User:NathanVegdahl/WeeklyReports/2023

2023

May 15-19

PR and Issue review

  • Narrowed down #106801 to a minimal reproduction case.

Development

  • More work poking at heat diffusion weighting failures.
  • Worked on tracking down #106801, to actually fix it. No luck yet.

Animation 2025

  • Continued work on the performance prototype, adding a basic model of ID Blocks and the ability to query the animation system for a specific ID Block.


May 8-12

PR and Issue review

  • Proposed a design for #103802, based on the idea of simply respecting handle selection in the graph editor.
  • Confirmed #76549 as still an issue, and marked as bug.
  • Closed #85350 as not a bug, but with notes about why it is indeed confusing.
  • Identified #86371 as already fixed in main, but still affecting the current LTS releases (which were referenced in the issue). Left a comment asking if anyone needs a backport to one or both of the LTS releases. Original reporter says they don't need a backport. Also discussed with the animation module and decided to not backport, because the backport would likely be involved, and the bug is easily worked around.
  • Confirmed #84609 as still an issue, and provided a partial explanation of what's happening and how (for now) users can work around it.
  • Reviewed and approved #107869.
  • Reviewed and approved #107173 for Blender 4.0 (since it's a breaking change for the Python API).

Development

  • Investigated a quick fix for #84609 (automatic weights failure), and although it fixed the reported issue, it also caused other problems. So it's not a viable fix. Following that, figured out (experimentally) a "fix" that allows automatic weights to succeed over-all, leaving only the problematic vertices unweighted. Certainly a better behavior, but I want to better understand what's going on and hopefully figure out a better fix.

Animation 2025

  • Worked on animation data performance prototype. Some promising results, but I'm not confident it models current Blender internals well enough yet.
  • Talked to Falk David about whether the proposed animation data model would work/conflict with the new Grease Pencil design. Conclusion: they'll work quite well together.
  • Talked briefly with Rik about some of the ideas for Animation 2025. Definitely want to talk more with the animators! Great feedback and skepticism, which will help push the design to be really strong and useful.


May 1-5

PR and Issue review

  • Confirmed #106410 and marked as bug.
  • Confirmed #106542 and marked as bug.
  • Closed #106980 as not a bug.
  • Looked into #103802. It's not a bug, but probably needs discussion as to what the desired behavior is.
  • Finished #107223 and got it merged.
  • Identified #106340 as a duplicate (already fixed).

Animation 2025

  • Finished the draft animation data proposal, and gave a presentation at the animation module meeting.
  • Started a "performance prototype" of the proposed animation data system. One of the unknowns of the proposed system is how fast we can make it. So this is just a quick-and-dirty outside-of-Blender skeleton to get a broad-strokes idea of what the performance might look like. This is explicitly not to be the code of the actual system (e.g. it isn't even in the right programming language).

April 24-28

PR and Issue review

Development

Animation 2025

  • Created an outline of a concrete animation system foundations proposal. Still need to write/draw it up properly to share with the module.

April 17-21