We need a separate array that we can change in during the parallel
group construction. That array tells where in each group the index
is added. Building this array is expensive, since construcing a new
`Array` fills its elements serially. There are two possible solutions:
1. Use a copy of the offsets to increment result indices directly
2. Rely on OS-optimized `calloc` instead of `malloc` and a copy/fill
Both depend on using `fetch_and_add` instead of `add_and_fetch`.
The vertex to corner and edge to corner map creation is optimized
by this commit, though the benefits will be useful elsewhere in the
future.
| | Before | 1. offsets copy | 2. calloc |
| -------- | ------- | --------------- | --------------- |
| Grid 1m | 3.1 ms | 1.9 ms (1.63x) | 1.8 ms (1.72x) |
| Grid 16m | 51.8 ms | 33.3 ms (1.55x) | 32.7 ms (1.58x) |
This commit implements the calloc solution, since it's slightly faster
and simpler. In the future, `Array` could do this optimization itself
when it detects that its fill value is just zero bytes.
Pull Request: https://projects.blender.org/blender/blender/pulls/112065
Assert that the partial is part of the particle system before
creating an index. This is a hint to developers that the cast
is safe and using a pointer offset isn't going to cause problems.
Issue raised by !111193.
Co-authored-by: Loren Osborn <linux_dr>
When using a pointer offset that may be out-of-bounds the integer
must not truncate the offset.
Issue raised by !111193.
Co-authored-by: Loren Osborn <linux_dr>
#111972 made sure copied sockets have unique identifiers, but it didn't
cover the case of copying a panel and its children.
- Added an optional `uid_gen` argument to item copy functions which
generates new unique identifiers for copied items. If not specified
the items will retain the original identifiers (e.g. when copying an
entire node tree).
- Removed the `copy_items` panel function from the API. Only used
internally and requires a uid generator now.
Pull Request: https://projects.blender.org/blender/blender/pulls/112074
Batch deletion of IDs could lead to deleting ObData ones while keeping
the related ShapeKeys.
Orphaned shape keys are not allowed anymore in Blender, they are checked
against in both file write and read code.
This would lead to assert (and crash) e.g. in the liboverride code.
This commit forcefully add shapekeys of deleted meshes, curves etc.,
when calling e.g. `BKE_id_multi_tagged_delete`.
Note that deleting the shapekey when deleting the obdata ID was already
implemented in single ID deletion (`BKE_id_delete` & co), in the
underlying private `id_free`. But this is skipped in `no main` case.
which is used by the batch deletion code for performance optimizations.
The basic idea is very simple. Whenever a supported menu is open, one can just
start typing and this opens a search that contains all the (nested) menu entries.
The main downside is that this collides with accelerator keys. Those are the
underlined characters in each menu. For now, we just enable this new searching
behavior in a few selected menus: Node Add Menu, View3D Add Menu and
Modifier Add Menu.
This new functionality can be enabled for a menu by setting
`bl_options = {'SEARCH_ON_KEY_PRESS'}` to true in the menu type.
The status bar shows `Type to search...` when a menu is opened that supports search.
Pull Request: https://projects.blender.org/blender/blender/pulls/110855
When data-block/asset previews were not stored in the thumbnail cache
yet (or were outdated), we'd read them from .blend files. This could
lead to random crashes (but quite reliable with a small number of
previews to be read).
Wasn't clearing runtime memory which could lead to the
`PRV_TAG_DEFFERED` bit being set. This meant we would try to free
deferred preview data since eefee47a8a, which was just garbage memory.
The `Library.tag` data is supposed to be runtime only, so needs to be
cleared when reading data from blendfile.
This fixes the 'need resync' status of a library staying around forever
once set, annoying visual warning in the Outliner.
Change the existing "Is Shade Smooth" node to be named "Is Face Smooth"
and add a new "Is Edge Smooth" node. Also give the "Set Shade Smooth"
node the ability to set face or edge smoothness.
The fact that the nodes process "smooth" data reversed from the builtin
"sharp" attributes can be reversed with versioning in a separate commit.
