The issue was that if the evaluated object doesn't have an active
layer, the canvas is offset by twice the objects position.
The reason was that the code scaled the whole transformation
matrix by two to match the code in 4.2, but then had to overwrite
the location of the transformation again to counteract the scaling.
The fix is to not counteract the scaling and just scale the 3x3 part
of the matrix instead. This way we can remove the part afterwards
that writes to the location of the transform.
Pull Request: https://projects.blender.org/blender/blender/pulls/130454
For flat caps, only one point is added in data.positions array for
endpoints of stroke. It results in incomplete outline. To resolve
this, add another point in new curves at the exact opposite position.
Pull Request: https://projects.blender.org/blender/blender/pulls/130437
When points are close together in 2D view space the eraser intersection
function does not have a valid direction vector (zero). This cancels out
terms in the computation that would still contribute to the result and
the function always returns zero intersections.
This patch adds a special case when the direction vector is zero,
so the point distance to the circle is still calculated (just not in
any particular direction).
Pull Request: https://projects.blender.org/blender/blender/pulls/130341
This is not implemented yet so remove the callback function which draws
channel color for GPv3 layers. Also `acf_gpl_channel_color` is a legacy function,
it may lead to different channel color as explained in #130370.
Pull Request: https://projects.blender.org/blender/blender/pulls/130445
The issue was that when the fill strokes were merged into the drawing,
the `fill_opacity` attribute would be created on the drawing
(when it didn't exist before) and initialized to 0. This meant that
all previously drawn fill strokes would disappear.
The fix manaully initializes the `fill_opacity` attribute on the
target geometry before calling `join_geometries`.
Ideally this could be something the attribute system handles
or the `join_geometries` function takes care of by getting
some information about how to initialize attributes.
Pull Request: https://projects.blender.org/blender/blender/pulls/130442
The crash was caused by the selection attribute not being
initialized properly. The size was 0.
The fix uses `ed::curves::ensure_selection_attribute` to ensure
the attribute is created and on the right domain.
Extrude should only work for a point selection, so the poll
function of the operator is updated to reflect this.
Pull Request: https://projects.blender.org/blender/blender/pulls/130333
1. Trace operator was missing a topology cache tag after replacing the stroke
data. This caused an invalid normals cache and crash in drawing code.
2. The drawings must not be manipulated outside the main thread. Job data now
stores a curves geometry array that is used to update the drawings at the end of the job.
3. Since the job data now stores an array it has to be allocated using MEM_new.
Co-authored-by: Falk David <falk@blender.org>
Pull Request: https://projects.blender.org/blender/blender/pulls/130313
wdas_cloud.blend file on 4.3+ raises an assert, due to reading some
VSE data:
readfile.cc:4798, blo_verify_data_address(), at
'MEM_allocN_len(new_address) >= expected_size' Corrupt .blend
file, unexpected data size.
This is caused by 30dbb7820d which removed completely unused
data from Sequence struct. But it turns out, Sequence serialization
inside scene_blend_read_data is "somewhat strange", to put it mildly,
and silently assumes that struct offset of seqbase and channels
will never change.
For now, restore the previously expected struct member offset,
ensure it stays like that via static_assert, and add notes on
how someone should fix this in a better way.
Pull Request: https://projects.blender.org/blender/blender/pulls/130296
On Linux we can't always determine topmost window under cursor,
therefore check for overlapping windows and, if so, only use the
active window. This allows dragging between windows as long as
they don't overlap
Pull Request: https://projects.blender.org/blender/blender/pulls/130242
Set unique name for `target_layer_name` to display in pop-up box when
creating a new layer. For that, `unique_layer_name()` is moved to member
function of `GreasePencil`.
Pull Request: https://projects.blender.org/blender/blender/pulls/130208
Similar to the handling of this cache in `flush_update_step` (after a
brush stroke calculation), the face normals cache needs to be cleared.
Sculpt mode doesn't have granular updates implemented for face corner
normals currently.
Pull Request: https://projects.blender.org/blender/blender/pulls/130248
Brush assets from the current file should be supported like other brush
assets.
This assert would fail on startup for example, when opening the
essentials file containing the default brush of this mode.
Similar issues to previous problems in `IMAGE_OT_open` and the
`BKE_image` API... Relative filepaths of linked editable IDs must be
relative to the ID's library blendfile, not the current 'local'
blendfile.
Use new `_in_lib` BKE_image API to re-use or load an image from/in the
expected library, including using the library path as root path for
relative file paths handling.
Using the new `BKE_image..._in_lib` API also allows to get rid of the
`check_exists` 'hack' introduced in `image_open_exec` (when support to
edit some linked data was added, in 5f9f3116db) to avoid re-using local
Image IDs for editable linked data (brush assets), since
`BKE_image_load_exists_in_lib` will now ensure to only re-use an Image
ID if it matches the target library.
