We don't use the callbacks that create virtual arrays from the custom data
anymore, they just add extra indirection. The only non-obvious case was
the crease attribute which had a setter function. Replace that with an
attribute validator like the other similar attributes.
Pull Request: https://projects.blender.org/blender/blender/pulls/107088
While embedded IDs are usually considered as private local data of their
owner ID, some areas of code, like the depsgraph, can consider them as
regular IDs in some aspects.
So when swapping IDs, also properly 'counter-swap' their potential
embedded IDs, such that the pointers to the embedded IDs remain as before
swapping, even though the data of the embedded IDs is swapped.
The main target of this change is memfile undo code. There, newly read
IDs are swapped with their oldder version, so that the old address
contains the new data. This allows to avoid rebuilding some of the
depsgraph. Doing the same thing for embedded IDs should reduce even
further the needs for depsgrah rebuilds on undo steps.
This commit also gives more control over the remapping of 'self' ID
pointers inside themselves.
Pull Request: https://projects.blender.org/blender/blender/pulls/107044
This is the first step for refactoring the lightcache system.
Each probe instance (as in `Object`) will now store its own baked data.
The data is currently stored in uncompressed readable format.
This introduces two new operators for baking to avoid confusion with
the previous light baking pipeline. These do nothing other than
creating empty caches that will be populated by EEVEE later on.
The DNA storage is made to be able to include multiple caches
in case of baked simulation over time but it isn't yet supported.
I prefer to keep the implementation simple for now as the long term
goals for this feature are uncertain.
There is still a type flag (`LightProbeObjectCache.cache_type`) that
will be used for versioning.
The naming convention of structs is a bit weird but that's all I
found in order to avoid interfering with the old scene light cache
that is still used by (old) EEVEE.
Related task #106449.
Pull Request: https://projects.blender.org/blender/blender/pulls/106808
This should not have much effective consequences with current code, but
fixes potential missed remappings for e.g. some nodetree pointers in the
node editor, or the `parent` pointer of collections to a scene's master
collection.
Implements #95966, as the final step of #95965.
This commit changes the storage of mesh edge vertex indices from the
`MEdge` type to the generic `int2` attribute type. This follows the
general design for geometry and the attribute system, where the data
storage type and the usage semantics are separated.
The main benefit of the change is reduced memory usage-- the
requirements of storing mesh edges is reduced by 1/3. For example,
this saves 8MB on a 1 million vertex grid. This also gives performance
benefits to any memory-bound mesh processing algorithm that uses edges.
Another benefit is that all of the edge's vertex indices are
contiguous. In a few cases, it's helpful to process all of them as
`Span<int>` rather than `Span<int2>`. Similarly, the type is more
likely to match a generic format used by a library, or code that
shouldn't know about specific Blender `Mesh` types.
Various Notes:
- The `.edge_verts` name is used to reflect a mapping between domains,
similar to `.corner_verts`, etc. The period means that it the data
shouldn't change arbitrarily by the user or procedural operations.
- `edge[0]` is now used instead of `edge.v1`
- Signed integers are used instead of unsigned to reduce the mixing
of signed-ness, which can be error prone.
- All of the previously used core mesh data types (`MVert`, `MEdge`,
`MLoop`, `MPoly` are now deprecated. Only generic types are used).
- The `vec2i` DNA type is used in the few C files where necessary.
Pull Request: https://projects.blender.org/blender/blender/pulls/106638
Animation data (including drivers) on inactive view layers now work. The
removal of such view layers was too optimistic; they are now kept
around. The bases are still removed, mostly for safety sake and to keep
the changes to a minimum.
`scene_remove_unused_view_layers()` has been renamed to
`scene_minimize_unused_view_layers()` to reflect its new functionality.
For compatibility with assumptions in other areas of the code, the
function still ensures the input view layer is at index 0.
This also introduces a new function
`BKE_view_layer_free_object_content(view_layer)`, which is a subset of
the functionality of `BKE_view_layer_free()`.
