Part of #118145.
There is some complexity in this area because the normals need to be
updated on the original geometry only when there is no deformation
or multires modifier. The simplest way to encapsulate that usage of
the original geometry for now was adding a separate function that
contains the lookup with a comment justifying it.
Started happening with 422dd9404f that introduced multi-threaded
conversions of src->dst (usually RGBA->YUV) format before encoding
the frame with ffmpeg. But the issue itself is not related to
multi-threading, but rather with the fact that AVFrame objects
started to be backed by an AVBuffer object (as that is needed for
threaded swscale to work).
Turns out, if a frame is backed by AVBuffer object, said buffer
might become "non writable" because it got shared (non-1 refcount).
And that happens with some ffmpeg video codecs, particularly PNG one.
Make sure to make the AVFrame objects writable inside
generate_video_frame. This follows official ffmpeg example
(doc/examples/encode_video.c) that explains why that is needed:
"the codec may have kept a reference to the frame in its internal
structures, that makes the frame unwritable. av_frame_make_writable()
checks that and allocates a new buffer for the frame only if necessary"
Pull Request: https://projects.blender.org/blender/blender/pulls/126317
Retrieve the relevant data directly from the original object's mesh
(or the evaluated SubdivCCG in that case) rather than using the
BVH tree's geometry points which we'd like to remove.
Pull Request: https://projects.blender.org/blender/blender/pulls/126284
No functional changes expected.
The function `BKE_action_get_item_transform_flags` was only ever used in a single place.
As such it could be a static function within the given file.
This reduction in scope will help with future refactors.
This is in service of fixing #126125 since the function in its current form
does not support layered actions.
Pull Request: https://projects.blender.org/blender/blender/pulls/126287
This avoids an issue in clang-cl on Windows ARM64 where an invalid pointer would be (re)used.
See #124182 for more details, but basically without this, when the vector of tiles inside changeset is resized, a reference to an invalid pointer (ie, the location in the old vector) is kept somewhere, which makes the test fail.
Pull Request: https://projects.blender.org/blender/blender/pulls/126083
The problem was the automatic instance deduplication. There were two
instance references before baking, both of which were referenced by instances.
When loading the bake, they were deduplicated, so there was only one geometry,
but the instances still referenced two.
The fix is to not do deduplication when loading instances from a bake.
* Adds two helper methods, the first replacing direct access to the
PBVHVertRef i value, the second replacing the previously removed
SCULPT_active_vertex_co_get
* Removes most trivial usage of active_vert_ref(), the remaining cases
are a bit more complex and need separate PRs
Pull Request: https://projects.blender.org/blender/blender/pulls/126186
This adds the ability to customize the default width of a group node that's
created for a node group. This feature works towards the goal of unifying the
features available to built-in nodes and node groups. We often customize the
width of built-in nodes from them to looks slightly better (e.g. to avoid
cut-off labels).
Pull Request: https://projects.blender.org/blender/blender/pulls/126054
The preview is only supported with deform modifiers which are
only supported for Mesh sculpting, so just write that implementation.
Also remove SculptSession::vert_positions which is now unnecessary.
Part of #118145.
Remove some uses of the `SculptSession` array in favor of retrieving
positions as necessary. Because of implicit sharing, avoiding unnecessary
mutable references to data can avoid copies and increased memory usage.
There are still more places to change before sculpt mode recieves that
benefit though.
Previously, values for `ID.flag` and `ID.tag` used the prefixes `LIB_` and
`LIB_TAG` respectively. This was somewhat confusing because it's not really
related to libraries in general. This patch changes the prefix to `ID_FLAG_` and
`ID_TAG_`. This makes it more obvious what they correspond to, simplifying code.
Pull Request: https://projects.blender.org/blender/blender/pulls/125811
When appending e.g. an object, it's generally expected that it is not just
appended but also added to the active collection. Something similar happens for
adding collections, or just geometry like a mesh data-block. This is called
"loose data instantiation" in the source code.
Which data is instantiated and how depends on context. Generally, it's best to
first do the low-level link or append operation and only afterwards do the
instantiation at a higher level where more context is known. For example, when
dragging an object asset into the 3D view, it's expected that the object is
added to the active collection. When dragging a node group into the node editor
that uses an object internally (e.g. as mesh storage), not so much.
