:
`blender/tests/data/modeling/geometry_nodes/attributes/attribute_statistics.blend`.
To support it, one solution could be to tag `BHead` in .blend files to indicate
that a specific pointer is shared and unique within the entire file. But that's
for another day.
Pull Request: https://projects.blender.org/blender/blender/pulls/122780
The buttons to upgrade or sync extensions called extension
operators via bpy.app.handlers, requiring awkward glue-code
which didn't show error reports in to the user.
Remove these operators and call the upgrade & sync operators directly.
