Some compositing applications do not support spaces and dots in layer names, and change
these to other symbols on import. This causes various compatibility issues, including
with Cryptomatte metadata. While technically those could be considered bugs in the
Cryptomatte implementation of other software, where they are not properly accounting for
that layer renaming, it's not ideal.
The OpenEXR channel naming convention is "layer.pass.channel". We get away with dots in
the layer name since we parse this from right to left, but it's a weak assumption.
Now we don't forbid using spaces or dots, and existing files are unchanged. But at
least by default names will be compatible, and hopefully other software catches up in
time to support more flexible layer names.
Ref T68924
OpenEXR uses "." to separate layers/passes/channels, so using AOV.001 is
a problem. Other applications will not be able to parse it correctly.
Default to AOV_001 instead, and don't allow using dots in AOV names.
Fixes T89991
Ref T73266
Ref D12871
Need to initialize components for the full Diffuse BSDF.
Steps to reproduce:
- Default cube scene
- Switch to Cycles renderer
- Enable OSL backend
- Start viewport render
- Observe cube being much black
Differential Revision: https://developer.blender.org/D12921
A Clang-Format configuration to make the closure definition block to
be properly recognized as such.
Also small wrapper macro to avoid comma in the actual definition code
which was causing unwanted indentation of parameters definition.
Requires Clang-Format 7 or newer. The version we ship in the libs is
12, so for recommended development setup it should all be good.
Differential Revision: https://developer.blender.org/D12920
The thread affinity setting in OIDN can break multithreading on some CPUs.
While this leads to somewhat worse performance on CPUs that do work correctly,
it's better than having some CPUs use only half the cores.
The lookup tables are to be initialized after device free.
On Linux was only noticeable when rendering default cube scene with
an extra assert. On Windows it was causing an assert in STL in debug
builds.
Differential Revision: https://developer.blender.org/D12918
We do this in other nodes to reduce overhead of using the same node more
than once. I don't think it will make a difference with index nodes
currently, but at least it's consistent.
noise.cc uses std::min and std::max without
including the algorithm header required.
Newer MSVC versions and GCC implicitly include
it somewhere, which isn't something we should
count on. Best to include what you use.
Even though these nodes do support fields internally, there are no
input string sockets that support fields currently, so removing
the diamond sockets for now helps avoid confusion.
Differential Revision: https://developer.blender.org/D12828
Split thumbnail job in 2 passes. First pass will render visible images
and second part renders set of "guaranteed" equally spaced images.
When viewing larger amount of strips, it is likely that only 1 or 2
images will be rendered in first pass, while in second pass it is up to
30 images.
This results (seemingly) in 3x better performance, but zooming before
set of guaranteed images is done will be slightly more inaccurate.
During rendering VSE cache is invalidated, so thumbnails would be
removed and thumbnail job would constantly restart.
Even if thumbnails would be preserved, resources should be dedicated
for rendering job.
When the 'Separate Children' option is selected, the children of
the selected collection are inserted into the geometry output
sorted alphabetically by name.
One item to note is that the rename function does not trigger a
depsgraph update. This means that the changes are not reflected
in the spreadsheet until another action triggers the update.
Differential Revision: https://developer.blender.org/D12907
Move the static functions higher in the file so they are usabl
for an upcoming patch, and make it use clearer names instead
of overloading a function name.
Pencil
Not naming the auto-generated vertexgroup after the selected bone was
just confusing (since the group would not have an effect), so now use
similar code that is used for meshes for greasepencil as well.
Maniphest Tasks: T92314
Differential Revision: https://developer.blender.org/D12906
* Rename struct KernelGlobals to struct KernelGlobalsCPU
* Add KernelGlobals, IntegratorState and ConstIntegratorState typedefs
that every device can define in its own way.
* Remove INTEGRATOR_STATE_ARGS and INTEGRATOR_STATE_PASS macros and
replace with these new typedefs.
* Add explicit state argument to INTEGRATOR_STATE and similar macros
In preparation for decoupling main and shadow paths.
Differential Revision: https://developer.blender.org/D12888
Instead of switch statements, make use of generic virtual arrays
so the code is shorter and easier to read.
Differential Revision: https://developer.blender.org/D12908
It seems possible to switch object selection (if `Lock Object Modes` is
turned off) and end up with an object that has a SculptSession but a
NULL PBVH.
(I was not able to repro from scratch, but file from the report was
clearly in that state).
This would crash in displaying scene statistics.
While there might be a deeper fix (making sure PBVH is available early
enough -- possibly using `BKE_sculpt_object_pbvh_ensure`,
`sculpt_update_object` or friends), there are also many checks in tools
for PBVH, so the situation seems to be somewhat vaild/expected also in
other places.
So to fix this, just check for a non-NULL PBVH, returning early
otherwise.
Note: this leaves us with displaying 0/0 Faces & Vertices in the borked
case until an operation takes place that updates the PBVH.
Maniphest Tasks: T92246
Differential Revision: https://developer.blender.org/D12904
The old code only worked when built-in nodes are only unregistered
at most once while Blender is running. However, this is not the case
when running certain unit tests such as `AbstractHierarchy*`
in `blender_test`.
Found by Sybren, thanks.
In an attempt to reduce scrolling. This can be re-visited if a kind of switch
between "compact" and "comfortable" UI size is implemented in the future.
On GCC in release mode (and maybe also debug mode without ASAN enabled),
allocating an `AssetLibraryService` will reuse the space that should
have just been freed. This made a test fail, as it was testing that new
memory was allocated and not some old instance reused.
To ensure that the calls that should allocate a new block of memory
return a unique pointer, I added some dummy allocation to the test.
No functional changes to Blender
Previously, every node had its own declaration. This isn't ideal, because
it's often the case that all nodes of the same type have the same declaration.
That's the case for all nodes using declarations currently. It will not be true
for e.g. group nodes in the future.
Sharing node declarations between nodes makes it a bit more efficient.
Differential Revision: https://developer.blender.org/D12898
This was caused by `DRW_view_data_texture_list_size_validate` which now
delete everything from the render engine.
This might change in the future but for now we just avoid calling it from
the render loop (when using DRW_cache_restart).
CD_PROP_COLOR vertex data is stored in scene linear while legacy vertex
colors are srgb, so both operators also need to do this conversion
Reviewed By: sergey
Maniphest Tasks: T79005
Differential Revision: https://developer.blender.org/D8320
