Make GPU_viewport_colorspace_set() const-crrect w.r.t view_settings.
Instead of doing in-place modifications of the view_settings argument
with restoring them later introduce new function for copying view
settings which keeps curve mapping unchanged in the destination.
Should be no functional changes.
Pull Request: https://projects.blender.org/blender/blender/pulls/138189
These are not the same. Debug Scope are for scoped
capture. These should be static to avoid flooding
debug tools with scopes. Since this code is
unlikely to be debugged, it is better to use
debug groups.
Fixes for #137856 & #137507 were workarounds to ignore the
PROP_PATH_SUPPORTS_BLEND_RELATIVE for CollectionExport however they
were error prone - especially #137856 as it relied on the surrounding
buttons.
Instead of adding logic to ignore the flag for this specific case,
add a property to the CollectionExport which can have it's flags
and use this instead of the operators "filepath" property.
Internally this wraps the same IDProperty data used by the operator.
Ref !137883
When tweaking envelope radii of (connected) envelope bones, we dont
always make sure that corresponding head and tail radii are in sync.
Using the `Bone Size` tool should take care of this (and it works to
some extend) but tweaking the radii through the sidepanel UI or the
Properties Editor (through RNA), then corresponding heads and tails
radii get out of sync.
Once we are in such unfortunate situation, then code in the transform
system's `recalcData_edit_armature` fails with wrong assumptions, it
propagates radii from children to parents which are unexpected / not in
use.
So one thing to do would be to add this syncing of radii to
`rna_Armature_editbone_transform_update`.
This alone would solve the "problem" in new files.
For existing files that are already out of sync we add versioning that
corrects this on file load
Pull Request: https://projects.blender.org/blender/blender/pulls/137599
This converts the public `uiLayoutColumnFlow` and `uiLayoutGridFlow` functions
to an object oriented API (an `uiLayout::column_flow` and `uiLayout::grid_flow` respectively),
matching the python API.
This reduces the difference between the C++ API with the python version,
its also helps while converting code from python to C++ code (or vice-versa),
making it almost seamless.
Both renamed functions now returns an `uiLayout` reference instead of a pointer.
New calls to this method should use references too.
Part of: #117604
Pull Request: https://projects.blender.org/blender/blender/pulls/138074
Previously UV sync-select in face mode would isolate UV selection
so dragging a face would disconnect it from adjacent faces.
Now the sticky mode is respected for face selection.
This is needed for #136817 so storing per-selection doesn't result
in a user visible change in behavior.
When a object matrix has a determinant of 0 (which typically means 0s on
one or more of the scale axes), the matrix inversion would become
invalid thus the inversion matrix will be quietly set to diag(0). This
would basically reset all vertices to the origin of the object after
applying the inversion matrix. This happens in the "apply rotation"
stage.
Now addtional warnings will be shown to the user when there are
objects that tries to go through such operation, and they will be
skipped.
-------
It will look like this when some objects failed, if you want to test, just scale some objects to 0 and apply rotation:
Pull Request: https://projects.blender.org/blender/blender/pulls/128273
Clamping behavior in the new slip operator (added in #137072) would
silently fail if any strip(s) in the selection did not have enough
underlying content to fill the strip bounds.
This patch allows the user to continue slipping along a clamp defined
by the strips that support it. A warning in the status bar
is also added in this case.
If no strips can clamp, the button is disabled entirely, with only the
warning present.
Images available in the PR.
Pull Request: https://projects.blender.org/blender/blender/pulls/137980
The code that updated the ray based on the constraint point was
incorrect in orthographic view. In this view, it's the ray origin that
changes, not the direction. This has now been fixed.
Additionally, as a further improvement, when snapping is not available,
a fallback point is now used as the constraint point.
Pull Request: https://projects.blender.org/blender/blender/pulls/137776
In all paint and sculpt modes, it was practically impossible for modal
operators to set a modal cursor, because the cursor was immediately
overwritten by the paint cursor. In this patch the paint cursor function
checks for a modal cursor first. When a modal cursor exists, it isn't
overwritten by the paint cursor.
Pull Request: https://projects.blender.org/blender/blender/pulls/137893
This commit reduces the Mesh paint BVH node limit on vertices from
10000 to 2500. In general, a lower value here improves performance for
both raycasting and brush strokes, as smaller leaf nodes mean more nodes
can be filtered in the broad sphere tests instead of needing each
vertex to be processed.
There are a number of considerations that go into profiling this value:
* This change is most impactful for large meshes, where the user is
going to be modifying a small total percentage of the model's area.
* Lower values may improve CPU performance, but overall responsiveness
as perceived by the user goes down with values much smaller than 2500.
This is likely due to GPU overhead, as each BVH node corresponds to a
VBO.
* A lower node limits increase the BVH building time.
The new limit of 2500 corresponds to the following performance changes
as compared to the prior baseline:
* 1.12x ~ 1.52x FPS increase depending on the ratio of brush to mesh
size.
* 1.27x ~ 2.88x performance increase for the brush evaluation.
* 1.31x slower BVH build times on a mesh with 10 million verts
(0.2864s to 0.3761s).
