- Use `StringRef` over `StringRefNull`, since it handles `nullptr` in
its constructor and gives more flexibility about the string's source.
- Change `const char *` to `StringRef` as well, for improved ergonomics
and efficiency.
- Move `add_interface_socket_from_node` to a .cc file. It's not small
or performance sensitive enough to be inlined.
- Avoid allocating empty strings for item descriptions.
- Allocate with `BLI_strdupn` to avoid unnecessary `strlen` calls.
- Move a bit more code into a proper namespace.
Pull Request: https://projects.blender.org/blender/blender/pulls/112721
`Restore item` button is not resetting `active` property when keyitem is registered
as `inctive` by default. To fix this, clear `inactive` flag then use default keyitem's
flag to correctly restore the keyitem.
Pull Request: https://projects.blender.org/blender/blender/pulls/112497
Support the property settings that were implemented some time ago.
As an aside, change the defaults for min and max to integers: integer
values can be implicitly cast to float, but the reverse results in an
error. Also, pointer properties don't support a default value.
Pull Request: https://projects.blender.org/blender/blender/pulls/112712
AgX and Filmic have different looks in the configuration, but the looks
share the same UI name and purpose. So we can switch to the
corresponding look when switching view transforms.
Thanks to Yiming Wu for helping track down the problem.
Pull Request: https://projects.blender.org/blender/blender/pulls/112686
The goal is to make the evaluation of repeat zones more efficient by making
use of the lazy-function evaluation system. Performance is improved in two ways:
* Unnecessary nodes are not evaluated anymore. E.g. if a repeat zone outputs two
geometries but only one of those is actually used, the other one will not be
computed anymore.
* Support evaluating different iteration indices at the same time on different threads.
It is possible that some uses of repeat zones become slower with this refactor,
especially when each iteration does very little work and there are a lot of iterations.
The old implementation was not optimized for this use case either but now there
is a bit more overhead constant overhead per iteration than before.
On the bright side, this change can result in some very significant speedups when
some computations can be skipped. See #112421 for some example setups.
There is one todo comment for adding back-links for socket usages. Properly linking
those up can result in better (shorter) life-times for anonymous attributes. Even without
that, performance is already better than before.
The implementation reuses the existing lazy-function graph system for the repeat
zone, by building a dynamically sized graph based on the number of iterations.
Building a lazy-function graph makes it possible to use the lazyness and multi-threading
features of the lazy-function graph executor. This is much easier than reimplementing
this behavior for repeat zones specifically (hence there was only single-threaded eager
execution before).
Pull Request: https://projects.blender.org/blender/blender/pulls/112421
Without this it was necessary to convert pasted text into spaces
for them to display properly.
Since tabs are a valid part of a string, it's incorrect to assume
tabs can always be expanded to spaces.
Pasted text and console output now display tabs properly,
the tab still uses spaces for indenting though.
Adds the ability to set text cursor position in the Console editor
by clicking with the mouse and many other expected operations such as
cut, select all & deselect on text entry.
Ref !108626
Using SDL's initialization logic, this is mainly a change for XDG
as LIBDECOR already required a configure event before accessing
the underlying XDG window.
While I didn't notice functional changes with this change window
flickering on startup remains an issue with some compositors
(KDE & river). Debugging these issues is simpler when both windowing
decoration systems work in a similar way & window configuration
is guaranteed to have run before the window is returned.
(via #xdg_surface_ack_configure).
Small style and layout tweaks of some tabs in Preferences.
Mainly inline styling for compact layouts, grouping labels and
adding icons where needed.
Purposely leaving aside complex areas such as Themes,
Add-ons, or Keymap. They require larger design decisions.
Pull Request: https://projects.blender.org/blender/blender/pulls/112393
See #103343.
We had attempted to move the file to C++ twice before but found failing
tests each time. To find the code with changed behavior between C and
C++, I temporarily added a header declaring all the functions and moved
them one by one. The problem ended up being one function's use of `exp`
which used floats instead of doubles in C++.
Pull Request: https://projects.blender.org/blender/blender/pulls/112693
The workaround for shader draw parameters isn't working for OpenGL. The
reason is that it isn't looking for the correct attribute to attach the
buffer to.
This fixes the issue that all objects where drawn using the unit matrix when
starting blender with `--debug-gpu-force-workarounds`.
Fixes#112641
Pull Request: https://projects.blender.org/blender/blender/pulls/112668
With a callback to node panels similar to the one for nodes. Used in the
Principled BSDF to place enums in the relevant panels.
Longer term we want to make enums into sockets, but even then there are
still potentially some types of buttons we want to have in panels.
Pull Request: https://projects.blender.org/blender/blender/pulls/112591
The layout for this is not great but consistent with existing
collapse/expand functionality there. There are ideas to replace this
entire material properties drawing longer term, but for 4.0 it's
important to show the panels in some way.
Pull Request: https://projects.blender.org/blender/blender/pulls/112591
Shortening names based on context is something we do in the user
interface throughout Blender. This is a simple automatic mechanism to do
it for node panels as well, which significantly improves the Principled
BSDF layout.
We want the properties and sockets to be identifiable outside of the
panel contextcontent, for example in the animation editor, API docs, or
when exposing a single socket as a group input. So making the actual
name shorter is not an option.
Pull Request: https://projects.blender.org/blender/blender/pulls/112591
Default to displaying nearly all nodes when the engine is not set to
Cycles or Eeeve. And instead hide nodes not supported in Eevee only
when Eevee is selected.
This way external engines can use the same menu, particularly useful
for upcoming support for conversion to MaterialX.
Only the Shader category was incomplete like this, so it can also be
seen as fixing an inconsistency with other categories.
Replace the import/export of armature layers with bone collections.
The Old Way:
- Export: Each bone would store which armature layer it was on.
- Import: All armature layers that contain at least a bone are shown.
The New Way:
- Export: Each armature contains a list of its bone collections,
including which one is active and which ones are visible.
- Export: Each bone stores which bone collection it is on.
- Import: the above data is simply used as-is.
Due to limitations of the current Collada importer code, each "extra"
tag can only occur once per Collada node. This means that it was
impossible to write a `<collection name="Bones">` tag for each bone
collection, as only one of those would actually be stored by the
importer for further processing. To work around this limitation, all
bone collection related tags store their values as newline-separated
strings. Example:
```
<node id="Armature">
<extra>
<technique profile="blender">
<collections sid="collections" type="string">Layer 1
Layer 3b
Group
That One Bone</collections>
<visible_collections sid="visible_collections" type="string">Layer 1
Layer 3b
That One Bone</visible_collections>
<active_collection sid="active_collection" type="string">That One Bone</active_collection>
</technique>
</extra>
</node>
```
Instead of adding the default "Bones" bone collection to every Armature
data-block, only add it for Armatures that are added in the 3D View with
Shift+A > Armature.
This simplifies things like importers, which otherwise would have needed
to remove that default collection before creating the imported ones.
Additionally, this also ensures that the one bone you get by default is
actually assigned to the default bone collection.
Move the 'set active bone collection by name' functionality from RNA
to the ANIM module, so that it can be used in other places.
No functional changes.
