Hi Niels, On 2018-08-14 17:53, Niels Thykier wrote: > Control: tags -1 moreinfo > > Aurelien Jarno: > > Package: release.debian.org > > Severity: normal > > User: release.debian....@packages.debian.org > > Usertags: binnmu > > > > Hi release team, > > > > One of the goal of the glibc team for buster is to get rid of the > > multiarch-support package, given that multiarch support in the dynamic > > loader is fully supported in debian for a long time. Debhelper used to > > add this dependency, but it's not the case anymore. > > > > clsync is one out of two packages left in buster, the other one, > > gtk2-engines-oxygen, should hopefully migrate in the next days. > > Unfortunately clsync has an RC bug in sid, which prevents it to migrate > > to buster. This RC bug is not present in buster, so the package hasn't > > been autoremoved from there. > > > > Therefore the easiest way to fix the issue is to trigger a binNMU in > > *buster*, against a "recent" debhelper: > > > > nmu clsync_0.4.1-1 . ANY . buster . -m "Rebuild to remove the > > pre-dependency on multiarch-support" > > > > Thanks, > > Aurelien > > > > [...] > > Hi Aurelien, > > If the goal is to make testing free of packages (Pre-)Depending on > multiarch-support
Yes, that's exactly the goal, that way we can just stop building the package in the next upload. There are still a few depending packages in sid/experimental, but given the bug reports have been sent with patches attached more than one year ago, I don't mind breaking them. > then removing clsync from testing is also an option > AFAICT. > > I am considering since clsync has had its FTBFS bug against sid for > almost 2 years now with no reaction from its maintainers. That solution has the same result than the binNMU, so that's also fine for me. I would leave the decision about what is the best way to achieve that to the release team. Thanks, Aurelien -- Aurelien Jarno GPG: 4096R/1DDD8C9B aurel...@aurel32.net http://www.aurel32.net
