For Sculpt Undo, certain operators will modify the topology count of the
mesh. These operators are handled separately from normal brush strokes,
and so having tests for an operator that uses this functionality is
beneficial in detecting regressions.
Pull Request: https://projects.blender.org/blender/blender/pulls/139249
Previously, when a socket was detected to be unused, it was just grayed out.
This patch adds support for automatically hiding unused sockets based on this
convention: Menu inputs control visibility while other inputs only control
whether something is grayed out.
More specifically, an input is visible if any of these conditions is met:
* It affects the output currently.
* It never affects the output. In this case its usage does not depend on any
menu input.
* It is used if all non-menu inputs are considered to be unknown.
In the future, we could support customizing which inputs are allowed to control
visibility. For now it's good to use the convention that Blender generally
follows itself.
As before, panels are grayed out if they only contain grayed out sockets and
panels are hidden when they don't contain any visible sockets.
Hiding inputs works in group nodes, the Geometry Nodes modifier and node
operators. In theory it will work for all node tree types, but since only
Geometry Nodes supports the Menu Switch node currently, this patch currently
only makes a difference there.
The implementation reuses the existing `SocketUsageInferencer` with a different
sets of inputs. So no new core-inferencing logic was needed.
Design task: #132706.
Pull Request: https://projects.blender.org/blender/blender/pulls/138186
Regression in [0] caused the poll function to compare the result with
nullptr instead of std::nullopt.
The operators exec function then de-referenced the the std::nullopt.
[0]: 550094b018
The particle system generates some particles with NaN values. The
set_if_different mechanism skipped copying those due to a refactor
in the matrix equality test. Revert that part of 689633d802 for now.
A better solution would be to improve handling of NaNs in Cycles,
and to find and fix the cause of the NaN in the particle system.
Pull Request: https://projects.blender.org/blender/blender/pulls/139238
Test was failing since efdda78175. The asset library path was now using
backslashes on Windows, and got compared to a path with forward slashes. Use
our higher level path comparison function to make sure they point to the same
directory.
When displaying an asset library that contains a back slash on macOS (or Linux,
I assume), resizing the asset browser would make all assets disappear, and
instead the message would show that indicates an invalid asset library path.
Convert all slashes to native format when initializing an asset library. This
might convert slashes that are valid parts of the file name, but this just
leads to an error about a not found asset library, which is better than
crashing. This is a typical tradeoff when dealing with cross platform paths.
Now that handles are effectively hidden with the "Tweak Handles" option
(and clear themselves if `use_restore_handle_selection` is set),
handle selection for multiple strips is quite unintuitive.
It is expected that, after selecting a group of strips, the user will
be able to tweak all of their handles, even if they are not connected,
consistent with other NLEs.
This patch enables that behavior by propagating handle selection to
all selected strips if "Tweak Handles" is on.
Pull Request: https://projects.blender.org/blender/blender/pulls/139075
Originally, `element_already_selected` would always return `true`
as long as the user clicked on a strip that was already selected,
regardless of its handle selection state.
This led to unexpected behavior when trying to tweak a strip:
E.g. taking a strip, selecting its left handle, then trying to tweak
the strip itself would actually tweak the left handle unexpectedly, even
though the cursor would be far enough away that this does not make sense.
A lot of this is very messy and I keep running into
the same weird edge-cases, so hopefully more documentation
will help for the time being, until I can rework the entire
selection system in 5.0.
This should make VSE code more readable and easier to understand from an
outside perspective.
The name was chosen to be `channel` rather than `channel_index` to keep
things short and concise -- it should be clear based on the context
whether we are talking about the strip's channel index (singular case,
`Strip::channel` or `SeqTimelineChannel::index`) vs. the channel list
(plural case, e.g. `Editing::channels`).
Pull Request: https://projects.blender.org/blender/blender/pulls/138919
Some Linux multi lib setups have a helper include file for Jemalloc that
in turn includes the actual header file. This makes our version regex fail.
As the Jemalloc version we are checking for is no longer in any of
the currently supported LTS linux distros, we can safely drop it.
Pull Request: https://projects.blender.org/blender/blender/pulls/139225
For default values, the fairly complex reasoning for the code is explained
in details in comments. The TL;DR: would be that this code is needed to
safely load blendfiles older than 2.83, which were saved without any DNA
type information for these default value data.
Storage data have always been written with DNA info (added in commits
3bae60d0c9 and 9d91bc38d3), so no need to add special handling for them.
Just use regular DNA struct reading.
Pull Request: https://projects.blender.org/blender/blender/pulls/139175
Previously, when reordering a tree view item to the end, one had to move it drop
it fairly precisely on the lower half of the last item in the list. This patch
makes it possible to just drop it below the last item too. The outliner has the
same functionality.
Pull Request: https://projects.blender.org/blender/blender/pulls/139025
Currently, nothing happens when trying to scroll while hovering a tree view that
is fully visible (i.e. does not have a scroll bar). This patch makes it so that
the scroll event is passed through in this case so that one can scroll the
entire region even when hovering the tree view. If the tree view does have a
scrollbar, then the scroll-event is still captured by that even if scroll all
the way up/down already.