While it's tempting to abstract the sharpness status into a single node,
face and edge smoothness are accessed separately in edit mode, and the
subtlety of interacting with data on different domains would make that
confusing. Instead, a separate "Is Shade Smooth" node group asset will
give all the sharp elements taking into account both builtin attributes.
The fact that sharpness is stored separately on two domains makes the
best design for simple operations non-obvious. For example, you should be
able to remove all sharpness or make everything flat with a single node.
The behavior depends on whether the two attributes exist and the
combination of values between the domains.
---
Pull Request: https://projects.blender.org/blender/blender/pulls/112029
This is a nasty gathering of several issues, main one being that 'local
collections' of 3DViews are still updated immediately instead of the
deferred update used for all other viewlayer cases.
The change in a16bcb6576 led to internal references to the Scene's
master collection to become invalid, which is the expected behavior.
But this turns the Scene's view_layers into invalid state too.
Ideally, there should never be resync of viewlayers of a scene being
deleted anyways.
For now, take the (hopefully!) safe approach of explicitely forbidding
any viewlayer update during ID deletion process, and deferring it at the
end of the process.
Note that this change may also give some marginal gerformance
improvements in some rare edge cases (like deleting a very heavy scene
with many collections and 'local collection' 3DViews ?).
The handle selection should be ignored when hidden.
Now set handle type considers all handles selected when handles
are hidden and the knot is selected.
Replace use_handles with an enum that optionally uses handles
except when the vertex (knot) is selected in which case it behaves
as if both handles are selected.
Needed for nurb curves not to change handle type when only the
center point is selected (as is done in the graph editor).
No functional changes.
There are a couple of functions that create rna pointers. For example
`RNA_main_pointer_create` and `RNA_pointer_create`. Currently, those
take an output parameter `r_ptr` as last argument. This patch changes
it so that the functions actually return a` PointerRNA` instead of using
the output parameters.
This has a few benefits:
* Output parameters should only be used when there is an actual benefit.
Otherwise, one should default to returning the value.
* It's simpler to use the API in the large majority of cases (note that this
patch reduces the number of lines of code).
* It allows the `PointerRNA` to be const on the call-site, if that is desired.
No performance regression has been measured in production files.
If one of these functions happened to be called in a hot loop where
there is a regression, the solution should be to use an inline function
there which allows the compiler to optimize it even better.
Pull Request: https://projects.blender.org/blender/blender/pulls/111976
This fixes a specific crash that would happen to me sometimes when changing
between files (the files may have been saved in older versions of the Blender).
The fix is just to null-check the icon lookup.
Pull Request: https://projects.blender.org/blender/blender/pulls/111670
Items in the node group interface need to have unique identifiers.
Copying an item (socket or panel) was not doing this, so the node groups
end up having sockets with the same identifier. Linking sockets was
broken because copies could not be distinguished.
Pull Request: https://projects.blender.org/blender/blender/pulls/111972
In 2.6 the old method of using bNodeSocket lists in bNodeTree directly
as group sockets was replaced with new group input/output nodes. This
required versioning to create those input/output nodes and then redirect
links to the new node sockets. Because creating nodes relies heavily on
node typeinfo this versioning was done in the `_after_linking` section
of the 2.6 versioning code, running after _all other versioning_
(including for much newer versions!) has already happended.
While typinfo is available at that point, doing such late versioning
causes severe problems when the data structure changes, as is the case
with the recent node panels patch (#111348). The new node group
interface also has versioning code for 4.0, but this runs _before_ the
`_after_linking` code for 2.6! Versioning for node panels expects
sockets in bNodeTree to not have any links pointing at them, but this is
not true for old 2.6 files which have not yet been fully versioned at
that point, because of the late versioning stage. Subsequently 2.6
`_after_linking` code crashes when trying to modify node links with
dangling pointers.
The solution here is to move the old versioning code out of the
`after_linking` stage to restore the expected versioning chain. This
requires creating nodes and node sockets without any typeinfo, but
luckily we only need to create simple known group input/output nodes
which don't have much complicated behavior.
