00:06 zmike: tarceri: are you still reviewing piglits or do we just yolo merge
00:11 tarceri: zmike: what piglits should I be looking at?
00:13 zmike: probably all within the last week
00:13 zmike: I merged a couple already
00:13 tarceri: The rule has generally been if nobody comments in a couple of weeks then its generally ok to merge
00:14 zmike: I started making people who submit AI patches also submit AI piglits
00:14 tarceri: ok see one there now will take a look,
00:14 zmike: the AI tax
00:14 tarceri: lol
00:15 zmike: if they look ok just hit merge
00:36 orowith2os: can someone review !44749 to make sure I'm not going insane?
00:36 orowith2os: it's not ready for merge at all (waiting on upstream EGL extension work) but I want to make sure I didn't fuck anything up.
00:36 orowith2os: I believe I may also go back over EGL and do a bunch of cleanups, because it's... a bit of a mess on the dri backend.
07:22 MrCooper: orowith2os: see https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/26528
07:25 orowith2os: MrCooper: what about it?
07:25 MrCooper: using FIFO & commit-timing protocols for EGL
07:26 orowith2os: ah
07:26 orowith2os: I'll put it in my notes
07:27 orowith2os: not sure if I'd be able to pick it up without {{support}} but I can see what I can do after my other WSI stuff.
07:29 orowith2os: MrCooper: did you have any other things to address on the gallium/x11 PR, btw?
07:30 orowith2os: !44924
07:30 MrCooper: I pointed it out because you asked about that, no expectation of you picking it up (would need to coordinate with Derek)
07:30 MrCooper: if I did, I would have raised them
07:31 orowith2os: gotcha
07:31 orowith2os: I'll either wait for somebody to r-b it or something, then.
07:31 orowith2os: wow, I can't speak properly right now.
07:34 orowith2os: MrCooper: do you have any particular feelings about checking for xwayland (using the XWAYLAND extension) and using ASYNC for mailbox, and ASYNC_MAY_TEAR for unpaced with tearing?
07:34 orowith2os: I might be able to implement EGL-Registry/!228 on XWayland like that, if that's fine.
07:35 orowith2os: (my efforts there have largely been focused on the Wayland side, since I don't think X11 clients care enough there ^^;)
07:35 MrCooper: not sure why Xwayland detection would be needed
07:38 orowith2os: mmmh, fair I suppose...
07:39 orowith2os: I guess I can always treat ASYNC as mailbox if ASYNC_MAY_TEAR is supported
12:00 zmike: alright but this is a bit much https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/45200
12:09 mareko: happy review
12:11 mareko: 2027 will be the year of zink
12:13 zmike: I meant because I read the commit logs and the timing of the commits and there is a certain consistency which makes me skeptical
12:24 zmike: one item the current AI policy also doesn't cover at all: I think now to get developer/merge access we should mandate that the patch count requirement be met solely with patches that do not use AI
12:26 mareko: I thought it was legit :/
12:26 zmike: it most certainly is not
12:52 CounterPillow: wow I'd hate to be obligated to deal with that.
12:52 zmike: I appreciate the enthusiasm but it really is a bit much
12:52 DragoonAethis: even the MR description has a specific smell to it, so
12:54 zmike: yup yup anyway we are actually really close to landing basic zink video I think, which is pretty exciting after 5 fucking years
12:54 CounterPillow: yeah thanks to this MR by Claude™
12:54 CounterPillow: okay I'll stop now
13:09 zmike: nowrep: any other comments?
13:12 nowrep: no, rest is good
13:12 zmike: nice
13:13 zmike: will check your last one out when I get home, then I guess I'll merge once the anv series lands (and the bp passes)
13:13 zmike: seems a bit aggressive to try jamming it into this release
14:22 MoeIcenowy: zmike: it sounds necessary but counting someone's non-AI commits isn't as easy as counting someone's commits ;-)
14:25 zmike: it isn't as difficult as you might think
14:26 karolherbst: could also do something else, because granting commit access to a critical repo just by doing 25 commits is... a low bar to clear.. like if somebody feels like they deserve commit rights, they file an issue and somebody is going vouch for them.. or not
14:26 zmike: I really thought it was 50
14:26 karolherbst: maybe I confuse it with something else that was 25
14:27 zmike: no, I think you're write as far as the text is concerned
14:27 zmike: but we've used 50 in a number of cases
14:27 zmike: and I think 50 is not such a high bar
14:28 karolherbst: I couldn't even find the text...