Pull Request: https://projects.blender.org/blender/blender/pulls/138930
Node UIs can now have panels. Some of those may need to have their
labels translated using translation contexts. PanelDeclarations
already had a translation_context member, this commit adds a way to
specify this context, and to use it for translation on drawing the
node.
Pull Request: https://projects.blender.org/blender/blender/pulls/139124
Reported on devtalk: MotionBuilder produced FBX files contain
"cameras" that are not really user visible cameras, but rather map to
MotionBuilder viewports. They are at root, have no child elements,
and have special names. There is also a "camera switcher";
ignore that too.
Pull Request: https://projects.blender.org/blender/blender/pulls/139204
This patch turns the options of the Crop node into inputs.
Instead of specifying the bounds of the crop, the inputs now specify the
Width and Height of the crop region. The Crop Image Size option was
renamed to Alpha Crop and inverted, so it now defaults to actual
cropping. The Relative option was removed, as it is now superseded by
the Relative To Pixel node, and removal was done to facilitate the
options to inputs project.
Reference #137223.
Pull Request: https://projects.blender.org/blender/blender/pulls/139163
The Relative To Pixel node considers the original resolution of the
image with no transformation, while it is more useful for it to consider
the realized size instead, because that's what later processing nodes
will be using.
If the cursor has never been over the mesh and a stroke is drawn that
starts off of it, passes over the mesh, and then ends off of the mesh,
then when the active index is accessed, it is possible that the BMesh
indices and tables are not initialized.
Pull Request: https://projects.blender.org/blender/blender/pulls/139184
This was used to cache a lazily evaluated viewer field in the past, but is
unused for a while now already. If we need it again, it can be implemented with
`GenericKey` instead of a custom version of that.
Pull Request: https://projects.blender.org/blender/blender/pulls/139192
A small quality of life change to indicate which slots are used.
This only applies to the slots menu, the UI List could show these too
however this requires extending the RNA API.
Ref: !138651
This converts the public `uiItemFullO_ptr` function to an object
oriented API (an `uiLayout::op` overload), matching recents changes
in the API.
Changes include rearranging the `IDProperty *properties` parameter to be
the last parameter. Now is optional (but will be removed), also instead
of using a return parameter the function now returns the pointer to
write properties.
Pull Request: https://projects.blender.org/blender/blender/pulls/139166
With the release of the brush assets project in 4.3, most brush
management moved out of the current .blend file into either the
essentials library or user-created libraries.
With this change, a number of workflows became more difficult:
* Handling a large library of texture and texture properties for brushes
due to ID linkage constraints.
* Having local tweaks to a brush without bloating the the asset library
and reducing discoverability.
This commit introduces an intermediate step to assist with both of the
preceding pain points. The `brush.asset_save_as` operator is extended
to allow saving into the current blend file via the `Duplicate Asset`
context menu entry.
Additionally, these features help ease authoring brush assets in general
from the UI instead of requiring manual datablock management.
This allows brushes to be stored alongside other data inside a specific
blend file.
Related to #129655 and [1].
[1] https://blender.community/c/rightclickselect/XYMA/
Pull Request: https://projects.blender.org/blender/blender/pulls/138105
This commit adds the `BKE_brush_duplicate` function that performs a deep
copy of a brush ID and all associated IDs into the current Main
database.
Unit tests are added with this case to ensure that grandchildren and
embedded data is handled correctly for both Sculpt and Grease Pencil
brushes, as the latter contains more ID references than other brush
types.
Related to #138105
Pull Request: https://projects.blender.org/blender/blender/pulls/138629
Since ec141ba3ff, commits from all branches are listed. Issue is that
commits from different repositories and branches will be mixed.
Personally I prefer grouping commits better, so I always edit the output
manually.
With this, commits will be grouped by branch, or PR if one can be found.
The branches and commits will then be printed under their target
repository. This isn't the owning repository of the branch or PR (often
people's personal fork), but the one it targets. This worked much better
in own tests.
Further:
- For some common repositories more readable names are used, e.g.
"Blender Manual" instead of "blender/blender-manual".
- Commits of the Blender repository are listed first, without grouping
under a `blender/blender`.
- Commits to each repository's main branch are listed first, directly
under the repository grouping.
- See PR for a textual mockup of the output format.
Pull Request: https://projects.blender.org/blender/blender/pulls/138615
Currently the script only includes commits to "main" branches. Much of
people's work is done in branches however, so developers either
customized the script, or they would manually go through own history to
find remaining commits (I asked some).
Even if some people prefer to only list main branch commits, it's much
easier to simply remove some commits than to add missing ones.
To be able to tell which branch a commit was on, this adds a " on
`branch-name`" after the commit hash. Personally I don't find this
optimal, I'd rather group commits under their repository/branch,
proposed separately in #138615.
Pull Request: https://projects.blender.org/blender/blender/pulls/138612
Currently, when using link-drag-search and searching for "Value" shows the bake
node first. This is annoying because it's rarely what one means. It's shown
first because it's a shorter search entry.
This patch reduces the weight of that entry so that the Value node shows up
first.
Pull Request: https://projects.blender.org/blender/blender/pulls/139156
