The issue was a floating point precision issue
between the number of the cycle and the time
within the cycle (`cycle` and `cyct`).
Due to that the `cycle` would advance to the
next whole number while the `cycle time`
is still slightly under that.
This is fixed by calculating the `cycle` with double precision.
This has a measurable performance
impact (for `fcm_cycles_time`)
| Before | After |
| - | - |
| 39ns | 44ns |
For fcurves with a cycle modifier,
this code is run at every evaluate call.
The only time when that would be an issue is
in the graph editor, where there is an evaluate
call for roughly every pixel for curves that have a modifier.
However even in that scenario the performance
is the same within run to run variance (for `graph_draw_curves`)
| Before | After |
| - | - |
| 91565ns | 91430ns |
Pull Request: https://projects.blender.org/blender/blender/pulls/123222
When using GPU_SAMPLER_EXTEND_MODE_EXTEND the incorrect sampler
was created, making the OCIO shader fail. This PR selects the
correct wrapping mode (CLAMP_TO_EDGE).
Pull Request: https://projects.blender.org/blender/blender/pulls/123234
This PR hooks up the vulkan backend with the render graph
for drawing. It can run Blender better than the previous
implementation so we also flipped it to be the default
implementation.
**Some highlights**
- Adds support for framebuffer load/store operations
- Adds support for framebuffer subpass transitions
- Fixes workbench shadows
- Performance is just below OpenGL performance when comparing
fps. But the screen feels more fluent when using complex
scenes.
- Current performance is without doing any optimizations so
will improve in the future.
- EEVEE will not crash but has artifacts and many parts that
require more work.
**Related to**
- #121648
- #118330
**Known Limitation**
- Similar to previous implementation resources can be freed when
still in use crashing Blender. This is typically the case when
playing back an animation or updating a material icon.
**Next steps**
- Remove old implementation
- Get EEVEE to work
- Fix double resource freeing
- Improve performance by identifying hotspots and change them
Pull Request: https://projects.blender.org/blender/blender/pulls/121787
Allow assigning integer values beyond int32 range to float/double
IDProperties. Extract the py object value into a temporary int64 value
in these cases.
These two lights have a second culling pass that
uses some shapes defined in view space.
These shapes need to have their handedness flipped
otherwise the test fails.
Fix#122614
Conceptually, these are the same as IDProps used for regular dynamic RNA
data (should have been done that way from the beginning). At least make
them statically typed, to avoid all kind of issues when the IDProp type
change and does not match expectations from the geometry nodes anymore.
Ref. #122743.
Pull Request: https://projects.blender.org/blender/blender/pulls/122891
All IDProperties generated as part of 'storage backend' for dynamic RNA
properties are now statically typed.
Also adds some basic unittests for the new statically typed IDProperty
system, based on py-defined RNA data.
Ref. #122743.
This implements (most of) the proposal in #122743:
* Add a new `IDP_FLAG_STATIC_TYPE` IDProperty flag.
* Update `BPy_IDProperty_Map_ValidateAndCreate` and related to never
change an existing property type if statically typed.
The biggest change happens in bpy assignement code, since instead of
replacing the old exisitng property by a newly created one, and copying
over a few settings, now the old property is kept if possible, and a new
one is only created if needed.
And in case the existing property is statically typed, if it cannot be
re-used to store the given value, and error is reported and it remains
unchanged.
`IDP_ARRAY` is also supported for basic numeric types, so 'vector'
properties and such work as expected. Lentgh is considered as part of
the static type (i.e. one can only assign a 3 components py sequence to
a 3-len array property, etc.).
Such in-place update is not yet implemented for `IDP_IDPARRAY` and
`IDP_GROUP` types. While important (especially the group one), they are
not that critical for the current issues related to changing IDProperty
types.
Variable `leftColor` is added to global variable `overlay_edit_smooth_color_iface` in current version.
In result it is added to every following shader using `overlay_edit_smooth_color_iface`.
Pull Request: https://projects.blender.org/blender/blender/pulls/122803
The hidden socket in the Viewer is not shown when a new link to it is created.
This is because the hidden property of the socket is not updated when a new
link to it is created.
Fixing it by simply manually update the property after a new link is created.
Pull Request: https://projects.blender.org/blender/blender/pulls/120503
This could happen for non manifold meshes (like a water plane)
when using the accurate method.
If last hit in the volume ray is a frontface, we now consider
everything behind it in volume.
I didn't think the BMesh extraction mode could arrive at the case where
only loose edges are requested, but turns out it can because of the
mesh wrapper system where the evaluated mesh is actually a BMesh with
deformed positions.
Related to #119999 and #122059.
In case the _source_ PropertyRNA was unset (i.e. its underlying
IDProperty storage did not exist), the copy operation would silently
fail.
In fact, the existing code handling IDProperties separately in
`RNA_property_copy` was pretty bad, since it would also bypass all the
RNA 'setting value' code (like custom setters, update handling).
Turns out, liboverride RNA apply code can already handle all of these
cases, so simply pass the raw 'unresolved' RNA property to it, and
remove all this special handling code from `RNA_property_copy`, solves
all the issues.
Reorder `Action::assign_id()` so that there is a clear "unassign any
previously-assigned binding" step first, and then a clear "assign the new
binding" step. Previously these were a bit too much intertwined.
Also `animrig::unassign_binding()` now simply calls into `Action::assign_id()`
instead of having somewhat overlapping responsibilities.
Finally the RNA setter for `AnimData.action_binding_handle` is changed
to always call `Action::assign_id()` for both assigning and unassigning
the handle. This ensures the API is properly used, instead of modifying
properties directly.
Pull Request: https://projects.blender.org/blender/blender/pulls/123184
