Was caused by D6068, which did not handle "MEM_PIXELS" memory
when not in background mode. Before that it always fell back to using
generic device memory, so restoring that behavior. In future this
should be changes to use OpenGL interop for optimal performance.
When there was no way to find the data-path from context
the shortcut was still being created.
It would evaluate to "context.(null)".
Now adding shortcuts will be disabled if the path can't be computed.
Large objects with many separate pieces became unstably slow
(run for hours and not finish).
The entire original mesh was being duplicated twice per loose part.
In own tests, millions of vertices and thousands of loose parts
now run in around 5-15 seconds.
This fixes the crash, but it does not fix the core issue. The PBVH should
always be available when an object is in sculpt mode and tools should
not need to check for that.
Reviewed By: jbakker
Maniphest Tasks: T70790
Differential Revision: https://developer.blender.org/D6063
Was cause by recent fix for T65134 which assigned original object's
proxy_from to an evaluated pointer.
This is because motion path depsgraph does not include proxies, so
the pointer in an evaluated object was kept pointing to an original
object.
Caused by rB46102cf4e0c4 [which removed the check if the image can
actually be loaded].
Maniphest Tasks: T70903
Reviewed By: campbellbarton
Differential Revision: https://developer.blender.org/D6089
The OptiX implementation wasn't trying to allocate memory on the host if device allocation failed, while the CUDA implementation did. This copies the implementation over to OptiX to remedy that.
Differential Revision: https://developer.blender.org/D6068
'PE_set_data' / 'PE_set_view3d_data' would give us a depsgraph already,
so use it.
Also fix access to PEData->depsgraph without calling 'PE_set_data' prior.
Addresses concern raised in rBcf2c09002fae.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D6067
When displaying the voxel size for an adaptive domain the resolution of the adaptive domain was used to calculate the world size of the voxel.
This patch changes this to use the initial size of the domain.
When using adaptive domain the overlay was not rendered in the right
place.
Thanks to sebbas for part of the patch!
Reviewed By: sebbas, fclem
Differential Revision: https://developer.blender.org/D6076
Caused by what appears to be a missing flush from evaluated bone back to
original, which then makes it so copy-on-write operation happening after
click (to synchronize selection flags) pushes original bone to its initial
position.
Differential Revision: https://developer.blender.org/D6051
All the single-value texture inputs of Principled BSDF node should use
non-color colorspace profile, not sRGB one (issue raised in
https://blender.stackexchange.com/questions/155617, thanks).
That also revealed another issue - since those color space settings are
stored at the image level itself, not the node one, we need to
duplicate those image data-blocks when we use same picture for e.g. base
color (sRGB) and specular (non-color) inputs...
For now using a basic mechanism for that, might generate several extra,
uneeded copies of the image ID, but that’s better than breaking custom
settings and such.
Note that while this will modify the behavior of the impporters using
that node wrapper, no change should be needed in IO add-ons themselves.
The auto-smoothing flag can now be used by artists when the Alembic file
does not contain custom loop normals.
- Auto-smoothing disabled: mesh is flat-shaded.
- Auto-smoothing enabled: works as usual; set angle to 180° to ensure
a 100% smoothed mesh.
When the relative mode was used, the calculation of the total number of vertices was not done and it was using the total number of vertices in the datablock. This worked for small files, but with complex files the time to allocate all the data was too long and the performance was very bad.
Now, for relative mode the real number of vertex is calculated.
Also fixed the same problem when onion and multiedit is enabled.
This was caused by 2 things: Shadow map bias and aliasing.
It made the expected depth of the shadowmap further than the surface
itself in some cases. In normal time this leads to light leaking on normal
shadow mapping but here we need to always have the shadowmap depth above
the shading point.
To fix this, we use a 5 tap inflate filter using the minimum depth of all
5 samples. Using these 5 taps, we can deduce entrance surface derivatives
and there orientation towards the light ray. We use these derivatives to
bias the depth to avoid wrong depth at depth discontinuity in the shadowmap.
This bias can lead to some shadowleaks that are less distracting than the
lightleaks it fixes.
We also add a small bias to counteract the shadowmap depth precision.
Before this patch you could go to a local view with a single object,
while you had other objects also in edit mode, and your operators would
affect all objects even the ones outside your local view (same for local
collection).
Differential Revision: https://developer.blender.org/D6064
Background dithering was introduced to solve banding issues on gradient backgrounds.
This patch will enable dithering based on the texture that is used for drawing.
Only when using a GPU_RGBA8 texture the dithering will be enabled.
This disables dithering for final rendering, vertex and texture paint modes.
Reviewed By: fclem
Differential Revision: https://developer.blender.org/D6056
