Hi Andreas! Andreas Enge <andr...@enge.fr> writes:
> could you just base your patch on master? The idea of world-rebuild is > that I pick a few patches myself (so as to keep control, otherwise there > will be too many) by adding them to the milestone. This one looks > interesting, as I wanted to take a few from debbugs. > > So I would suggest that you either declare this pull request for master, > or close and reopen it on master. > > world-rebuild itself has so far not moved; I just branched off of master > at some point in time to be able to create the merge request on QA. > It still needs a bit of waiting anyway. There are no patches on world-rebuild yet, so rebasing on master is “free”. That’s actually why I didn’t bother to branch my MR directly off world-rebuild, although now that I know this approach causes so many Codeberg notifications, I’ll avoid any attempts at cleverness and always branch directly off the target. Since world-rebuild has not yet diverged, I believe there’s no need to retarget or reopen the PR. You can rebase world-rebuild on cyrus-sasl-xoauth2 (AGit branch, so I guess the remote refspec to fetch would be ‘refs/pull/2013/head’) or cherry-pick the patches if you specifically want world-rebuild to diverge from where it is now for some reason. Let me know if I misunderstand anything! > If possible, it would be good to do a “build -P 1”; I try to add only > patches for which all distance 1 dependents build (which, due to the > state of master, it not that easy - sometimes they do not build for the > current packages either). I gave this a shot, but the first build, compface, failed immediately. According to ‘guix weather’, it’s not built on berlin or bordeaux either. Perhaps a candidate for removal since it hasn’t been meaningfully updated since Danny (CC’d) added it in 2017? The home page 404s. Looking at the definition, I’m not even sure how it depends on cyrus-sasl. I tried with ‘–keep-going’, and things generally seem to be building, but it’s quite a lot for my little machine, so I had to stop. Unfortunately there’s no summary after you interrupt, but I saw Samba succeeded at least. Perhaps the simplest approach in the world rebuild case is to apply patches to the staging branch somewhat aggressively so we can let Cuirass do the hard work of determining if new failures exist? If so, you can aggressively drop them and I can investigate further. —Liam
