Re: Testing for backports of security/stable updates
Greetings, On Thu, May 17, 2012 at 01:34:05PM -0500, Micah Gersten wrote: > On 05/17/2012 01:27 PM, Evan Broder wrote: > > As part of trying to do a slightly better job of handling security and > > stable updates, I'd like to propose reducing our testing requirements > > for those backports. > > > > Proposal: For stable and security updates of backported packages which > > are not new upstream versions, it is not necessary to test reverse > > dependencies. The only testing required is builds/installs/runs of the > > package itself. > > […] > I would prefer testing of a subset of the reverse dependencies > (preferably the popular dependencies if any exist). While the backport > itself might not be an issue, there might have been changes in the > destination series which might have broken something. Having said that, > perhaps if we have stuff with good test suite coverage, we could rely on > those. We could say this: For cherry-picks / minimal updates (based on backporter assessment), the rdep testing requirement is waived for those rdepends which have not changed since the original backport testing (or which have themselves only had minimal updates). I guess this would require a script to figure out which versions were available when the backport happened, which wouldn't be entirely accurate if updates took place around the time the backport happened. Cheers, -- Iain Lane [ i...@orangesquash.org.uk ] Debian Developer [ la...@debian.org ] Ubuntu Developer [ la...@ubuntu.com ] PhD student [ i...@cs.nott.ac.uk ] signature.asc Description: Digital signature -- ubuntu-backports mailing list ubuntu-backports@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-backports
Re: Testing for backports of security/stable updates
On 05/17/2012 01:27 PM, Evan Broder wrote: > As part of trying to do a slightly better job of handling security and > stable updates, I'd like to propose reducing our testing requirements > for those backports. > > Proposal: For stable and security updates of backported packages which > are not new upstream versions, it is not necessary to test reverse > dependencies. The only testing required is builds/installs/runs of the > package itself. > > Objections to making this official backports policy? > > - Evan > I would prefer testing of a subset of the reverse dependencies (preferably the popular dependencies if any exist). While the backport itself might not be an issue, there might have been changes in the destination series which might have broken something. Having said that, perhaps if we have stuff with good test suite coverage, we could rely on those. Micah -- ubuntu-backports mailing list ubuntu-backports@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-backports
Testing for backports of security/stable updates
As part of trying to do a slightly better job of handling security and stable updates, I'd like to propose reducing our testing requirements for those backports. Proposal: For stable and security updates of backported packages which are not new upstream versions, it is not necessary to test reverse dependencies. The only testing required is builds/installs/runs of the package itself. Objections to making this official backports policy? - Evan -- ubuntu-backports mailing list ubuntu-backports@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-backports
