tir. 8. sep. 2020 kl. 14:29 skrev Thomas Goirand <z...@debian.org>: > > On 9/8/20 2:01 PM, Håvard Flaget Aasen wrote: > > On Tue, 8 Sep 2020 13:40:25 +0200 Thomas Goirand <z...@debian.org> wrote: > >> Hi, > >> > >> Since nobody seem to care about this, I've uploaded disabling that > >> failing test to DELAYED/2. Indeed, to me, it looked like a broken test, > >> rather than a broken package. Otherwise, a long list of reverse > >> dependency will be removed from testing in 8 days. > >> > >> Debdiff of the change attached. Let me know if you would like me to > >> cancel the upload. > >> > >> Cheers, > >> > >> Thomas Goirand (zigo) > > > > Thanks for this Thomas but I actually have a package up for RFS to > > address this issue. > > I must admit I removed all tests during build. Because of the same > > reason. same testsuite is being run as an autopkgtest which succeeds. > > Previous there has also been some issues with these tests bug: [761325] > > I intend to reopen this bug. > > > > Can I persuade you to review and upload my own package instead of doing > > an NMU? bug: [965370] package in mentors [3] > > > > Thanks, > > H??vard > > > > [761325] https://bugs.debian.org/761325 > > [965370] https://bugs.debian.org/965370 > > [3] https://mentors.debian.net/package/wand/ > > > > Hi Håvard, >
Hi Thomas, > I'm very much surprised by this. Why didn't you write in the bug report > then, saying you've addressed the issue and was searching for a sponsor? I guess the reason for that is my workflow, I work from a forked repo and don't push it to the official until it has been released into Debian. I will change this from now on, so the BTS can track when I close a bug in salsa, and make it easier to collaborate with others. > I had the impression that nobody cared, you know... And you can't expect > everyone to closely monitor mentors. I don't expect it, but for me, and many others, it's the only way to get a sponsor. There is a pane in the package tracker page, that informs about the RFS. > > I do not agree that removing all tests is the solution, when only a > single one fails. I would happily sponsor your upload, but *not* if it's > removing all tests. I do agree with you that disabling all the tests isn't the best option, but keep in mind that I enabled the testsuite with the last release, and it didn't last much more than two months before a test failed, because of the test, not the package. The testsuite had been disabled for four years before that. The autopkgtest also uses the same testsuite, which strangely enough succeeds. If my reasoning isn't enough, I will comply and change the package to only disable that one test. > > Cheers, > > Thomas Goirand (zigo) Regards, Håvard
