Implement a basic undo system making use of implicit
sharing to avoid copies for stored attribute arrays.
Functionally this has no benefit over memfile undo,
but current undo system architecture requires each
edit mode to implement an undo type.
(Reviewed and co-authored by Hans Goudey)
Pull Request: https://projects.blender.org/blender/blender/pulls/134623
When needing a friendly title and icon for windows and screen areas we
use SpaceType callbacks. For Node editors this could crash for custom
node trees that are unregistered. node_tree_type_find returns nullptr
since this custom tree type is not in the map. In this case just return
translated "Node Editor".
Pull Request: https://projects.blender.org/blender/blender/pulls/134692
Previously we generally expected CustomData layers to have implicit
sharing info, but we didn't require it. This PR clarifies that we do
require layers with non-null data to have implicit sharing info. This
generally makes code simpler because we don't have to have a separate
code path for non-shared layers. For example, it makes the "totelem"
arguments for layer freeing functions unnecessary, since shared data
knows how to free itself. Those arguments are removed in this PR.
Pull Request: https://projects.blender.org/blender/blender/pulls/134578
This PR changes the default user preferences so that the Translation
Options are enabled. Otherwise with fresh install changing language
only translates tooltips.
Pull Request: https://projects.blender.org/blender/blender/pulls/134649
194e233d86 caused a discussion in the chat about the initialization
behavior of `MEM_new()`, and agreement was to not rely on
zero-initialization ever. Noted this in the API comment now.
Some people found the existing comment useful but it still left some
questions. Tried to clarify that now.
This is a crucial memory management function, it's important to have
behavior documented well, even if a full explanation is out-of-scope.
Also added another link in case people want to check more details.
Pull Request: https://projects.blender.org/blender/blender/pulls/134577
The compositor leaks memory in certain setups where a transformed result
is linked to two inputs in the same pixel node. This happens due to an
overestimation in the computation of reference counts of those results.
Since pixel operations might share the same input for multiple links in
the node tree, the reference count should be corrected to take that
sharing into account.
This sharing was previously accounted for as part of releasing inputs in
the pixel operation, but input processors didn't take that into account,
so the realize on domain input processor would leak memory. So to fix
this, we correct the reference count at the evaluator level instead,
such that input processors can safely operate on the correct reference
count.
Pull Request: https://projects.blender.org/blender/blender/pulls/134666
This avoid the dependency on the draw module and remove the
access to a global variable.
# Conflicts:
# source/blender/editors/space_sequencer/space_sequencer.cc
All 2D vectors related to image transform code were changed to float2.
Previously, it was decided, that 4x4 matrix should be used for 2D
affine transform, but this is changed to 3x3 now.
Texture painting code did rely on `IMB_transform` with 4x4 matrix.
To avoid large changes, I have added function
`BLI_rctf_transform_calc_m3_pivot_min`.
Main motivation is cleaner code - ease of use of c++ API, and avoiding
returning values by arguments.
Pull Request: https://projects.blender.org/blender/blender/pulls/133692
Fixed use of lists using multiple arguments - `list` only
support a single argument.
Noticed working with fake-bpy-module - providing multiple
arguments to `list` resulted in typing error (e.g. "Too many type
arguments provided for "list"; expected 1 but received 2").
Pull Request: https://projects.blender.org/blender/blender/pulls/134663
The main issue was that the code was using
`foreach_editable_drawing_with_automask` which shouldn't have been
the case since there is no "masking" to be done when pasting strokes.
Additionally the code was computing the view position of every
pasted point and then computing the offset to move the points to the
mouse position.
Instead, we compute the center of the 3d bounds of the pasted strokes
and then project only the center to the view. From there we can offset
all the points to move them under the mouse cursor.
It's a bit weird to do this projection of the center to the view and
I believe we mainly do it to reuse the `DeltaProjectionFunc`.
There are cases where the projection of the center to the view will fail
which can lead to unexpected behavior on the user end. For now we
just leave the pasted strokes where they were copied from, which
isn't ideal. But this can be improved later.
Pull Request: https://projects.blender.org/blender/blender/pulls/134589
The vulkan backend was implemented with async in mind, however the one place
where Blender uses for async was implemented blocking. This PR splits the
readback into flushing the command and waiting for readback.
**Performance**
Improvement of animation playback performance of shader balls.blend is around 10%.
Shader balls.blend frame: 1-100, 10 x animation playback
| Branch | Total time | Average time |
| -------------------- | ---------- | ------------ |
| blender-v4.4-release | 26851 ms | 2685 ms |
| This PR | 23675 ms | 2367 ms |
Pull Request: https://projects.blender.org/blender/blender/pulls/134227
This ends up being very simple because point cloud has
no connectivity information and each element is just one
position.
Also implement the 3D transform gizmo.
The geometry deformation system isn't implemented in this
commit. That can be tacked later.
Avoid calling `GPU_indexbuf_add_line_verts` and the triangle
version of that function. It's faster to avoid function calls and
just write to the data arrays directly. I did some very rough tests
and observed about a 10% improvement in runtime for the
entire index buffer creation process.
This patch modifies the logic behind handling NDOF device events on
macOS so that it can benefit from the `Blender` profile available in
3DConnexion driver v10.8.7 and later.
A new device command was introduced: `kConnexionCmdAppEvent`, which is
sent by the driver upong getting an apprioriate NDOF device button
input. This allow the driver to consumes all NDOF device input and then
send appropriate app events based on its configuration instead of
forwarding raw data to the application directly.
When using 3DConnexion driver versions prior to v10.8.7, the behavior
is unchanged. This approach allows for supporting all of the SpaceMouse
Enterprise buttons, long presses included (solving issue #119206 on macOS)
Co-authored-by: Sergey Sharybin <sergey@blender.org>
Pull Request: https://projects.blender.org/blender/blender/pulls/126694
The warning to check if two numbers are nearly equal when comparing
floats didn't take negative numbers into account, causing -1.0 & 0.0
for e.g. to be considered not equal but close.
This also resolves a divide by zero.
