09:27 lucaceresoli: Hello! This super simple series is still missing R-by on many patches:
09:28 lucaceresoli: https://lore.kernel.org/lkml/20260831-drm-bridge-drm_panel-cleanup-v3-0-50efd8168b3e@bootlin.com/
09:28 lucaceresoli: What can I do? Can I apply it anyway?
09:28 lucaceresoli: mripard: ^
09:32 karolherbst: what's the implication of VK_FORMAT_UNDEFINED on attachments? I'm seeing clears on them and I see that e.g. radv skips the clear if they encounter that format. Is there a solid reason on why that's legal? Or does VK_FORMAT_UNDEFINED basically means "well.. if the format is undefined, I can do whatever I want"?
09:47 mripard: no, you can't apply it anyway
09:48 mripard: ping people to get them
12:55 lucaceresoli: mripard: OK, makes sense, I'll try another time
12:56 mripard: I'm not the only "people", try with the driver maintainers first and then you can try the bridge maintainers as well :)
13:25 mareko: I've found a use for AI that improves productivity while it doesn't actually do any coding
13:26 zmike: shitposting ?
13:28 mareko: to AI: review all commits in origin/main..HEAD (it usually discovers several real important issues)
13:29 mareko: yesterday it prevented hours of debugging, so that's quite something
13:30 zmike: cool
13:39 pixelcluster: bisect couldn't have helped narrow it down?
14:28 mareko: 10 bugs? I'd rather not
14:32 zmike: just because it's possible to do long division by hand doesn't mean it's the best way to do it
14:58 zmike: mareko: it seems like u_blitter is failing when trying to blit 2darray -> 3d with nonzero src layer
14:58 zmike: does this sound familiar?
14:58 zmike: seems like the layer is correct in blitter_draw_tex() at least
15:08 mareko: nope
15:08 mareko: it sounds like a VK driver bug to me
15:35 zmike: how did you know ?