On Mon, 12 Dec 2016, Chris Lamb wrote: > Santiago Vila wrote: > > > I guess this makes the bug not to be RC after all, not because > > it simply "does not usually happen" in the official autobuilders > > Not sure I agree with that logic, but that's a debate for elsewhere > and another time. :)
Note: I do not agree with such logic either :-). That was really not an endorsement of the (wrong) rule saying "if it does not happen on buildd.debian.org, it's not RC". Instead, it is just an acknowledgement of the fact that if we build packages with TZ=Etc/UTC, as it happens in buildd.debian.org, then the build failure will never happen. In other words, I consider Etc/UTC to be part of the "standard to build Debian packages". > However, more important is the question is: Does the fact that > python-babel FTBFS in such conditions imply that it is broken at > runtime for our users running under such timezones? Yes. That's a very good point indeed! So: Only in case the test failure does not mean that the program will not work properly for the users in such timezone, as explained by Chris: Does the proposed patch look good enough to be applied, so that we can finally fix this bug? Thanks.
