Due to rounding errors in `wm_gesture_straightline_do_angle_snap`, it
could give out a snapped result that's meant to be 45 degrees but does
not have equal X/Y distances after converted to integer. Since 45
degrees can be used to make a equal length structure, it's best to make
sure that resulting X/Y values are the same if the angle is multiple of
45 degrees.
This fix would be effective for all tools that uses this "straight line
gizmo".
Pull Request: https://projects.blender.org/blender/blender/pulls/128683
Due to rounding errors in `wm_gesture_straightline_do_angle_snap`, it
could give out a snapped result that's meant to be 45 degrees but does
not have equal X/Y distances after converted to integer. Since 45
degrees can be used to make a equal length structure, it's best to make
sure that resulting X/Y values are the same if the angle is multiple of
45 degrees.
This fix would be effective for all tools that uses this "straight line
gizmo".
Pull Request: https://projects.blender.org/blender/blender/pulls/128683
This patch fixes broken connections and effect strips which would arise
in some cases when trying to split connected strips with effects.
This happened in case an effect strip was being split -- first, its
effect chain would be added to a VectorSet of seqs, but then, only the
connected strips of the original seq (in this case the effect strip)
would be added to the VectorSet. Fix by iterating over the entire list
and adding connected strips.
Additionally:
- Add a final SEQ_iterator_set_expand() for SEQ_query_effect_chain() in
case these connected strips had effects themselves.
- Move the seq_edit_split_operation_permitted() check after the entire
chain is evaluated to avoid bugs w/ locked strips being editable
before.
Pull Request: https://projects.blender.org/blender/blender/pulls/128670
The issue was that if the seqslide was cancelled such that the
`new_frame` of the left handle was greater than its previous
right_handle (before the cancel), `SEQ_time_left_handle_frame_set()`
would clamp this value and so the left handle would not properly restore
itself.
Fix by updating both handles at once (right handle first) if both
handles are selected and the cancel jump travels in the right direction.
This also seems to fix the situation where strips w/ both handles
selected would be erroneously detected as overlapping when moving them
rapidly side to side.
Pull Request: https://projects.blender.org/blender/blender/pulls/128740
If points are beind the camera, we don't really want to erase them. This
patch marks invalid coordinates thus preventing them from intersecting
with a eraser.
The reason for using a large value to indicate "invalid coordinate"s are:
- No need to further break down the way we process `src_to_dist` point matching array for `hard_eraser` `soft_eraser`, makes the entire logic much easier.
- No eraser is gonna be touching such a large coordinate of `1e20`.
Technically there's this case where if a segment crosses the near or far clipping plane (to handle this correctly, you'll need to split that segment into two at the clipping plane position and it increases complexity a lot), and then you will have undefined erasing behaviour, however the worse case is that the one segment was completely removed, and in such case I think it's acceptable.
Pull Request: https://projects.blender.org/blender/blender/pulls/128738
PBVH is null when redoing the move operation. To fix this,
update active object data from evaluated object before sending
undo_push.
Also fixed the crash in cache_init() after updating depsgraph
Pull Request: https://projects.blender.org/blender/blender/pulls/128625
Part of the brush assets project followups, see #116337.
Based on feedback, it seems important to indicate to the user when a brush has
unsaved changes.
There's no reliable updating mechanism we can use or hook into here, except for
RNA "update" callbacks. Brush data gets changed in many places in code, the only
way to do this seems manual tagging every time a brush property gets changed.
This PR introduces `BKE_brush_tag_unsaved_changes()` for this. I spent some time
going through all brush properties to ensure changes call the tagging function.
A known limitation with this will be that changes to dependencies won't be
indicated in the brush. E.g. Changing the texture attached to a brush won't make
the brush be indicated as changed.
The UI to indicate the changed brushes is being discussed still, see #128846.
Pull Request: https://projects.blender.org/blender/blender/pulls/128845
While splitting areas the two parts are shown with rounded outlines,
but some of the underlying content can show through in the corners,
spoiling the effect. This PR just erases anything in those corners,
making it look cleaner and rounder.
Accidentally merged to main first with c3ea941273.
While splitting areas the two parts are shown with rounded outlines,
but some of the underlying content can show through in the corners,
spoiling the effect. This PR just erases anything in those corners,
making it look cleaner and rounder.
Pull Request: https://projects.blender.org/blender/blender/pulls/128901
Print in `RNA_struct_free` would crash in ASAN builds because they can often
access memory (struct identifier) that is owned by Python/BPY, and
therefore already freed at exit time.