This commit also simplifies/makes `ED_image_filesel_detect_sequences`
signature more to the point. It replaces the `Main` parameter by a mere
`root_path` string. The given `bmain` was indeed only used to extract
(again) a base for relative paths.
NOTE: The general idea of this fix is the same as in PR !130058, thanks
to Sean Kim for the investigation and initial fix.
Pull Request: https://projects.blender.org/blender/blender/pulls/130197
Previously, we had to fetch all asset previews at once. However the preview
caching system was limited at 8192 previews, so with large asset libraries, many
previews would get dropped. Instead, allow selectively requesting previews and
only request previews of items displayed on screen. This should fix the issue,
and reduce the time it takes to display brush assets in general, especially with
bigger libraries. This should also reduce memory usage quite a lot when
displaying the asset shelf, since not all asset previews are loaded anymore.
Pull Request: https://projects.blender.org/blender/blender/pulls/130105
This commit refactors `BKE_image_load` and `BKE_image_load_exists` APIs:
* Remove the `_ex` versions (the 'exist' boolean return pointer can have
default `nullptr` value instead).
* Add `_in_lib` versions, which match signature and behavior of the
generic ID creation code to allow to find or create a new image ID
directly in a library 'namespace' (as linked data).
This is required by upcommig fixes for Brush Assets, which are using
linked but editable data-blocks.
Fix#130194: When trying to add a new image from the UI (e.g. for a new
texture) for a local ID, if that same exact image was already loaded by
a linked ID, it would trigger an assert in `BKE_id_move_to_same_lib`,
because `BKE_image_load_exists` would return the matching linked Image ID
instead of creating a new local one. In release builds with no assert,
it would result in making a linked ID 'local', while still being used by
original other linked data.
Passing around the intended final destination of the new Image (local or
in a given library) allows `BKE_image_load_exists` to be more specific
when searching for an already loaded matching image ID, and ensures that
a new local Image ID is created in the case described above.
Pull Request: https://projects.blender.org/blender/blender/pulls/130195
Partial fix for #129267.
Avoid creating an entry in the file list cache for every asset when iterating
over the available assets. For this, allow passing a pre-filter function to the
asset iterator, to filter based on `AssetRepresentation` data. The asset shelf
can then pre-filter based on asset type and the active catalog.
Really, the cached file entry is only needed to access the preview image of the
asset, so it can be avoided most of the time.
While this should still be solved properly as outlined in #122439, this is
a general improvement and fix the mentioned report in a minimal way.
Was hoping this would be sufficient to fix#129267, but the whole of
!130105 will be needed.
Pull Request: https://projects.blender.org/blender/blender/pulls/130166
Caused by 71398478b2.
`parsubstr` is nullptr in Layer struct. This results in crash when an
object is parented to a new layer. This can be fixed by passing layer
member function `parent_bone_name()`. It handles the null pointer.
Pull Request: https://projects.blender.org/blender/blender/pulls/130205
* Early return in case the given `id` is already in the same library as
its target `owner_id` (in addition to both being local data).
* Assert with message AND return in case the given `id` is a linked ID
from a different library.
The second point somewhat mitigates the severity of #130194.
There is actually a valid (rare) case where a Library ID can have a `0`
versionfile value during blendfile loading: if that library only
contains 'indirect weak' linked data.
These IDs are only actually read in the library blendfile if they are
also used by some other data, otherwise they are purposedly dropped and
removed, and the library blendfile can end up never being read at all.
This would lead to a completely empty library on blendfile loading,
which can then be removed from the data-base.
NOTE: this _does not_ affect cases where the whole library is missing,
or some linked IDs from that library are not found anymore.
The data-block selector filters out local data-blocks when trying to
assign to linked data. A lot of data-block management relies on this.
Since these data-blocks are created for the use with brushes only,
create them under the same library that the brush is from, as indirect
data. The UI allows selecting these data-blocks then. We do the same for
textures attached to a brush, and some other data-block types.
Also add paint-curves to the editable data-block types, so they can be
edited despite being considered linked data.
Pull Request: https://projects.blender.org/blender/blender/pulls/130184
The operator to move a channel to the top revered the layer order
by iterating from the top to the bottom and moving selected channels
to the top. Iterate from the bottom to the top instead to keep the right
layer ordering.
Pull Request: https://projects.blender.org/blender/blender/pulls/128793
The cache mechanism for line art is changed during migration to GPv3,
however the code path failed to handle following cases which could lead
to a few problems:
- Line art cache isn't deleted after last line art modifier because it
coule be hidden, causing memory leaks.
- A modifier inside a multiple line art modifier sequence that doesn't
use cache would prematurely delete line art cache, causing subsequent
line art modifier to give empty result.