14:28 zmike: ask AI
14:28 MrCooper: IIRC the threshold for access to Valve games on Steam was 25 Mesa commits :)
14:28 karolherbst: but I think it's a bit dated.. like.. if somebody wants it, that person should file a ticket and somebody says okay or not... we can really do it as simple as that :D
14:29 zmike: yeah maybe
14:29 karolherbst: we could say "ask if you have at least 25" or so though
14:29 karolherbst: *only
14:29 karolherbst: yeah.. getting keys was gated at 25 for sure
14:44 dcbaker: I mean, most mesa contributors work for one of a handful of companies, so those companies/teams can kinda vouch for their people
14:44 dcbaker: and the rest end up being drive by contributors, or proflific
14:45 dcbaker: als, PSA: I'm planning to start the branchpoint process around 13:00 PDT, which is roughly 4 hours from now
14:45 dcbaker: If there's any huge things you'd like me to wait on going through marge before I do that please let me know ASAP
14:46 MoeIcenowy: dcbaker: sigh another team at my employer is sending slops out
14:46 austriancoder: dcbaker: I have the impression eric_engestrom is doing the release too: https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/45199#note_3712366
14:47 karolherbst: yeah.. not sure how much people are willing to trust those companies these days
14:48 zmike: just use AI to review AI :p
14:48 dcbaker: austriancoder: There's also a 26.2 release scheduled for today
14:48 austriancoder: dcbaker: ahh ..
14:48 dcbaker: karolherbst: I should have been clear, I trust the existing people from said companies to vouch for other people from their company
14:48 karolherbst: ohh yeah
14:53 glehmann: apparently we have 230 people with merge access
14:53 zmike: jesus
14:53 glehmann: that's 130 more than I expected
14:53 karolherbst: yeah....
14:54 karolherbst: like.. I don't want to say that we might overdo it here... or that anybody doesn't know how to secure their system... on the other hand... that's kinda a risky thing on its own 🙃 We should at least figure out how/when to prune that list a bit
15:43 mareko: AI-generated review comments (as quoted text with attribution) should be allowed on AI-generated MRs
15:44 zmike: tbh I don't really care if the policy allows for this or not, I'm gonna do it anyway
16:03 DragoonAethis: How does Marge deal with assignment actually?
16:04 DragoonAethis: Does it check for who assigned a MR for a merge?
16:27 glehmann: only developers can assign people to MRs, so I don't think marge needs to check anything
16:56 orowith2os: <dcbaker> "als, PSA: I'm planning to..." <- Can I put in a wishlist item 🥺
18:52 zmike: nowrep: I'll probably ask you to re-review the video stuff tomorrow; I asked my LLM to do a max depth review of the branch today and it found a ton of stuff (half of it was based off radeonsi code from 5 years ago that you have since rewritten) which I'm having it fix
18:54 dcbaker: Yves Auxier (she/her) (Vague and Factual) (Threatening): let me know what you need. I can either make sure it lands, or I can pull it back if doesn't land before I branch
18:55 orowith2os: @dcbaker it was this pull request, I already asked daniels to review it, but it's all fairly simple. I asked Michel and there wasn't anything from his end of concern either.
18:55 orowith2os: https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/44924
18:58 orowith2os: dcbaker: also to clarify, this shouldn't block release, it's just a QOL thing that I would like to land ASAP ^^;
19:00 dcbaker: PSA: If there are MRs that are currently in the marge queue that you want in the release, please add them to this milestone: https://gitlab.freedesktop.org/mesa/mesa/-/milestones/58#tab-merge-requests
19:00 dcbaker: I will pull those all in manually for you after the branchpoint
19:00 dcbaker: If they haven't gone to marge, just add the Backport-to: 26.3 tag in commit messages as normal
19:01 dcbaker: Yves Auxier (she/her) (Vague and Factual) (Threatening): If you add the `Backport-to:` tag that MR will autoomatically get picked up ^
19:02 orowith2os: TIL, thanks ^^
19:02 dcbaker: Yves Auxier (she/her) (Vague and Factual) (Threatening): we also use `fixes:` tags like the kernel, let me see if I can find those docs...
19:04 dcbaker: https://docs.mesa3d.org/submittingpatches.html#the-fixes-tag
19:06 dcbaker: I did move all of the MRs that are currently on the marge queue and are a release blocker to this milestone
20:08 alyssa: dcbaker: Ironically I think that deprioritizes them getting merged since (at least a few years ago) marge sorted by last updated and not by time of assignment
20:08 alyssa: ;P
20:09 omrash: If that's true then just leave a comment on every other assigned MR :)
20:09 alyssa: omrash: you joke, but-