Bezier curve position evaluation expects the handle position attributes
to exist and doesn't handle the case where they don't. Swith to using
a utility function to evaluate each curve type so Bezier evaluation can
stop early in that case.
Default settings changed for Sculpt mode's
Paint Brush, Smear Brush, and Smear Brush.
~~This includes updates of PR review #105691.~~
Updated to only include commits relevant to this PR.
Reviewed by: Joseph Eagar & Julian Kaspar
Pull Request: #105759
commit 7aa5e65dcbda862dcb17ecfc6727eb241a12c316
Merge: c08a9ec19f 7c9e493da55
Author: Joseph Eagar <joeedh@gmail.com>
Date: Sun Apr 16 15:11:53 2023 -0700
Merge branch 'main' of https://projects.blender.org/ChengduLittleA/blender into ChengduLittleA-main
commit 7c9e493da55a4adbfa2415b711e6d0daa2720ad9
Author: YimingWu <xp8110@outlook.com>
Date: Fri Mar 31 17:46:32 2023 +0800
Fix#106358: Handles null evaluated object when entering sculpting workspace.
The setup where everything in the scene is invisible/not enabled could
trigger a crash when switched to sculpting workspace, triggered when
opening the file.
This patch handles such situation.
Decompose most of the nesting in the code to make the code
more consistent along the line of program execution.
Mainly achieved through:
- Remove redundant else
- Invert condition
- Add temporary variable to redistribute and name conditions
Pull Request: https://projects.blender.org/blender/blender/pulls/105509
Cleaned up sculpt code to store lists of `PBVHNodes` with
`blender::Vector` instead of simple pointer arrays. This is much
simpler and eliminates memory leaks caused by forgetting to free
the result of `BKE_pbvh_search_gather`.
Notes:
* `BKE_pbvh_search_gather` is now `blender::pbvh::search_gather`.
* `FilterCache` and `ExpandCache` have ownership over their .nodes
members; as a result they're no longer pure C structs and
are allocated with `MEM_new`/`MEM_delete`.
* The word 'totnode' no longer occurs anywhere in
`source/blender/editors/sculpt_paint`
Todo (not for this PR): create a new properly C++ task API for sculpt
(with lambdas) and use it for brushes.
Pull Request: https://projects.blender.org/blender/blender/pulls/106884
Embedded IDs did not benefit from any of the recent optimizations
(especially for undo case) when writing regular IDs (cleaning up of some
pure runtime data that would generate a lot of fake 'changed on undo'
status).
Now factor out of `write_file_handle` this part of the code generating
temp ID copy with cleaned-up data for writing, and expose it in BLO API
such that IDs owning embedded ones can also use it.
Code there was fairly naive and simple, missing some ID pointers,
sometimes improperly accessing non-ID data as IDs (usual dear Outliner
tree element usages of its 'ID' pointer...).
And code was especially quite severely broken in case these UI ID
usages were processed in a non-readonly context (i.e. if some of these
ID pointers were expected to be modified).
Code has been updated following existing very similar code in
`lib_link_workspace_layout_restore` from `readfile.cc`.
Similar to 7eee378ecc, this change decreases memory usage and
improves performance when copying curves and meshes without changing
their topology. The same change used for custom data layers is applied
to face and curve offset indices, which aren't stored as a custom data
layer.
The implicit sharing info for the offsets is stored in the mesh and
curve runtime structs, since it doesn't need to be written to files
directly. When changing the offsets pointer directly, the sharing info
must be updated accordingly. To make that easier, a few utility
functions take care of common operations like making an array mutable,
resizing an array, and creating sharing info for allocated data.
This commit also clarifies the intention to not allocate the offsets
at all when there are no curves/faces. That slightly complicates some
of the logic, but there's no reason for the single `0` integer to be
allocated.
Pull Request: https://projects.blender.org/blender/blender/pulls/106907
This commits adds some new, specific flags to further control ID
remapping process (like and option to skip user refcounting completely).
It also adds a new function to do 'raw' remapping, without any extra
post-processing, depsgraph tagging, etc. This is not used currently, but
will soon be needed by readfile post-processing code changes.