This patch adds a new `BKE_blendfile_link_append_instantiate_loose` method that
does the "default instantiation", i.e. the one we want to happen when using the
general link/append operator to import data. Instead of calling this directly
from low level `BKE_blendfile_append`, it's now called in `wm_link_append_exec`
(and other places where desired).
Furthermore, `view3d_ob_drop_copy_external_asset` does not use this anymore, but
explicitly adds only the dragged object to the active collection. Some places,
e.g. when linking/appending using the Python API, don't do any instantiation at
all. They didn't do this before either, but it was way less obvious (it used
`BLO_library_link_params_init` instead of
`BLO_library_link_params_init_with_context`). Now there is always an explicit
call to do instantiation in higher level code.
Pull Request: https://projects.blender.org/blender/blender/pulls/125814
This is getting in the way of refactoring to the BVH. It hasn't been
working for over a year and we want to reevaluate the approach
before committing to this anyway.
When (de)selecting the active material, use the active slot
falling back to the first-used slot (for non-active objects).
Resolve regression in [0].
Ref !125948
[0]: 296d05060d
Part of #118145.
- Specialize the algorithm for each geometry type.
- Store the cache in an array instead of an attribute. Attributes are
meant to store user-edited data, not derived data like caches.
- Calculate the island IDs with a parallel disjoint set instead of a
flood-fill algorithm.
- Avoid storing the island ID array if there is only one island, saving
1 byte per vertex.
Pull Request: https://projects.blender.org/blender/blender/pulls/125907
The RNA path that is generated for Editor properties were mostly
incomplete (e.g. the viewport overlay settings).
- in python tooltips
- from the `Copy (Full) Data Path` operator
- python methods `path_from_id`, `path_resolve` returned incomplete paths
Since a space (editor) is ultimately owned by the screen, we now add the
missing "areas[x].spaces[x]" subpath to any editor.
Nested structs (like the viewport overlay) also need to include this
subpath.
Some editor related structs are not tied to a single space though (and
these cases are therefor not resolved yet):
- Dopesheet is referenced from SpaceGraph, SpaceAction and SpaceNla
- View3DShading is referenced from SpaceView3D but render engines as
well
- FileSelectParams / FileAssetSelectParams / FileAssetSelectIDFilter
need more investigation
NOTE: in case of the VSE, to make this work this also changes the
overlays to be tied to SpaceSeq (the only editor using them)
NOTE: since the above is now in place, adding VSE overlay props to Quick
favourites is now made possible as well (was a leftover from !116604)
Fixes#124527, #113489
Pull Request: https://projects.blender.org/blender/blender/pulls/125365
Together with changes from upstream audaspace, this should fix the
issue. The change slightly increases the duration of the scrubbing
sound to ensure that the audio buffer size gets played back at least
before it stops playback.
The File Output node doesn't provide an option to save byte formats like
PNG in a space that is not sRGB. This is problematic for data images
like normal maps, which need to be saved as non-color.
This patch adds a Color Space option to the File Output node to allows
users to override the assumed color space. This also adds a new global
Save As Render option that is used if Use Node Format is enabled.
Pull Request: https://projects.blender.org/blender/blender/pulls/124238
PR #125332 made changes to the format of the version string displayed
on the title bar and in status bar. These changes included not showing
the patch level if zero. Unfortunately this also changes the version
as displayed on the command-line with "--version". This particular
change was approved quite hastily so this PR just reverses that, always
shows zero patch level and therefore shows the command-line version as
before. If we want to hide patch zero we can debate this specifically
later.
Pull Request: https://projects.blender.org/blender/blender/pulls/125788
Part of #118145.
Instead of storing a separately allocated array for each BVH node
for the temporary "color buffer" colors meant for mixing during a
stroke, just store an array the size of the whole mesh. Though this
is wasteful in terms of memory usage, plenty of other brushes store
mesh-sized arrays already, and it should make more sense as BVH
nodes get smaller too. After this commit, the BVH tree has no
specific code for color attributes anymore.
