On Mon, 22 May 2023, Jelle van der Waa wrote:
Hi,
Hi,
On 22/05/2023 17:11, Erich Eckner wrote:
1st: It would have been nice to have some testing setup around, so we could have adapted our scripts beforehand. But maybe, the test repositories were around and I just didn't look hard enough. No heart feelings, here - our users are used to somewhat bigger latencies :-D
We had a test setup, it was posted on arch-dev-public maybe it wasn't viisble enough for archlinux32.
Yeah, foxboron already told me that on irc. I guess, it's not really "your fault", that we didn't notice it - my spare time is shrinking constantly.
2nd: The mapping from pkgbase to the source repository is not clear. Most packages' repository is named exactly like the pkgbase. But when the pkgbase contains a "+", it is sometimes transcribed as "plus" (e.g. for libc++) and sometimes as "-" (e.g. for gtk2+extra). This makes automatic pulling of the needed repository unclean (we have to try several transcriptions, until one works).
Gitlab does not allow certain project names so we have a special mapping in our tooling sadly :( See these multiple sed's:
https://gitlab.archlinux.org/archlinux/devtools/-/blob/master/src/lib/api/gi...
Ah, that's, what I need - thank you! :-) Naively, I was expecting a "repository-name column" next to the pkgbase, version, version-with-mangled-epoch and git-commit-hash columns in the state repository.
Good luck with getting arch32!
Thanks, much appreciated :-) We're also through with the git migration - now follows the package migration >:-) regards, Erich