Several Package updates not seen by ALHP although updated on Arch weeks ago #292
Reference in New Issue
Block a user
Delete Branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
All those packages were updated on Arch on April 9, but status.alhp.dev still lists the version before as current version. I think there are more of those, but those are the only ones I see after recently disabling alhp repo's for some days.
Example:
https://status.alhp.dev/?pkgbase=confuse&exact=
https://archlinux.org/packages/?sort=&q=confuse&maintainer=&flagged=
another one:
winetrickswas updated to 20260125-2 on May 14, but ALHP only sees 20260125-1https://archlinux.org/packages/extra/any/winetricks/
https://status.alhp.dev/?pkgbase=winetricks
I had some time to investigate, seems like the upstream state repo is out of sync in case of those packages. Will see what I can get upstream to see if this is a systemic problem or not. At least alhp seems to do everything correctly in this case.
example for confuse: https://gitlab.archlinux.org/archlinux/packaging/state/-/blob/main/extra-x86_64/confuse
In the case of winetricks, the arch changed to
anysince it doesn't produce any binaries, is it related?Could be.
-anymoves are still pretty buggy because the state seems to duplicate in such cases. The bot does try to determine the correct statefile to use but its still not perfect.Should be fixed on the alhp side. ALHP now compares the upstream alpm version against the db, which catches both the missing pkgrel bumps and -any moves like winetricks.
Caught a lot more drift than just confuse/jbig2dec/libmspack/signond, all queued. winetricks dropped since its any-arch now.
Still want to follow this up upstream though, when the time allows for it.
Thanks.
Totally missed the arch - change of
winetricksI think I remember reading some conversation on Arch's IRC devops or projects channel that they had issues with states not syncing or matching but that was weeks ago and I thought they fixed that. Can't find the conversation though because something in my ZNC server does not work as expected. Also can't find an issue on gitlab.