00:24 <airlied> daniels: yeah I hit that as I go past

02:28 <alyssa> " New accounts can only file issues in existing projects. "

02:28 <alyssa> about that :-(

11:48 <zmike> seems like swrast ci is dead again

12:06 <daniels> can confirm

12:06 <daniels> zmike: can you please push in an MR to kill the jobs again? you should just be able to revert okias's revert

12:07 <zmike> daniels: https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21836

12:08 <zmike> feel free to assign

12:12 <daniels> done, thanks

12:12 <zmike> time to merge all my untested lavapipe patches

12:20 <daniels> classic you

12:20 <daniels> though, tbf, in my slaughter piglit run overnight for panfrost+a618+zink-tu, the only one which didn't demonstrate any new flakes was zink-tu

12:21 <zmike> got a lot of flakes added on that one already

12:21 <daniels> well good adding, because it was literally flawless across 200 runs

12:21 <zmike> that's a lot of runs

12:22 <daniels> tbf that was only because I forgot to ^C the stress test before I went to bed

12:22 <zmike> haha

12:23 <daniels> oh, I take it back, dEQP-GLES3.functional.texture.specification.texstorage3d.format.depth_component16_2d_array did flake once on zink-tu

12:23 <zmike> f

12:39 <DavidHeidelberg[m]> :D

12:57 <mohamexiety> daniels: hello! sorry for the ping but I am having a really weird problem where connection to freedesktop is really slow and I can't seem to figure out why. trying to clone any repo and I can't get over 30 KB/s transfer no matter the time I try in. I asked here a few days ago but couldn't get a reponse

12:58 <mohamexiety> this is only for freedesktop. cloning any other outside repo / regular internet usage / etc are fine, so this seems exclusive to fdo somehow.

12:58 <mohamexiety> talking to other devs (I only know mesa people), nobody else has this issue. also, I am in Egypt if this helps isolate the issue

13:25 <mohamexiety> hm, trying again right now I am getting 300 - 700 KB/s which is much more serviceable than 30 KB/s, but still a fair bit lower than my peak (3.5 MB/s). for reference, cloning git gives me 1.5 - 2 MB/s

13:26 <__tim> what exactly do you mean by 'connection to freedesktop'? which service/host/port?

13:27 <daniels> mohamexiety: are you cloning from gitlab.freedesktop.org or anongit.freedesktop.org?

13:30 <mohamexiety> daniels: gitlab.freedesktop.org

13:30 <daniels> mohamexiety: mtr output would help

13:40 <mohamexiety> daniels: https://pastebin.com/NWKsWUzx (mtr -r -c 10 gitlab.freedesktop.org)

14:03 <__tim> reminds me of https://withinboredom.info/blog/2022/12/29/golang-is-evil-on-shitty-networks/

14:05 <daniels> mohamexiety: mmm yeah, I would suspect it's just some bad routing along the way - 150ms doesn't seem totally outlandish from Egypt to the west coast, but that does look to me like there's probably an overloaded peering link in the middle somewhere, and the provider has decided to not prioritise our host

14:05 <daniels> mohamexiety: I'm afraid there isn't anything we can do about that, but if you have some way of bouncing through an alternate host somewhere, that should probably work around it

14:18 <mohamexiety> daniels: alright then, got it. thanks a lot!

14:19 <daniels> np, good luck

14:30 <bentiss> heh, looks like someone opened the hose on sitewrangler :)

18:37 <alyssa> https://gitlab.freedesktop.org/mesa/mesa/-/jobs/37865829

18:37 <alyssa> what happened here?

18:39 <alyssa> looks like it passed, why marked fail?

18:41 <daniels> alyssa: it's collapsed in 'artifacts: prepare' - uploading them returned a 503

18:41 <daniels> it's immaterial since your pipeline would've timed out anyway

18:42 <daniels> (there's a huge stampede of jobs coming in on a Friday night, and GitLab distributes jobs based on random object hash value rather than queue time, so it's possible for them to get starved for too long; I've just typed out a patch for that)

18:43 <DavidHeidelberg[m]> alyssa: sadly gitlab cannot uncollapse sections when error occurs.. it sucks

18:43 <alyssa> woof, ok, nw

18:43 <DavidHeidelberg[m]> ...really hard.

18:43 <alyssa> just wasn't sure if it was something wrong with my MR i needed to fix or not

18:47 <eric_engestrom> https://lists.freedesktop.org/archives/freedesktop/ is missing an email received 4h ago

18:47 <eric_engestrom> incidentally, I wanted to link to that email here, since it's another user whose account was deactivated and they can't log back in

18:48 <eric_engestrom> bentiss: maybe ^

19:30 <bentiss> ok, that's enough with those: 1. I reverted the "Deactivate dormant users after a period of inactivity", and 2. I am reactivating all accounts

20:39 <Venemo> hi guys, can someone please help me figure out what is happening here? https://gitlab.freedesktop.org/mesa/mesa/-/jobs/37872690

20:39 <Venemo> it seems that it doesn't even run any tests

20:41 <eric_engestrom> Venemo: looks like a network error on the device; the file downloads fine for me

20:42 <eric_engestrom> I know you've already retried, but my advice is to retry again :]

20:43 <eric_engestrom> mupuf: navi10 runners failing to download ^