Blender Git Commit Log

GPencil: Add new X-Ray option for Canvas Grid

In some situations is good to have a grid visible anot beeing occulde by meshes.

By default is OFF.

Differential Revision: https://developer.blender.org/D7721

Revert "GPencil: Add new X-Ray option for Canvas Grid"

This reverts commit fd025b0b7a46c5e56f348b73cf989c6bbddec707.

Revert "GPencil: Change tooltip"

This reverts commit 7b688a6b04bc2033694f4742f06eef28ab023715.

GPencil: Change tooltip

GPencil: Add new X-Ray option for Canvas Grid

In some situations is good to have a grid visible anot beeing occulde by meshes.

By default is OFF.

Differential Revision: https://developer.blender.org/D7721

Merge branch 'blender-v2.83-release'

Fix T75365: memory leak when writing PNG fails due to lack of disk space

Fix wrong Volume object step size tooltip

Fix T76556: Apply parenting when joining parent into object

The current behavior isn't necessarily "incorrect," but it's unintuitive
and confusing. A simple fix is to apply parentinv before finishing the
operator.

However, there may still be issues when the object's parent has a parent.

Reviewed By: brecht

Differential Revision: https://developer.blender.org/D7723

Refactor: Move brush foreach_id to new IDTypeInfo structure.

Refactor: Move lightprobe foreach_id to new IDTypeInfo structure.

Cleanup: Keep IDTypeInfo separated from ID type specific API.

Also do not define IDTypeInfo callback when default generic code is
enough...

Refactor: Move speaker foreach_id to new IDTypeInfo structure.

Refactor: Move world foreach_id to new IDTypeInfo structure.

Cleanup: Keep IDTypeInfo separated from ID type specific API.

Also do not define IDTypeInfo callback when default generic code is
enough...

Refactor: Move shapekey foreach_id to new IDTypeInfo structure.

Refactor: Move camera foreach_id to new IDTypeInfo structure.

Cleanup: Do not mix IDTypeInfo code with rest of an ID type specific API...

Merge branch 'blender-v2.83-release'

Fix T75574: Instant crash: averaging custom normals on model.

Fairly critical code mistake actually, since it uses malloca,
BLI_smallstack should *never* be declared inside a loop...

Also optimized handling of the `loop_weight` heapsimple, we can also
only create and use a single one for all edited objects...

Found two other operators potentially affected by same issue (split
normals, and weld edges into faces).