On Monday, July 9, 2012, Thibaut VARENE <vare...@debian.org> wrote: > On Fri, Jul 6, 2012 at 4:32 PM, Sean McGovern <gsean...@gmail.com> wrote: >> >> >> On Thursday, July 5, 2012, Måns Rullgård <m...@mansr.com> wrote: >>> Sean McGovern <gsean...@gmail.com> writes: >>> >>>> On Thursday, July 5, 2012, Måns Rullgård <m...@mansr.com> wrote: >>>>> Sean McGovern <gsean...@gmail.com> writes: >>>>> >>>>>> Hi folks, >>>>>> >>>>>> It looks like the test failures we are seeing on these 2 build >>>>>> configurations are false negatives. The checksums are correct, but the >>>>>> filename text is missing a * in front. This must be a faulty regexp in >>>>>> tests/md5.sh >>>>> >>>>> Which variant does it end up using? >>>>> >>>> >>>> I don't have these setups myself (they're run by Thibaut VARÈNE) , but I >>>> think this broke relatively recently -- which suggests it's the 2nd >>>> variant >>>> that uses 'command' that we changed in 00c78a0a. >>> >>> Without knowing which command is used and what its output actually looks >>> like, there is nothing I can do. >>> >> >> Hello Thibaut, > > Hi, > >> Is there any chance you could verify which of the MD5 checksum commands in >> tests/md5.she is being used on these PPC Darwin FATE runs? > > I'm not sure what I'm supposed to do there. Care to guide me a little? ;) > Note that Mik has full access to this machine and to the testsuite > running on this machine. >
Hi Thibaut and Michael, In the Libav source tree, tests/md5.sh contains 4 different ways of getting an MD5 checksum for comparison with a known value. Recently, these tests have been "failing" on PPC Darwin (my x86_64 versions run on 10.6 Snow Leopard don't have the same failure). I need to know which of the 4 commands are being used on that run. This would most likely involve commenting them out one-by-one and running 1 of the failing FATE tests. -- Sean McG.
_______________________________________________ libav-devel mailing list libav-devel@libav.org https://lists.libav.org/mailman/listinfo/libav-devel