On Mon, 2013-04-22 at 19:06:52 +0200, Michael Biebl wrote: > Am 22.04.2013 17:24, schrieb Guillem Jover: > > Other problems might be when maintainer scripts use the running dpkg > > architecture instead of the package architecture (DPKG_MAINTSCRIPT_ARCH) > > to decide when to do stuff, for example in libglib2.0-0.postrm.
> I do think we handle all cases correctly now. If you see any issues, > please let us know. The following snippet seems suspect, although not a big issue anyway as it's just removing a cache file, although it might leave files behind on remove/purge: ,--- if [ -d /usr/lib/gio/modules ]; then # Purge the cache if [ $(dpkg --print-architecture) = amd64 ]; then rm -f /usr/lib/gio/modules/giomodule.cache rmdir -p --ignore-fail-on-non-empty /usr/lib/gio/modules fi fi `--- The running architecture of dpkg is irrelevant, you want to check against the package architecture set by dpkg as the environment variable DPKG_MAINTSCRIPT_ARCH. Because you could have installed libglib2.0-0:arm64 and libglib2.0-0:i386 on a native amd64 and the architecture will never match. But then, in that case, the whole check would be unnecessary, because DPKG_MAINTSCRIPT_ARCH should always match the package architecture replaced at build time; instead the package should only remove these when there's no other package instances left, regardless of the package architecture. Either that or I misunderstood the purpose of that code. :) (Unrelated but, why are those cache files not under /var/cache/ instead?) Thanks, Guillem -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20130422185101.ga12...@gaara.hadrons.org