This speeds up saving `070_0100.anim.blend` from the Heist project
from ~3s to ~300ms by adding PROPOVERRIDE_IGNORE in a few
places. It's not completely obvious to me when `PROPOVERRIDE_IGNORE`
should be used and when it shouldn't. Given that the same is done for
meshes already, it seems correct.
Pull Request: blender/blender#107196
Add texture usage flags for textures which are used as texture views
or require texture views for backing implementation.
Authored by Apple: Michael Parkin-White
Pull Request: blender/blender#107163
Apply compilation fixes for Metal compatibility.
This includes explicit type casts, packed data types
where vec3 alignment is inconsistent, constructor replacement
with factory function.
The Metal shader generator also needs knowledge of when bound
resources are fundamental data types, so
SHADOWS_TILE_DATA_PACKED must be described as uint in
ShaderCreateInfo.
Authored by Apple: Michael Parkin-White
Pull Request: blender/blender#107178
Caused by 6b8cdd5979.
Above commit introduced element tagging for boundary calculations but
only cleared them properly on all faces if the new `Preserve Seams`
option was chosen. We cannot be sure about the state of element tags
from prior operators though, so correct the culprit check to also only
be in effect if the new `Preserve Seams` option was chosen.
Pull Request: blender/blender#107161
Leverage new API call in Metal to increase the number of threads
dedicated to concurrent shader compilation. First step to improve
parallel compilation times when multiple engines are active.
Would also enable an increase in worker threads for shader
compilation jobs within the DRWManager.
Note that this is only available in the latest
version of macOS Ventura (13.3).
Authored by Apple: Michael Parkin-White
Pull Request: blender/blender#106625
The result of detecting if a quad should flip the default 0-2 split
when tessellated only used a pre-calculated normal when available,
since the method of detecting the flip was different, the check for a
concave face could change depending on the existence of polygon-normals.
In practice this meant cycles render preview could use a different
tessellation than the GPU display.
While [0] exposed the bug, it's an inherent problem with having 2
methods of detecting concave quads.
Remove is_quad_flip_v3_first_third_fast_with_normal(..) and always
use is_quad_flip_v3_first_third_fast(..), because having to calculate
the normal inline has significant overhead.
Note that "bow-tie" quads may now render with a subdivision in a
different direction although they must be very distorted with both
triangles along the 0-2 split pointing away from each other.
Thanks to @HooglyBoogly for investigating the issue.
[0]: 16fbadde36.
While this didn't cause bugs, initialize members to avoid problems
in the future.
GHOST::SystemX11
- m_keyboard_vector
- m_keycode_last_repeat_key
GHOST::Window
- m_cursorGrabInitPos
- m_userData
Generally, one does not know if the sharing info is currently shared
and should therefore be const. Better keep it const almost all the
time and only remove the constness when absolutely necessary
and the code has checked that it is valid.
The main challenge is to avoid dangling pointers. Currently, the lifetime of socket
declarations is somewhat unbounded (at least we didn't restrict it explicitly yet).
Therefore, storing non-owning pointers in it tricky. For ID pointers one could
potentially use the foreach-id iterator to update pointers in declarations as well,
but that's a bit out of scope and might not be the right solution anyway, since it's
not obvious that all node declarations are reachable from IDs stored in `bmain`.
The solution now is to use a callback that retrieves the right ID pointer when it
is used. The important thing is that the callback does not capture any potentially
dangling pointer either.
Pull Request: blender/blender#107179
When running oneAPI with AoT binaries, on hardware that's not compatible with
these, recompilation could have been missing from the kernels loading phase and
happen during execution instead.
These changes fixes it, any kernel compilation will now happen during the
kernels loading phase.
It was broken in two ways:
- bpy_bmlayercollection_iter passed PY_SSIZE_T_MIN, while
PY_SSIZE_T_MAX was needed.
- bpy_bmlayercollection_subscript_slice() contained an
off-by-one error.
Pull Request: blender/blender#107165
Completes the TODO in GHOST_SystemPathsCocoa::addToSystemRecentFiles
Also renames the filename parameter to the more appropriate filepath.
The recently opened/saved file will now also show up in:
- Blender Dock icon > Right click.
- Three finger swipe down in Open Blender i.e., App Expose
Based on a earlier contribution by @jenkm.
Pull Request: blender/blender#107174
After e45ed69349 we need to remove the existing attribute
when adding the rest position before evaluating modifiers. Also, adding
the rest position attribute was completely skipped in edit mode.
Pull Request: blender/blender#107190
The crash can occur in the following situations:
- Attempt to open a corrupted EXR file
- Attempt to save an EXR file under a non-existing directory.
The root cause is not really clear: for some reason the OpenEXE API on
the Blender side can not catch OpenEXE exceptions by a constant
reference to a std::exception, although it can by a constant reference
to an Iex::BaseExc.
This does not seem to be an issue with the OpenEXR library itself as
the idiff tool from our SVN folder catches the exceptions correctly.
It is also not caused by the symbols_apple.map as erasing it does not
make the problem go away.
It could still be some compiler/visibility flag which we were unable
to nail down yet.