Pull Request: https://projects.blender.org/blender/blender/pulls/111704
Previously, BKE level preview image code was in `BKE_icons.h` and `icons.hh`.
While these types are related, I always found this quite hard to navigate since
preview image stuff was just in the middle of icon functions. Plus, people
don't expect preview image functions in icon files, the relationship is not
obvious.
Instead, use focused files that make it easy to quickly navigate them
and see what they are dealing with.
Pull Request: https://projects.blender.org/blender/blender/pulls/111709
This commit disambiguates the following messages:
- Sequencer effect strip types: use "Sequence" context in relevant
places, as that is already extracted as part of the
`sequencer_prop_effect_types` enum, and more specific.
- "Language" (a natural or programming language)
- "Flat" (gender)
- "Smooth" (action or amount -- very partial disambiguation for now
because this is all over the place)
It also extracts these messages:
- Newly created Freestyle data
- LineStyle
- LineSet
- Modifiers
- "Registering panel class:" error message from RNA registration
- "Node must be run as tool" error message from tool geometry nodes
Ref #43295
Pull Request: https://projects.blender.org/blender/blender/pulls/111146
The f12e9f32b5 patch introduced a new improved method of blending
dual quaternion transformations to handle combined scale and rotation
better. However, the changes were not complete:
* The new math ignored crazyspace computations, which need to compute
a complete transformation matrix. As an optimization, the new method
avoided fully computing the scale component, so the matrix would
have no scale or shear.
* The Armature constraint is supposed to behave identically to the
modifier, and it was not updated. The constraint also requires
computing a complete matrix.
This change extracts the new math into a utility function, change
the optimization to be controlled by a parameter, and use the new
function in the constraint.
Pull Request: https://projects.blender.org/blender/blender/pulls/111759
The main goal here is to rename things in a way that makes sense for
simulation baking, but also for the upcoming bake node.
This also removes some versioning code from 3.6 which initialized the
default bake path. Baked data from back then can't be loaded anymore
anyway, and the way the default path is generated is different now as well.
Pull Request: https://projects.blender.org/blender/blender/pulls/111845
Blob stands for "binary large object" and is a known term. I used to use the term `bdata`
to mean "binary data", mainly because I didn't think of a better name. Blob is a much
better name as it captures the intend of those files much better.
This change breaks existing bakes, because I rename the folder from `bdata` to `blobs`.
I think that is ok, because I also just broke bakes two days ago in a larger refactor
(e92c59bc9b).
Pull Request: https://projects.blender.org/blender/blender/pulls/111822
The file in question does not have the default (internal) root panel
flag for allowing child panels. This was probably saved before the flag
was added. A simple versioning snippet takes care of this case.
Pull Request: https://projects.blender.org/blender/blender/pulls/111828
The hash tables and vector blenlib headers were pulling many more
headers than they actually need, including the C base math header,
our C string API header, and the StringRef header. All of this
potentially slows down compilation and polutes autocomplete
with unrelated information.
Also remove the `ListBase` constructor for `Vector`. It wasn't used
much, and making it easy to use `ListBase` isn't worth it for the
same reasons mentioned above.
It turns out a lot of files depended on indirect includes of
`BLI_string.h` and `BLI_listbase.h`, so those are fixed here.
Pull Request: https://projects.blender.org/blender/blender/pulls/111801
Panels could have parents if the parent was the root panel.
Anyway, the assert isn't really helpful here, better to keep that
just for higher level editing functions.
Mostly the issues were because drag and drop moving to a different
index in the parent didn't take into account the root panel. Also add
a null check for safety similar to b36367e663.
When entering sculpt mode, the normals cache is shared with the
evaluated mesh. `SharedCache::ensure()` un-shares the cache in order
to update it, so the pointers in the pbvh must be reset.
Also bump the recorded minimum 3.6 subversion for opening files, to
correspond to a fix in the 3.6 branch that prevents bones from seemingly
disappearing when opening files with 4.x bone collections.