There is also some small cleanups and reorganization in that area of code,
the main noticeable change being the switch from a short to an int for
the flags controlling remapping code (using short here does not give
any benefit, and makes it harder to switch to integers when it becomes
necessary).
No change in behaviors are expected from this commit.
This type will be used to store mesh edges in #106638, but it could
be used for anything else too. This commit adds support for:
- The new type in the Python API
- Editing the type in the edit mode "Attribute Set" operator
- Rendering the type in EEVEE and Cycles for all geometry types
- Geometry nodes attribute interpolation and mixing
- Viewing the type in the spreadsheet and using row filters
The attribute uses the `blender::int2` type in most code, and
the `vec2i` DNA type in C code when necessary. The enum names
are based on `INT32_2D` for consistency with `INT8` and `INT32`.
Pull Request: https://projects.blender.org/blender/blender/pulls/106677
This is the case e.g. of the `parent` collection pointer of collections
children of a scene's master collection, or some nodetree pointers in
the UI data (node editor).
Right now handling of this new flag is exactly the same as in owning
embedded case, the distiction between both usages will happen in future
commits.
This commit is expected to have no behavioral change at all.
Similar to e35f971da1. We aren't meant to rely on the
zero-initiatialization on creation anyway. I observed a small (a few
percent) decrease in minimum runtime in the geometry nodes points node.
The position attribute has special meaning for point clouds, and
meshes and curves have access methods for the attribute as well.
This saves boilerplate and gives more consistency between types.
This integrates the new implicit-sharing system (from fbcddfcd68)
with `CustomData`. Now the potentially long arrays referenced by custom
data layers can be shared between different systems but most importantly
between different geometries. This makes e.g. copying a mesh much cheaper
because none of the attributes has to be copied. Only when an attribute
is modified does it have to be copied.
Also see the original design task: #95845.
This reduces memory and improves performance by avoiding unnecessary
data copies. For example, the used memory after loading a highly
subdivided mesh is reduced from 2.4GB to 1.79GB. This is about 25%
less which is the expected amount because in `main` there are 4 copies
of the data:
1. The original data which is allocated when the file is loaded.
2. The copy for the depsgraph allocated during depsgraph evaluation.
3. The copy for the undo system allocated when the first undo step is
created right after loading the file.
4. GPU buffers allocated for drawing.
This patch only gets rid of copy number 2 for the depsgraph. In theory
the other copies can be removed as part of follow up PRs as well though.
-----
The patch has three main components:
* Slightly modified `CustomData` API to make it work better with implicit
sharing:
* `CD_REFERENCE` and `CD_DUPLICATE` have been removed because they are
meaningless when implicit-sharing is used.
* `CD_ASSIGN` has been removed as well because it's not an allocation
type anyway. The functionality of using existing arrays as custom
data layers has not been removed though.
* This can still be done with `CustomData_add_layer_with_data` which
also has a new argument that allows passing in information about
whether the array is shared.
* `CD_FLAG_NOFREE` has been removed because it's no longer necessary. It
only existed because of `CD_REFERENCE`.
* `CustomData_copy` and `CustomData_merge` have been split up into a
functions that do copy the actual attribute values and those that do
not. The latter functions now have the `_layout` suffix
(e.g. `CustomData_copy_layout`).
* Changes in `customdata.cc` to make it actually use implicit-sharing.
* Changes in various other files to adapt to the changes in `BKE_customdata.h`.
Pull Request: https://projects.blender.org/blender/blender/pulls/106228
I cleaned up the cube brush tip code quite a bit; more remains
to be done. There is a new function to initialize cube
tip matrices, SCULPT_cube_tip_init. It's currently only
used by the paint brush, I'll need to do a bit of testing
before using it for clay strips and multiplane scrape.
Note: SCULPT_cube_tip_init uses the brush local matrix code
to avoid code duplication (and to take advantage of the debouncing
that is done there).
This data should be initialized when building the mesh. If the arrays
aren't already initialized, there is already a bug. Theoretically this
could improve performance, since we can use `malloc` instead of
`calloc`. In practice I observed less than a 1% difference though.