The proposed solution is to add catch-all cases, mimicking the OIIO
tools. This solves the problem with the downside is that there are
no friendly error messages in the terminal. Those messages could be
brought as part of the workaround by additionally catching the
Iex::BaseExc exception. But probably nobody relies on those error
prints anyway, so added complexity in the code is likely does not
worth it.
Pull Request: blender/blender#107184
Animation: Adds a new "Parent Space" Orientation option for the Transformation Gizmo.
---
For child targets (objects, bones, etc) being able to transform in parent space is a desired feature (especially when it comes to rigging / animation).
For objects:
* with a parent, the gizmo orients to it's parents orientation
* without a parent, the gizmo orients to Global space
For Armatures:
* Child bone shows parent's space regardless if "Local Location" is set for parent bone
* For root bone **without** "Local Location" set, use the armature objects space.
* For root bone **with** "Local Location" set, use local bone space.
---
No new transformation orientation code needs to be written, we can achieve the desired results be using the existing `transform_orientations_create_from_axis`, `ED_getTransformOrientationMatrix`, and `unit_m3` methods. To do this, we check to see if the bone has a parent, if so, we use the bones pose matrix (`pose_mat`). This is done similarly for objects using the parent's object matrix (`object_to_world`).
Pull Request: blender/blender#104724
Scene.cpp and Geometry.cpp are large file it can be broken up into smaller easier to handle files. This change has been broken out from #105403 to make understanding the changes easier.
geometry.cpp is broken up into:
1. geometry.cpp
2. geometry_attributes.cpp
3. geometry_bvh.cpp
4. geometry_mesh.cpp
scene.h & scene.cpp is broken into:
1. scene.h
2. scene.cpp
3. devicescene.h
4. devicescene.cpp
Pull Request: blender/blender#107079
When a parent item was renamed, the `TreeView` was doing everything as
expected, however `AssetCatalogService::update_catalog_path` is supposed
to also update the catalog paths of all sub-catalogs [which it does --
but it does not tag sub-catalogs as having unsaved changes, resulting in
wrong saving of catalogs afterwards, meaning the parent item was saved
with the old name and a new item with the new name was created].
Now also tag sub-catalogs for having unsaved changes.
This should also go into 3.3 LTS
Pull Request: blender/blender#107121
The selection (box select, click select...) had an offset when selecting keys in the timeline.
That was because the function to get the NLA mapping ignored the timeline.
Pull Request: blender/blender#106904
The following properties were not library overrideable, but now are
* Line Thickness
* Color
* Custom Color Checkbox
Pull Request: blender/blender#106959
Currently when a baked curve is in the Graph Editor and normalization is enabled, it doesn't work.
It even throws a warning.
This patch adds the missing logic to normalize baked FCurves within a preview range.
Pull Request: blender/blender#106890
When using Xcode version 14.3 on Apple Silicon hardware a number of
regression tests fails. This change fixes this problem.
The root cause comes to the floating point contraction. It was already
disabled for GCC on Linux, but not for Clang on neither of Linux or
macOS.
Also corrected the comment about Clang default, as it as set to on
somewhere in 2021.
Pull Request: blender/blender#107136
Occlusion query buffers not being cleared to zero resulted in
erroneoues selection in certain situations.
Authored by Apple: Michael Parkin-White
Pull Request: blender/blender#107135
Shader source requires explicit conversions and shader address
space qualifers in certain places in order to compile for Metal.
We also require constructors for a number of default struct types.
Authored by Apple: Michael Parkin-White
Pull Request: blender/blender#106219
This patch adds several tools and options to the weight paint mode of Grease Pencil.
* Blur tool: smooths out vertex weights, by calculating a gaussian blur of adjacent vertices.
* Average tool: painting the average weight from all weights under the brush.
* Smear tool: smudges weights by grabbing the weights under the brush and 'dragging' them.
* With the + and - icons in the toolbar, the user can easily switch between adding and subtracting weight while drawing weights.
* With shortcut `D` you can toggle between these two.
* The auto-normalize options ensures that all bone-deforming vertex groups add up to 1.0 while weight painting.
* With `Ctrl-F` a radial control for weight is invoked (in addition to the radial controls for brush size and strength).
* With `Ctrl-RMB` the user can sample the weight. This sets the brush Weight from the weight under the cursor.
* When painting weights in vertex groups for bones, the user can quickly switch to another vertex group by clicking on a bone with `Ctrl-LMB`.
For this to work, follow these steps:
* Select the armature and switch to Pose Mode.
* Select your Grease Pencil object and switch immediately to Weight Paint Mode.
* Select a bone in the armature with `Ctrl-LMB`. The corresponding vertex group is automatically activated.
Pull Request: blender/blender#106663
This patch implements the ID Mask node for the realtime compositor.
The node can be implemented as a GPU shader operation when the
anti-aliasing option is disabled, which is something we should do when
the evaluator allows nodes be executed as both standard and GPU shader
operations.
Pull Request: blender/blender#106593
`make format` uses autopep8 to format Python, using a list of paths
specified in `tools/utils_maintenance/autopep8_format_paths.py`. The
scripts folder used to be a submodule inside release, but it is now at
the root of the blender repo.
This commit adds `scripts` to the list of paths to format.
Ref !107143
The comment didn't really make sense, since the removed code did the
same thing as the CustomData function anyway, and that's already done
in `mesh_init_data`.
