13:53 <zmike> kusma: iirc I was using piglit/bin/teximage-colors for most of the l/a testing

13:54 <kusma> zmike: Right, but that doesn't change the fact that we don't use l/a formats at all...

13:54 <kusma> We just fall back to the state-tracker expanding them to RGBA

13:54 <zmike> wonder if I did some last minute refactor before merging

13:56 <kusma> Perhaps.. If you did something similar to the opposite of https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18596/diffs?commit_id=b49dffee9ae644bc1da0c415ab54642bfdaa0079, perhaps you were testing it before ;)

13:57 <zmike> could be

13:57 <kusma> But doing that now just breaks badly, which is what I'm trying to fix... However, the dEQP tests are yelling at me now...

13:58 <kusma> And I don't really understand why, the swizzles we end up with should as far as I understand be the right ones, but oh no, dEQP does not like.

13:58 <kusma> I need to leave now, but I'll take another look later tonight.

13:58 <zmike> 🙏