10:43 <glehmann> why can this happen on a project where I have the developer role?

10:43 <glehmann> > Your comment could not be updated because your comment has been recognized as spam. please, change the content to proceed..

10:47 <mupuf> glehmann: because gitlab is dumb like that :s

10:48 <mupuf> someone will have to look into how to hook into the spam detection and disable it for any internal users

10:51 <bentiss> glehmann: submitted as ham, hopefully this will help in the future. And yeah, as mupuf said. (and sorry)

10:51 <glehmann> ham?

10:51 <bentiss> not spam

10:52 <bentiss> don't ask me for the whole story, but IIRC it's because of those ham in a can thingy you can find in the US

10:53 <glehmann> ah, thanks

14:37 <martink> hey all. How does one dry-run a ci job on a forked Mesa, i.e. sans doing an actual MR?

15:07 <eric_engestrom> martink: when you push your changes to your fork, gitlab will create a pipeline with the jobs for your change (note that it only looks at what changed since your last push, unlike on MRs, so you might expect some jobs and not actually get them; if that's the case, add a dummy commit with a dummy change in a file core to the work on your branch)

15:07 <eric_engestrom> martink: once that pipeline is created and contains the jobs you want, you can run `bin/ci/ci_run_n_monitor.sh --target 'job-regex|for-the-jobs|you-care-about'`; add `--force-manual` to also run manual jobs that match the regex. The script will monitor the jobs and tell you their state, and if your terminal supports it you can click on the job name to open the full log (otherwise, manually go to your fork, pipelines, and

15:07 <eric_engestrom> open the one you're running)

15:19 <martink> eric_engestrom, thank you for the elaborate response

19:17 <eric_engestrom> bentiss: could you look at the marge logs? it's completely misbehaving :(

19:19 <eric_engestrom> https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/26372 it's claiming the commit is already in main when that's not the case

19:19 <eric_engestrom> and now it looks like it's adding that commit to all the other MRs it's trying to merge

19:19 <eric_engestrom> https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/26354

19:19 <eric_engestrom> https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/26369

19:19 <eric_engestrom> and when the pipeline passes, it fails with "I couldn't merge this branch: had some issue with GitLab, check my logs..."

19:35 <bentiss> eric_engestrom: not sure what happened. I am kicking it to restart it in the hope it'll fix itself by doing a fresh restart

19:35 <eric_engestrom> thanks!

19:36 <eric_engestrom> oh btw, are we running marge from https://gitlab.freedesktop.org/freedesktop/marge-bot or marge from https://gitlab.com/marge-org/marge-bot?

19:36 <eric_engestrom> asking because I sent this MR upstream: https://gitlab.com/marge-org/marge-bot/-/merge_requests/411

19:37 <eric_engestrom> (to have marge cancel pipelines when it gives up on an MR, to avoid compounding the infrastructure overload)

19:39 <bentiss> eric_engestrom: we are running https://gitlab.freedesktop.org/freedesktop/fdo-bots/-/blob/main/values/marge-bot/equinix-DC.yaml.gotmpl?ref_type=heads -> registry.freedesktop.org/freedesktop/marge-bot/marge-bot:0.10.1-new-upstream-pipeline-id-part-of-fix

19:44 <bentiss> That image is 6 months old, so unlikely to have your MR in it

20:00 <jenatali> Looks like restarting it worked

20:21 <eric_engestrom> indeed, looks like it's working normally now

20:23 <eric_engestrom> bentiss: regarding my MR, it's not merged yet (waiting on someone to review it) so this doesn't apply yet, but I was wondering if I should backport it to our fork or if we just need to bump to the latest upstream

22:34 <eric_engestrom> bentiss: looks like Marge is back to the same broken state; can you try to nuke its working directory?

22:34 <eric_engestrom> (or tomorrow, it's late already)

23:29 <daniels> mm, I don't understand why that's happening as it did pull origin and try to rebase on origin/main, with those showing no errors ...