On Mon Apr 28 12:50:03 2025 +0000, Huw Davies wrote:
Do your tests in commit 3 show that the changes in commit 2 are necessary? What I'd suggest is to swap the order of commits 2 and 3, associated with commit 3 would be the removal of some `todo_wine`s.
I went ahead and made the tests very first commit, also I restored original logic in the tests where eM21 = -eM12 * 2.0f, that shows more failures in the current code. Is this what you had in mind?