This patch changes the operation space of the Flip node from the global
space to the local space. This means that the Flip node will now flip
the image without changing its location.
For example, snapping or pivot point.
The slowdown happens in complex files with drivers which use scene
properties as input. Can be seen in the demo file from #117335.
This change adds ability to mark RNA property as required to bring
the evaluated ID up-to-date, but not do possibly expensive tag for
drivers.
The tool settings possibly can be tagged as do-not-do-deg-tag at
all, but safer seems to follow the current behavior a bit closer.
Pull Request: https://projects.blender.org/blender/blender/pulls/117405
Align the blur filter with the CPU compositor, which uses box filter.
The pre filter was already using box filter, while the post filter was
using gaussian filter.
Additionally, GPU compositor was doing twice amount of blur due to the
mismatch between radius and size semantic: node setting is in size,
while the separable blur expects radius.
Pull Request: https://projects.blender.org/blender/blender/pulls/117454
The calculation of the matte could lead to a division-by-zero
leading to completely different results on different graphics
backends.
Perform the same checks of the input and key saturation before
performing division as it is done in the CPU compositors.
The `node keying matte` test is still failing on Metal, but the
look of the result is much closer to what it would be.
The Tiled default is the least commonly used mapping mode, switch the
default to something that is more commonly used.
Do it for all modes, since according to Julien it is reasonable default
behavior.
Pull Request: https://projects.blender.org/blender/blender/pulls/116933
This is needed so extensions repositories can reference
user-script directories without them having hard-coded paths
which will be invalid when upgrading a Blender version.
Smooth modifier ported to Grease Pencil 3.0
Exposed `smooth_curve_attribute()` from `grease_pencil_edit.cc`
to achieve the smoothing effect. It will not be the exact same result
as the old algorithm.
Pull Request: https://projects.blender.org/blender/blender/pulls/116975
In discussions about #93551, some users mentioned it would be helpful
to have the ability to apply modifiers from the outliner. Adding that
ability is quite simple, and it makes sense since we already have the
option to remove modifiers there.
---
Pull Request: https://projects.blender.org/blender/blender/pulls/117349
If the mix factor means the old values won't be used at all,
the existing corner normals from before the modifier don't
matter and we can skip computing them.
Dragging on a color value slider in the color picker will assert for
an invalid string offset. We now using std::string for uiBut->str, the
string is empty and we are checking against str[1]. This replaces that
with an std::basic_string::find which handles this fine.
Pull Request: https://projects.blender.org/blender/blender/pulls/117455
When the object info node was originally created, the transform of an
object with a negative scale would also have negative scale. At some
point that behavior changed though, and the matrix always had positive
scale.
The negative scale is the more expected behavior (see #104926), and
66dc0ebf2e restored that behavior. But node setups started to
rely on the non-negative behavior in the meantime. So this commit adds
an absolute value node with versioning.
The versioning is added to the 401.10 file subversion bump, since
that's when the behavior changed back to the current state.
Pull Request: https://projects.blender.org/blender/blender/pulls/117425
There are still node versioning bits running after linking, for example
`version_geometry_nodes_replace_transfer_attribute_node`. This code
relies on node types being untouched at that point.
Move the unknown node type replacement from `ntreeSetTypes` into the
`ntreeUpdateAllNew` function. This is called _after_ late versioning.
Pull Request: https://projects.blender.org/blender/blender/pulls/117441
When a shader performs a geometry shader injectoin to work around
features that are not supported natively on the GPU (viewport,
barycentric coordinates, layered rendering), linking would fail.
The reason was that the geometry shader was stored in a slot that was
patched by the specialization constants, resulting in an empty geometry
shader. An empty shader can be compiled, but doesn't match the interface
with other stages, so the linking would fail.
This fixes the issue that EEVEE crashed on Intel iGPUs. These GPUs
don't support viewports.
Pull Request: https://projects.blender.org/blender/blender/pulls/117440
The compositor Texture node differ between CPU and GPU. That's because
CPU falls back to the texture intensity as the alpha channel if no alpha
channel exists. And it also broadcast the texture intensity to all color
channels if no RGB evaluation exist. So adjust the GPU implementation to
follow that.
Rename: anim -> ImBufAnim
Rename: anim_index -> ImBufAnimIndex
There were cases where removing redundant "struct" qualifier caused
a warning since the name of the struct member was also anim.
