https://builder.blender.org/admin/#/builders/57/builds/1462

Merge master

Hello! Can we add simple stuff like Loop slide to Bevel Modal Map, it doesn't have to be keymapped by default, but it is really annoying not to have the ability to toggle it with a hot key and having to always click in the bottom left corner to toggle it off and on, and when modeling hardsurface stuff you switch these modes quite often, please add it to modal map.

CC @Germano Cavalcante (mano-wii)
CC @Campbell Barton (campbellbarton)

Can confirm.

Thanks for the update. I can confirm too. This is actually a regression introduced between 98a04ed45242 - 8d9d5da13706
Not bisected yet but I suspect rB5045968f24b5: Revert "Gizmo: optimize intersection tests, fix selection bias" is the breaking commit (reverts cc8fe1a1cbc6 too)
cc @Campbell Barton (campbellbarton)

Resolved by:
e441e21d74f9
64984126a2b6
bb8488c62c9a
fb7ef40006b9
f35d671f466d

Tested version: 3.2.2, branch: master, commit date: 2022-08-02 18:15, hash: rBbcfdb14560e7. The same problem is present

The patch LGTM, still not so keen on modifiers having to create hide data even when it's not needed (object mode display for e.g.). But this could be handled as part of a separate patch (if at all - the additional memory use is fairly minor).

• Don't save hide attributes in files at all (since they're still written in the old format anyway)
• Parallelize layer copying
• Fix attribute names in assert
• Disallow procedural access
• Rename .hide_face to .hide_poly
• Merge branch 'master' into refactor-mesh-hide-generic

The convention to use "." + snake-case for built-in custom-data should be documented, suggest: CustomDataLayer.name although it could be elsewhere.

I'll do that separately, if you don't mind. The convention is already used (see allow_procedural_attribute_access).

Prefer the name .hide_poly for the polygon layer, as there is face data which is separate (and the term poly in general when assigning variables etc).

The current name is meant to be consistent with the name of the attribute domain, I'm fine with using "poly" too. (I hope we can replace the particle system by 4.0 so this stops being a problem).

Commit message should note the file size increase from this change (roughly).

File sizes shouldn't need to increase with this patch, since we still write and read with the legacy format (until 4.0).

Generally seems fine, some minor notes only.

+1 to fix this.

@Cian Jinks (Hobbes) thanks!

@item412 (item412) My bad, I forgot to include it in the corrective release list! I will see if 3.2.3 is a thing and I will add it to that list.

Blender version 3.2.2, the issue is still present

There seems to be a problem with passing vertexCos into MOD_deform_mesh_eval_get in this case (instead of NULL like all other modifiers do).
This sets me->runtime.edit_data->vertexCos which then causes issues ensuring normals (BKE_mesh_vertex_normals_ensure)?
mvert_len and mpoly_len coming into BKE_mesh_calc_normals_poly_and_vertex is garbled (much too high).

Caused by rBdeaff945d0b9: Mesh: skip conversion from edit-mesh to mesh in edit-mode