Age | Commit message (Collapse) | Author |
|
|
|
|
|
Patch by Valentin (Poulpator)
Differential Revision: https://developer.blender.org/D5468
|
|
Differential Revision: https://developer.blender.org/D5369
|
|
This removes the recursive conversion of material using old blend modes.
With the approval of @brecht
|
|
We were storing the 'item' part in the RNA path of the override op
itself, which will not work with IDs because overriding local one might
not have the same name as the linked one (when some local ID with same
name already exists).
Now we are properly handling this using the expected actual override
operation, which stores necessary data (names or indices) from both
local and linked data.
|
|
|
|
This is useless (as diffing on `objects` + `children` shall be enough),
and potentially very time consuming in case of heavy hierarchy of
collections.
|
|
This checkbox replaces the "Disabled" and "Enabled" entries in the
filter drop-down. As a result, it now takes a single click to limit the
shown entries to enabled add-ons only. This is also an actual flag in
the preferences, and thus its state is saved between runs on Blender (in
contrast to the filter, which is always reset to "All").
Reviewed by: brecht, billreynish
|
|
|
|
This fixes the glitching hairs described in T56408, T63534, and possibly
also T63534.
The fix consists of returning the original mesh (i.e. as visible in edit
mode) when constructing the ORCO mesh. This allows a static set of
coordinates to be used when computing the child hair positions.
The original mesh is only returned when it has the same topology (at
least same number of vertices, loops, and polys. It's up the author of
the Alembic file to ensure stable geometry when it's desired to be
compatible with Blender's hair system.
Reviewers: mont29, brecht
Differential Revision: https://developer.blender.org/D5492
|
|
The `BKE_material_pop_id()` and `BKE_material_clear_id()` functions had
a parameter `update_data` that, when `false`, would cause the mesh polys
to keep their material index, even when the indexed material slots were
removed. This behaviour was never used in the C code and not supported
by the drawing code, making polygons disappear and causing crashes. The
Python binding in RNA, however, defaulted to `update_data=False`.
This commit removes the `update_data` parameter altogether, and makes
the functions always fix up the material indices.
Reviewed by: mont29, brecht
|
|
There could be a more direct way to initialize a gizmo in a view,
for now this resolves the crash.
|
|
This was a mistake in rB87629b2a7443
`BKE_object_shapekey_free` would never return `true`, so DEG updates
would not happen...
So `ob->shapenr` was not up-to-date etc., leading to crash in
`BKE_mesh_nomain_to_mesh` shapekey handling...
Reviewers: brecht
Maniphest Tasks: T68710
Differential Revision: https://developer.blender.org/D5501
|
|
|
|
While speedup is non-linear, it gives ~30% speedup for ~6 million verts.
D3993 by @Al with edits.
|
|
This is 2.7x behavior, while there are plans to improve on this,
committing in case larger changes take longer.
Without this it's not easy to select object data without changing modes.
See D5493
|
|
Avoids calculating the visible part of a region whenever
on-screen overlays are drawn.
|
|
|
|
|
|
Issue introduced (more like exposed) in b7f86ff72273.
|
|
Complex nodetrees could fire infinite recursion with previous algo.
Now using another gset we make sure we can only evaluate a tree once.
|
|
|
|
|
|
`'ParticleSystem': named type definition in parentheses`
And prevent the need for struct `Object` to be defined.
|
|
|
|
objects inside the rect
But in the future the selection code may also be used in object mode (eg for snapping).
So to avoid using too much VRAM resources, it is good to avoid drawing all objects in the viewport.
The solution was to create an array with only objects that are detected within the selection area.
If the selection operator is modal, objects already detected are not removed from the array until view3d is moved or orbited.
To detect the object, its BoundBox is tested.
Since the Select Engine does not have a dedicated depth texture, whenever a new object is "found" the depth of the objects in the array already drawn is redrawn.
Reviewers: campbellbarton, fclem
Reviewed By: fclem
Differential Revision: https://developer.blender.org/D5435
|
|
For testing intersection with frustrum planes without having to transform all bound box vertices into global space.
|
|
Originally D5020 by @deadpin, refactored to make the change simpler.
|
|
- When no change is performed on a float button, cancel the action.
- Move left/right clicks into the same block.
- Replace ambiguous names: temp, tempf.
|
|
Forgot to properly maintain the edge for faces while
dissolving edges.
|
|
|
|
center of the world
|
|
|
|
As the video editor mode is not Render mode the VFX was omitted.
Now, the mode is only checked for View3D.
|
|
In rB9c010c44f420, DRW_mesh_batch_cache_create_requested was changed to
take a scene (instead of toolsettings directly), but
DRW_draw_depth_object is calling this with with NULL, just checking for
this seems to fix...
Reviewers: fclem
Maniphest Tasks: T68675
Differential Revision: https://developer.blender.org/D5488
|
|
|
|
|
|
Regression from 6148ed8cf9ca
|
|
Missed in recent commit.
|
|
|
|
Just enable some commented-out code from rB636289b755f6ce (disabled at
that time because we were too close of 2.79 release...).
Issue raised in D5486, thanks.
|
|
|
|
Reminder not to leave in unused code, or comment why it should be kept.
|
|
|
|
Differential Revision: https://developer.blender.org/D5484
|
|
|
|
This reduces the risks of hash collision while maintaining a small number
of character per attrib.
|
|
This is done because they don't have the same update frequency. UV can be
persistent even on geometry update (ex: skinned object) but tangents
can change if the normals change.
Also the name buffer per vbo was too small to contain all names.
|
|
This remove code duplication and use base63 encoding of the hash.
Use mumur hash to have more randomness.
|