Further performance data and comparison videos can be seen in the
associated PR.
Pull Request: https://projects.blender.org/blender/blender/pulls/137871
Previously, the bundled essential brushes in Sculpt Mode were changed
from the Scrape, Fill, and Flatten types to the Plane brush type. This
commit does the same thing programmatically and removes the now old code
related to these obsolete brushes.
Resolves#134076
Pull Request: https://projects.blender.org/blender/blender/pulls/138022
Adjust the logic in `view_zoomstep_apply_ex` to ensure that spaces
without the new flag `V2D_ZOOM_IGNORE_KEEPOFS` will still zoom if they
do not also have `keepofs` flags set.
This means spaces like Geometry Nodes can zoom again.
The correct order is View Space -> World Space -> Object Space, which
implies that the view to world matrix `cache->vc->rv3d->viewinv` has to
be applied before `ob.world_to_object`.
Introduced in 5a21fa0ada
Pull Request: https://projects.blender.org/blender/blender/pulls/138146
Allow configuring the inline buffer capacity for the slots array, and
add an inline buffer for the keys vector. Previously there was always
an allocation when adding an element.
The inline capacity is manually configured in a few places as part of
this commit.
Pull Request: https://projects.blender.org/blender/blender/pulls/136461
Two separate fixes are needed to resolve the crash. First is in the
store named attribute node, which incorrectly retrieved an attribute
value validator without checking the field type. The other is various
missing checks for the case when the evaluated positions for a
specific curve are empty.
Pull Request: https://projects.blender.org/blender/blender/pulls/137861
Create an option in `View2D.flag` that allows us to have separate
behavior for area resizing (left aligned with `V2D_KEEPOFS_X`) and
zooming (centered with the keepofs flags disabled). Add this to the
versioning code.
Pull Request: https://projects.blender.org/blender/blender/pulls/138041
Caused by 06f6d77979.
Legacy curve objects have not taken the radius into account in the
past. Better to stick with that behavior and only affect the new
curve object type. This only affects non-instanced legacy curve
objects; generally generated curves are presented to the depsgraph
iterator as the new object type.
Pull Request: https://projects.blender.org/blender/blender/pulls/137852
See #129009 for context.
The preprocessor parses metadata and writes a header file containing
an inline function that inits the `GPUSource` with the metadata.
These header files are then included inside `gpu_shader_dependency.cc`.
This still keep the usage of the `metadata` enums and classes to avoid
pulling the whole blender module inside the preprocessor executable.
This speeds-up startup time in Debug build:
`gpu_shader_dependency_init`
- Before : 37ms
- After : 4ms
I didn't measure release, but it is unlikely to be noticeable (in the
order of 4ms > 1ms).
Pull Request: https://projects.blender.org/blender/blender/pulls/138070
This avoid recreating the GPU context for each individual
tests. This reduces the overhead drastically.
Excluding static_shaders and texture_pool tests I get for GPUVulkanTest:
`Before: 129 tests from 1 test suite ran. (26304 ms total) `
`After: 129 tests from 1 test suite ran. (6965 ms total) `
Including static_shaders and texture_pool tests I get for GPUMetalTest:
`Before: 124 tests from 1 test suite ran. (54654 ms total)`
`After: 124 tests from 1 test suite ran. (1870 ms total)`
Given the tests are run twice for the workarounds versions, the
speedup can be multiplied by 2.
Overall tests time is still largely dominated by shader compilation time.
However, there is still 3x improvement using this patch:
Including static_shaders and texture_pool tests I get for GPUVulkanTest,
GPUVulkanWorkaroundTest, GPUOpenGLTest, GPUOpenGLWorkaroundTest:
`Before: 516 tests from 4 test suites ran. (318878 ms total)`
`After: 516 tests from 4 test suites ran. (106593 ms total)`
Pull Request: https://projects.blender.org/blender/blender/pulls/138097
This improves the user experience, because the data-flow is not interrupted
temporarily. The automatically inserted links are exactly the same that the
Evaluate Closure node uses when no closure is connected.
Pull Request: https://projects.blender.org/blender/blender/pulls/138125
The corner-cost bias used when calculating the shortest path meant
the path could cross over itself.
Resolve by tagging vertices as well as edges to prevent stepping across
vertices used by edges that have been added to the heap.
This removes the minimum thickness clamping in the shader.
The reason why this was clamped in the first place was to reduce
aliasing artifacts. With the new super sampling method for
rendering, this should no longer be an issue.
Note: This can break visual compatibility, but the previous radii
were arguably "wrong". This essentially fixes this and renders
the strokes with the actual radii from the attribute.
Previous files that might have relied on the clamping will have
to be updated.
Pull Request: https://projects.blender.org/blender/blender/pulls/138119
Metal and Vulkan don't support line smoothing. There is a workaround
implemented. This workaround is only enabled when linesmooth value is
larger than 1. However When using smooth lines it should also be used.
This is fixed by adding a `GPU_line_smooth_get` function for getting the
current line smooth state.
Pull Request: https://projects.blender.org/blender/blender/pulls/138123
