Will do some testing of this and the other pending PR tomorrow, otherwise small comments, nothing major.
I think it's worth adding a comment to the doc here, or maybe in the file, saying that the invert_strength is what differentiates the BLOB and CREASE brush. Alternatively, I don't think it would be a bad idea to make two functions:
I noticed the sculpt_project_v3_cache_init function has a conditional checking against FLT_EPSILON to seemingly prevent it from causing len_sq_inv_neg to be extremely large if the value is nearly 0. I think we probably need a similar check somewhere.
Minor nit: Any reason to prefer writing this out over calling pow? Here and below on line 277
As far as I understand it, the main reason we add the Scene parameter here and elsewhereis so that we eventually can call this method so we can get either the brush value or the unified value. This doesn't seem to me like a thing that we want long term, though passing it in is certainly better than retrieving it from SculptSession -> ViewContext -> Scene
Doesn't this comment contradict the BLI_NOINLINE?
32a78e480f
Merge branch 'main' of projects.blender.org:blender/blender into smooth-mask-brush-refactor
Minor nit: It seems like the only reason we pass down the entire Sculpt object is so we can access the .flags for clipping - do we want to extract this somehow?
We can just skip
SculptOrigVertDatacompletely, usingSpan<float3> positions = undo::get_node(...)->position
My only issue with this is that I think having an explicit named method like…