- When the first line art modifier is hidden (in viewport/render), the
cache is not created correctly, leading to crashes.
Now the new code logic addresses these problems properly by:
- Making sure the last visible line art modifier deletes cache.
- Giving a fresh cache pointer for modifiers that doesn't use global
cache.
- Line art cache is correctly ensured when there are modifiers that
are hidden
Pull Request: https://projects.blender.org/blender/blender/pulls/129953
We need to be careful about not just looking up active tool data from
context, because this code path may also be run from the Properties,
which use the tools of the 3D view. So make sure related code falls back
to the 3D view, like it's already done in some other places.
Caused by removing scene sound from "main scene" when other strip using
the same scene is not rendering sequencer strips.
Only remove sound when all strips using particular scene are not
rendering sequencer strips.
Pull Request: https://projects.blender.org/blender/blender/pulls/129768
The fill tool didn't write to the `fill_opacity` attribute leading
to non-filled strokes in some cases.
This makes sure that the attribute is created with full opacities.
Pull Request: https://projects.blender.org/blender/blender/pulls/130177
NOTE: This also required some changes to Cycles code itself, who is now
directly including `BKE_image.hh` instead of declaring a few prototypes
of these functions in its `blender/utils.h` header (due to C++ functions
names mangling, this was not working anymore).
Pull Request: https://projects.blender.org/blender/blender/pulls/130174
Reverts part of 0872e898bc.
In 4.2, the fill tool had a small paint cursor that could
be turned off using the "Display Cursor" setting.
This makes it so that the cursor for the fill tool behaves
the same as in 4.2.
Turns out the sampling module was not initialized when
computing lighting which resulted in undefined shadow location.
In some cases even producing NaN position which would make the
shadow disappear.
Basically, anything returned by `sampling_rng_nD_get` was returning
garbage, and on some occasion, it could be a negative number and
produce NaN if passed into `sample_disk` or any `sqrt`.
ASAN did not catch it because the uninitialized data load
was done on the GPU.
I ran the render tests and, minor difference aside, they all match.
#### Follow Up:
This raises the concern that we have no assertion for correctly
using some modules. At the very least, we should have some
mechanism to catch uninitialized buffer. Investigating this
was a pain au chocolat.
Pull Request: https://projects.blender.org/blender/blender/pulls/130150
The length modifier was not handling 2 point curves correctly.
This change does two things:
1) Fix the crash by copying the original indices for 2 point curves into `dst_to_src_point`.
2) Fallback to `extend_curves_straight` for extending 2 point curves.
Pull Request: https://projects.blender.org/blender/blender/pulls/130122
The Double Edge Mask node uses too much memory when it is not connected
to anything in the Full-Frame compositor. That's because it tries to
allocate a buffer of size 131072x131072, which is due to the cumulation
of multiple issues.
1. The Full-Frame compositor evaluates nodes that are not connected to
the output.
2. The constant folder assigns a bounds of +-65536 to its output memory
buffers as a maximum possible value.
3. Then the MemoryBuffer::inflate method doesn't know about the constant
folder special value, so it tries to allocate a buffer of that size.
The safest fix for now is to clear the output and return early if all
inputs are single value.
The problem is that this issue might happen for other nodes that also
satisfy the 3 issues above, that is, uses inflate and can be constant
folded. I manually checked for such nodes and there are no nodes that
satisfy all three, so hopefully this is isolated to the Double Edge Mask
node.
Pull Request: https://projects.blender.org/blender/blender/pulls/130163
Mistake in e494a44024
Even though collection/object PointerRNA argument is same in most cases,
use `layer_or_collection_prop` in if() before assigning value to
collection property (`base_or_object_prop`argument value is null in function
call from `_isolate_flag()`).
Pull Request: https://projects.blender.org/blender/blender/pulls/129250
Instead of looping through all drawings in the vertex group kernel
function, loop through all/visible drawings in the operators and
pass a single drawing to `remove_from_vertex_group`.
Continuing !129789.
Co-authored-by: Falk David <falk@blender.org>
Pull Request: https://projects.blender.org/blender/blender/pulls/129890
Add some missing `nullptr` and `ale->type` checks to ensure the
assmptions the rest of the code makes holds.
The direct cause of this issue is likely
822907a68d. The root cause is the hard to
predict behaviour of the animation filtering system combined with a lot
of pointer casts without testing whether the data is actually of the
expected type.
Pull Request: https://projects.blender.org/blender/blender/pulls/130148
`set_active_node` call from `grease_pencil_copy_data` resets inactive
layer again as locked before executing the drawing code. Since property
is only operated on layers, it is possible to remove the code from
`set_active_node()` and only have it in `set_active_layer()`.
Pull Request: https://projects.blender.org/blender/blender/pulls/129465