Use uppercase type name to conform with other types names.
Ref !117394
This PR renames the UI names of EEVEE render engines, due to the decision to
postpone EEVEE-Next to Blender 4.2.
* `EEVEE` is now named `EEVEE-Next`.
* `EEVEE (Legacy)` is now named `EEVEE`.
Pull Request: https://projects.blender.org/blender/blender/pulls/117437
This has been broken since 2.8.
The highlevel reason for this is that the "info file" (saved as frame
zero) was not written anymore.
Reason for that in turn is that when calling BKE_ptcache_write outside
the bake job, the PTCacheID calldata (a psys in our case) is not the
evaluated particle system, causing a check for totpoint to fail (these
were always zero).
Deeper reasoning is unclear, no further investigations were done as to
why/when this happened.
The solution proposed here is using the evaluated psys when writing the
info frame (and this is isolated to just this spot and only to particle
systems). File then gets
written and can/will be read when using this as an external disk cache
in another file.
Pull Request: https://projects.blender.org/blender/blender/pulls/117401
Caused by 089c389b5c. Previously the draw string was created
by appending text inside the buffer multiple times. Now it's formatted
in one step. Part of that change was missing for factor display.
The Tonemap node has a wrong luminance scale. This is because the
parallel reduction shader for logarithmic sum had a wrong identity
value. In particular, its identity was set to 0.0, but since its
initialization macro computed the log, the zero becomes a rather large
negative value.
To fix this, the general structure of the parallel reduction shader was
changed such that the identity is used as is, and not passed to the
INITIALIZE or LOAD macros. This simplifies the implementation and even
avoid the extra texel fetches at the boundary.
GreasePencil v3 requires some additional processing on its data before
writing to blendfile, and after reading from it.
Extract this from generic `read/write` code, and put it into dedicated
callbacks in Cpp classes that need it (layer groups and leaves
currently).
NOTE: Work done while exploring possibility to use 'memfile' read/write
code for GP editmode undo (!117072), but ended up taking a different
approach for it, so this PR is purely a 'no expected changes in behavior'
refactor at this point.
Pull Request: https://projects.blender.org/blender/blender/pulls/117413
This enables raw access to the property, which enables raw array access
to the property through ByteIntAttribute.data, speeding up Python API
access through foreach_get/foreach_set (bpy_rna.cc#foreach_getset).
Pull Request: https://projects.blender.org/blender/blender/pulls/117409
This PR improves the place when shader stages are attached to glPrograms.
Previously it was done when shaders stages where created, in the function
create_shader_stage.
This PR will attach the shader stages inside link program.
Ensuring that create_shader_stage doesn't alter the program, which isn't
clear in its name.
Pull Request: https://projects.blender.org/blender/blender/pulls/117407
`UUID` generally stands for "universally unique identifier". The session identifier that
we use is neither universally unique, nor does it follow the standard. Therefor, the term
"session uuid" is confusing and should be replaced.
In #116888 we briefly talked about a better name and ended up with "session uid".
The reason for "uid" instead of "id" is that the latter is a very overloaded term in Blender
already.
This patch changes all uses of "uuid" to "uid" where it's used in the context of a
"session uid". It's not always trivial to see whether a specific mention of "uuid" refers
to an actual uuid or something else. Therefore, I might have missed some renames.
I can't think of an automated way to differentiate the case.
BMesh also uses the term "uuid" sometimes in a the wrong context (e.g. `UUIDFaceStepItem`)
but there it also does not mean "session uid", so it's *not* changed by this patch.
Pull Request: https://projects.blender.org/blender/blender/pulls/117350
Armature Edit mode has its own undo handling, which didn't store the
number of root bone collections in the armature. Now it does.
An alternative would be to not store this in the undo step, but rather
loop over all bone collections to find the minimum `child_index`. This
would be linear in the number of bone collections, though, so I chose to
use a few more bytes of memory to make the undo system a little faster.
The relations update code does tagging needed to ensure that the
array of bases is updated when relations are updated. It was done
by tagging the Scene ID node, and potentially recursing into all
dependent depsgraph nodes. If there is a driver on a scene property
it was unnecessarily re-evaluated.
This solves the slow behavior of adding objects in the test file
from #117335.
Pull Request: https://projects.blender.org/blender/blender/pulls/117403
