Empty hair geometry in Cycles may still report having one curve, even when
there are no actual segments in that curve. This caused an attempt to build
an acceleration structure with zero primitives, which due to other setup
OptiX rejected with an error. Fix that by checking the number of segments
rather than the number of curves in the hair geometry, since the former will
always be zero for empty geometry.
Pull Request: https://projects.blender.org/blender/blender/pulls/115044
This changes a bunch of nodes that have a data type drop-down to using a dynamic
node declaration that changes based on the selected data type instead of always having
all sockets. This greatly simplifies the code and is less weird than having suffixes on
socket identifiers.
Backward compatibility and forward compatibility remain due to #113497 and #113984.
One user-visible change is that changing the data type in these nodes does not break
the link anymore.
It may be necessary to bring back some functionality from link-drag-search afterwards.
Pull Request: https://projects.blender.org/blender/blender/pulls/113553
No functional changes.
Move the function `autokeyframe_pose` from
`transform_convert_armature.cc` to `keyframing_auto.cc`
The function has a lot of overlap with `autokeyframe_object` so it
can potentially be simplified later on.
Pull Request: https://projects.blender.org/blender/blender/pulls/115051
No functional changes.
The function `insert_vert_fcurve` also initialized a `BezTriple` from the flags it got.
Extract that logic into a separate function so it can be reused elsewhere
Pull Request: https://projects.blender.org/blender/blender/pulls/115039
Slightly non-trivial linked collections hierarchy could cause 'Make
LibOverride' operation from 3DView fail, reporting that more than one
potential root collections were found.
Logic here was way to simple, now this error should only happen when
there are effectively more than one potential 'good' root collection
(i.e. when two unrelated collections both contain (directly or not) the
selected objects(s).
There was two problems because these operators were not supported in GPv3:
* If you remove the slot, the material was not reassigned and the material index attribute could get a wrong index. This also affected the unused slots.
* The Unused Materials operator was executed, but it was doing nothing because there was not support for new GPv3 object.
This PR adds support for new Grease Pencil object.
Pull Request: https://projects.blender.org/blender/blender/pulls/114850
Adds the Show All Materials operator.
Also includes a refresh of despgraph when the hide prop changes.
Note: The actual render engine does not support hiding by materials.
Related to #114997
Pull Request: https://projects.blender.org/blender/blender/pulls/115001
The compositor currently executes when editing node trees that are not
compositor related, like geometry and shader node trees.
That's because the node space listener always tags the compositor to
execute for any node edit notifier. To fix this, we add the ID of the
edited node tree as a reference to the edit notifier and only tag for
an update when the space tree matches the reference one.
Pull Request: https://projects.blender.org/blender/blender/pulls/114978
Currently we keep track of 3 command buffers. Data transfer, compute and
graphics. The reason is that data transfer and compute commands cannot
be recorded in a command buffer that has an active render pass.
This PR simplifies the implementation by combining data transfer and
compute commands as there is no need to separate those in individual
command buffers.
This is in preparation of improving submission performance.
Pull Request: https://projects.blender.org/blender/blender/pulls/115033
Use float3, float3x3, and Array for data used for mesh crazyspace
calculation. Propagate the change wherever necessary to not add
more casting to the old C types.
Because `ObjectRuntime` (and therefore `DEGObjectIterData`) became
non-trivial structs, the code that swaps iterators for RNA depsgraph
object iteration had to be changed a bit to be more friendly to C++
memory semantics.
Pull Request: https://projects.blender.org/blender/blender/pulls/114998
In Blender a context should not be shared between threads. In Vulkan a
command pool must not be shared between threads. In the current
implementation the command pool are stored on device level and could
therefore be shared between multiple context which made the implementation
not matching these rules.
This PR moves the command pool from device to command buffers where it
would not conflict between other contexts. This PR doesn't make the Vulkan
backend fully multithreaded. The access to the queue is still missing.
Pull Request: https://projects.blender.org/blender/blender/pulls/114977
This patch allows access to the depth pass in the Viewport Compositor.
Since the depth information require full precision, making use of the
pass requires the full precision option for now. In the future, this
pass will always be stored using full precision regardless of the
precision option.
In 89e3ba4e25, the preprocess step was removed because the
normals are lazily calculated. But now we need to fetch the source normals
from the correct placed-- they still were looked for in `CD_NORMAL`.
When using `SCOPED_TIMER` or `SCOPED_TIMER_AVERAGED`
the display would switch from ns to ms
once the value is over 0.1 ms with a precision of 1.
So when the timer value is hovering in the range of 0.1 - 0.2 ms it is
not giving any useful information.
Fix this by adding another digit to the precision of ms.
Pull Request: https://projects.blender.org/blender/blender/pulls/114724
Caused by bfe6128748.
The split frame logic actually became inverted in the update:
when the cursor position is true, then the frame is supposed
to be initialized in the invoke().
The menu does not use cursor position, so the code path where
the frame is retrieved from the RNA properties was used, and
the frame was never initialized in that case.
Pull Request: https://projects.blender.org/blender/blender/pulls/114973
Seems that the ExponentWithLinearTransform with offset=0 does not render
properly on GPU on Apply Silicon GPUs. Likely, it is possible to use plain
ExponentTransform in this case, which should also be faster.
Candidate for 4.0.1 corrective release.
Pull Request: https://projects.blender.org/blender/blender/pulls/114853
Exposed by 16e4eeb9c0.
In (rare) occasions, a Curve's `CharInfo` can be NULL.
This also crashed going into editmode on affected text objects in
previous versions, however opening and rendering these files was still
fine.
In 4.0, opening those kind of files crashed.
It is not entirely clear how this can happen (there are comments in code
dating back to 97df61a7e5 mentioning "old file", there might be other
reasons, too), still, avoiding the crash on file open seems like good
practice.
Pull Request: https://projects.blender.org/blender/blender/pulls/114916
