Le Sat, Jul 11, 2015 at 11:35:38PM -0700, Afif Elghraoui a écrit : > > I was talking about relying on the user to run the tests and manage CI, > but that seems to be not what you intended and that this particular test > runner was just broken. If it is possible to achieve both-- allowing > automated testing and providing the user the ability to manually test > the package, then I don't have such strong feelings about it.
Hi Afif, for some packages it is doable to run the tests at build time and also distribute them in a binary package that is used for the autopkgtests and for users to troubleshoot their own installations. See samtools for example. In other cases (Perl packages, R packages), the tests seem to be too much part of the build system to be easily packaged exhaustively in a stand-alone way. Or it is just that I never found how... For python-pysam, I think that I could start the packaged test suite in the past. Sorry that I did not have time to follow your work in details; is this package is now broken then it makes sense to remove it, otherwise I would argue that it is useful and should be kept. Lastly, sometimes for some pacakges (for instance bedtools), the regression tests never passed fully on Debian. This makes them less useful, but at least, they can be monitored to spot regressions, where even more tests would fail. Have a nice day, -- Charles -- To UNSUBSCRIBE, email to debian-med-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/20150712110555.gb1...@falafel.plessy.net