On 25 May 2010 01:28, Lucas Nussbaum <lu...@lucas-nussbaum.net> wrote: > I've just done an archive rebuild. 459 packages failed. 340 are new > failures (they didn't fail two weeks ago). I've attached the full list > of new failures.
I've started an archive-wide checkbashisms check on the source packages. As of the time of writing, it has checked 9415 source packages, and with some commands magic I'm counting 1113 packages with bashisms in configure scripts. There are of course some false positives among them. There's a total of 2758 source packages with some sort of bashism in any /bin/sh script in the source package (which might or might: be used during the build process or included in the binary packages.) The numbers are possibly higher because I fixed some bugs in checkbashisms' parsing code > How do you want to proceed? Should I file bugs on all those packages, or > wait until a "fixed" dash is uploaded? I can't speak for the RT, but they should really be treated as RC. > IMHO, the most reasonable option is to keep the current package as is, > to change the shell in the buildd chroots to bash temporarily, and to > file those bugs as RC. That way, people can easily test and fix them, > and the buildds are not affected. I agree. I will later send a RFH to -devel with the list of affected packages, pointing to a location where the full results can be found. (will start another archive-wide check against the binary packages with the improved checkbashisms too, will try to fast-process it. I.e. only review results if there's no goal-dash br already filed against the package.) Cheers, -- Raphael Geissert - Debian Developer www.debian.org - get.debian.net -- To UNSUBSCRIBE, email to debian-bugs-rc-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
