This is preparation for #106228, the crash does not happen in `main` currently.
The issue was `positions_for_write` might make a copy of the data if it was
shared. This operation is not thread safe on a single mesh. The solution is to
get write-access to the positions only once before threading starts.
For example
```
OIIOOutputDriver::~OIIOOutputDriver()
{
}
```
becomes
```
OIIOOutputDriver::~OIIOOutputDriver() {}
```
Saves quite some vertical space, which is especially handy for
constructors.
Pull Request: https://projects.blender.org/blender/blender/pulls/105594
The UI needs persistent pointers to catalog paths that it can pass
around via context. It was trying to manage this in own storage, but
this didn't work. Not entirely sure why - the menus get redrawn
continuously while assets load, and the root menu rebuilds the mentioned
storage every time. Maybe the submenus redraw a bit later, or the
keyboard navigation handling still accessed data from a previous redraw
somehow.
Either way, instead of managing own catalog path storage, point into the
asset system, which already has persistent storage for the catalog
paths.
Pull Request: https://projects.blender.org/blender/blender/pulls/106237
Don't install libraries in blender.shared and don't generate manifest
as it's not working and not important to have a clean directory here.
Always detect MSVC_REDIST_DIR as it is used by oneAPI. The way
InstallRequiredSystemLibraries is called it already only sets some
variables and doesn't install anything by itself.
Pull Request: https://projects.blender.org/blender/blender/pulls/105999
The overall goal is to use implicit-sharing in many places in Blender
that currently do unnecessary copies. See #95845 for more details.
This commit only adds the base data structures in blenlib and uses those
in `GeometrySet` and `AnonymousAttributeID`, which used a more ad-hoc
version of implicit sharing already. #105994 lists some more places where
support for implicit sharing can be added (most notably: custom data layers).
Pull Request: https://projects.blender.org/blender/blender/pulls/105994
Previous fix/hack a60f651502 made it so that all 'Fake User' linked
IDs are considered as directly linked, so that some IDs (like linked
Text) references are not lost on file save becaus they are not
effectively used.
However, this created unwanted references to all linked assets data,
because an asset always has its fake user set. This ended up keeping
lots of references to sub-assets when appending from the new Essentials
node groups.
NOTE: This is adding another hack to the collection, the whole handling
of 'which unused linked data reference to keep on file save' needs to be
reworked for Blender 3.6.
No user visible changes expected.
With this, empty rows will be added to the tree view so that the
background box is at least a few lines high (like with UI lists). If the
view is used as a drop target, data can be dropped on these empty rows
too then.
This was requested for the Cycles light linking project.
This PR adds support for clearing framebuffers and scissor testing.
Tweaks had to be made to VKTexture to keep track of its layout
on the device. Based on the actual call the layout can switch
to a more optimum layout.
For example during upload of a texture the texture will be converted
to a transfer destination optimized layout. When reading from the
texture it will be converted to a transfer source optimized layout.
The order of the attachments in the framebuffer follows the next rules
- When only color attachments are there the color attachments will
be placed in the slot they are defined. This way it will match
the ShaderCreateInfo binding location.
- When a stencil/depth attachment is added it will be placed
right after the color attachments. When there isn't a color
attachment it will be the first attachment.
Pull Request: https://projects.blender.org/blender/blender/pulls/106044
A few small refactors of the `draw_setting_widget()` function for drawing
animation channel widgets (like 'modifier mute', etc.)
- Use `const` where appropriate.
- Move some variable declarations to their actual use, to clarify their
scope and make it possible to use `const`.
- Return earlier where possible, which should improve performance as some
unnecessary code gets skipped.
No functional changes.
Improvements to packing efficiency and updating UV packing API.
* Add #pack_islands_alpaca_rotate, the fast alpaca-style packer with rotation.
* Update heuristic for sorting islands based on longest edge.
(Improves pack efficiency when rotation is enabled.)
* Add `aspect_y` to UV Packing API, to support non-square materials in the future.
Pull Request: https://projects.blender.org/blender/blender/pulls/105977
When all of the input meshes have no loose edges, the output won't have
any either, but finding that out for the output is much more expensive.
Copy this tag to the output mesh to save processing time when drawing.
In my test creating 1 million instances of a 16 face mesh, this changed
the time for the draw cache loose geometry extraction from 220 ms to
131 ms.
For now, only do this when the loose edges of the inputs are already
cached (primitive nodes, for example). This avoids paying the cost of
calculation when we don't know it will be necessary.
It doesn't take much time to know whether all of the curves have more
than a single point, and if that's the case, we know for sure there are
no loose edges in the mesh. This can save time when creating the mesh
draw cache. The time for finding loose geometry goes from 180 ms to
104 ms for a large molecular nodes test file.
For the single-threaded part of the curve to mesh conversion, avoid
writing to two logically separate arrays in the same loop. Instead
run the calculation on two separate threads if the output is large.
Timing for `calculate_result_offsets` when processing 1 million curves:
Before: Average: 10.4 ms, Min: 9.4 ms
After: Average: 9.2 ms, Min: 7.7 ms
Now that mesh primitive nodes calculate their bounds in constant time,
it's worth updating the bounds eagerly when translating a geometry
since doing so should be very cheap and might save a more significant
amount of time if they're needed later.
Also slightly unify the translation functions for the different types.
Pull Request: https://projects.blender.org/blender/blender/pulls/106159
Previous code was using BKE_mesh_uv_vert_map_get_vert to somewhat
detect identical UV values on mesh vertices. But this was not
handling the case when separate mesh vertices still use the same UV
values (e.g. cube with all six faces mapped to whole image).
Replace usage of BKE_mesh_uv_vert_map_get_vert with a simple "unique UV
value" map, very similar to how OBJ normals are calculated for export.
Measurements on a somewhat extreme case: exporting "Rungholt" Minecraft
level from https://casual-effects.com/data time (Win10, Ryzen 5950X)
goes 2.9sec -> 2.2sec, and resulting file size 486MB -> 231MB. This
particular model has a lot of mesh faces mapping to the same places
in UV map.
On less extreme cases the timings are similar between old and new
code, with a tiny speedup in most tests I've tried.
Allow clicking the "up" and "down" arrows to increment/decrement by a
frame for Video Sequencer strip Frame Start, Strip Offset Start, and
Strip Offset End.
Pull Request: https://projects.blender.org/blender/blender/pulls/106117
"Rotate Normals" is a changeable operation like any other and does not
need to be hardcoded.
An advantage of exposing this modal is that the shortcut key now
appears in the header when rotating an edited mesh.
Avoid calling `restoreTransObjects`, `resetTransModal`,
`resetTransRestrictions` and `initSnapping` for each change.
Also bring `applyMouseInput` close to the rest of the code.
No functional changes.
The lack of organization in the values of the constants that indicate
the properties was prone to errors.
One error that can be noticed is that the values of
`P_ORIENT_AXIS_ORTHO` and `P_VIEW2D_EDGE_PAN` were the same.
(not really a problem though)
