Processed: Re: Bug#1035383: unblock (pre-approval): brial/1.2.11-2.1
Processing control commands: > tags -1 confirmed moreinfo Bug #1035383 [release.debian.org] unblock (pre-approval): brial/1.2.11-2.1 Added tag(s) confirmed and moreinfo. -- 1035383: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1035383 Debian Bug Tracking System Contact ow...@bugs.debian.org with problems
Bug#1035383: unblock (pre-approval): brial/1.2.11-2.1
Control: tags -1 confirmed moreinfo Hi, [I was hoping another Release Team member would shim in, but alas. Please read till the end.] On 02-05-2023 16:11, plugwash wrote: Elbrus replied to my bug report, challangeing why I had filed it as rc, I explained my position and he seemed somewhat but not totally convinced. Although I agree with you on the normal approach, I explicitly note here that this source builds *multiple* arch specific binaries and only one fails to install on several architectures where it builds fine. Manually maintaining a list of architectures is a PITA. I would like to ask for a release team ruling on this bug. If the release agree it is rc and should be fixed, I am happy to make an upload doing so. On the other hand if the release team decide that it is not rc and should not be fixed at this stage in the release process I'm happy to abide by that descision. I think we have a way forward, because it's claimed that the binary python3-brial should be removed in the not too distant future, so this architecture hard-coding is a temporary solution. (However, there are two reverse dependencies of the package, so that will need to be solved in trixie). Having said that, I want to hear other opinions for the general case. Paul OpenPGP_signature Description: OpenPGP digital signature
Bug#1035383: unblock (pre-approval): brial/1.2.11-2.1
Package: release.debian.org Severity: normal User: release.debian@packages.debian.org Usertags: unblock It was discovered about a month ago by Bastian Germann that python3-brial needs python3-sage, and he added a dependency. Unfortunately this left the package uninstallable on about half of release architectures. Normally this would block migration to testing, but elbrus forced the package in. I filed a bug 1034443 with grave severity for this based on the following understanding. * An uninstallable package is unusable * The "is this package unusable" criteria is applied to each binary package individually and for packages that are built seperately for multiple architectures is applied on each arhictecture individually. Or to put it another way my understanding the criteria is applied to each "deb" individually. I don't think these are explicitly stated anywhere, but they are consistent with my experiance of how things are typically done in Debian. They are consistent with the state of testing (other than python3-brial there are no uninstallable arch-specific binary packages in testing) and they are consistent with the rules britney normally enforces for testing migration. Elbrus replied to my bug report, challangeing why I had filed it as rc, I explained my position and he seemed somewhat but not totally convinced. I would like to ask for a release team ruling on this bug. If the release agree it is rc and should be fixed, I am happy to make an upload doing so. On the other hand if the release team decide that it is not rc and should not be fixed at this stage in the release process I'm happy to abide by that descision. The debdiff for my proposed upload can be found at https://bugs.debian.org/cgi-bin/bugreport.cgi?att=2;bug=1034443;filename=brial.debdiff;msg=40 unblock brial/1.2.11-2.1