If a `DrawGroup` contained both inverted and non-inverted scale
the command generate shader would output the `resource_id`
content at conflicting indices. This is because the number of
instances stored inside the `DrawGroup` are the original
count before multiview. Actually, only `start` was taking the
multi-view count into account.
We cannot modify the value on CPU otherwise it would increase
the instance count for each submission. So the fix is to
pass the view count to the command generate shader and
multiply the instance count where needed.
Fix#128085
Pull Request: https://projects.blender.org/blender/blender/pulls/128854
This is likely caused by local_ray_up being degenerate with
very small radii. This is a temporary fix and should be revisited
later.
The issue is that the real fix is likely to have a higher
performance cost.
Fix#124636
Avoid processing function arguments directly and instead create
temporary variables for adjustments. This is done to make it easier to
maintain identical code to CPU.
- Replace blender::Map with an array to lookup heap-nodes by edge index.
- Only create the array when the topology influence is used.
These gives ~5-9% overall speedup to operator function.
The main goal is to simplify adding support for nested node panels. The patch
makes use of the updated recursive node declarations introduced in
6ffc585fb8.
The main changes are:
* Rewritten node drawing in a way that makes ui design decisions like panel
visibility and margins more explicit. Especially the handling of margins is
much better now imo. Previously, it was very hard to change the margin for
specific cases without accidentally breaking other situations. Now each
possible case has an explicit margin. This needs a few more lines of code but
is much easier to work with.
* Rewritten node drawing in panel (sidebar + material properties) using the new
ways to iterate over the declaration.
* It's possible to add custom layouts at any point in the node declaration now.
This also replaces the need for having a `draw_buttons` callback for panels.
Pull Request: https://projects.blender.org/blender/blender/pulls/128822
Building proxies for images was always saving them as JPG files.
However when input images are floating point (e.g. EXR files), this
loses both range and precision, making the resulting proxies
not really be useful for anything where you'd be using EXR files.
Change this to save float image proxies as EXR instead (FP16 data,
lossy DWAA compression). In my quick tests this does result in about
3x-4x larger proxy file size compared to JPG, however at 50% DWAA
quality that is still 10-30x smaller than original EXRs would be.
Changed proxy image loading to explicitly tell to load metadata. JPGs
were always loading it anyway, but EXRs only load when instructed to
do so.
Pull Request: https://projects.blender.org/blender/blender/pulls/128835
This patch implements the domain realization algorithm for the new CPU
compositor. Only nearest interpolation with no wrapping is implemented
at the moment.
A new sampling method was added to the result class and some relevant
methods were moved into inline functions.
Sequencer preview area drawing, when displaying float (EXR/HDR) final
image using GPU based display transform, was converting input float32
data to float16 data for the GPU texture, and the GPU was displaying
that float16 texture.
However this texture is displayed for just one frame, and the cost of
doing the conversion was pretty high. Just send incoming float32
data to the GPU instead.
Playback of EXR sequence at 2048x858 resolution, on
Ryzen 5950X / RTX 3080Ti: 28fps -> 70fps (time to do just the GL
texture upload: 28ms->3ms)
Pull Request: https://projects.blender.org/blender/blender/pulls/128829
When using OpenGL on Intel ARC the driver reports a max 3d allowed size
of 2048. The volume probe will create a texture that doesn't fit in this
dimension when selecting a probe size of 512 or 1024 MB.
This PR will reshape the volume pool atlas texture until it found a shape
that is optimal and fit on the device. When reshaping selects a different
pool size a warning message will be displayed as it might change the
visual quality.
When reshaping the smallest row size will be selected in order
to improve the occupancy. Reshaping will only happens when a
different setting is set in the `Performance->Memory->Light Probes Volume Pool`.
NOTE: Needs to be backported to 4.2
Pull Request: https://projects.blender.org/blender/blender/pulls/128877
The error message related to shadow updates in `eevee_shadow.cc`
currently contains a typo:
`"Error: Too many shadow updates, some shadow might be incorrect."`
This sentence should use the plural form of "shadows" to correctly
describe the context.
Fixing this typo ensures clarity and correctness in the error message,
providing developers and users with the appropriate feedback when
encountering shadow update issues.
Pull Request: https://projects.blender.org/blender/blender/pulls/128865
Technically a regression in [0] although it matches behavior prior
to v3.3 going back to 2.7x. The fix for #101883 [1] added a check
for the vertex number changing after the number was zeroed,
causing the shape key to be cleared in most cases.
While the handling of shape-keys in OBJECT_OT_convert wasn't well
defined - clearing the shape-key means in the evaluated coordinates
are always used so it is preferable.
Now the operator ensures the old (un-evaluated) shape-key isn't used.
[0]: 0791f53029
[1]: be32882e1c
