Ansgar Burchardt wrote: >Though it might make sense to move kfreebsd-* to ports.d.o. That was >planned to happen for hurd-i386 too.
That might be for the best… if not for this ultra-annoying shortcoming of mini-dak which hurts me on x32, and hurt me on m68k even more: when a new version of a library comes out, say src:gdbm now ships libgdbm5 and used to ship libgdbm3, mini-dak (contrary to dak) drops libgdbm3 (one run) after accepting the new version. This is a really big showstopper for the ideal procedure of basically letting buildds churn unattended and looking at failing builds only. Porters must manually look at the list of BD-Uninstallable packages on wuiet every once in a while and do their manual best to fix that, using porter uploads. (The old packages are still there on snapshot.d.o and can easily be fed into cowbuilder using a hook script.) So, unless someone who has knowledge of mini-dak (how it works and the programming language it’s written in) and fixes this bug, please don’t move architectures to ports for as long as it can be avoided. That being said: libcomerr2 in quinn-diff currently is a bit of in a WTF state, isn’t it? (That being said, on ports, we also sometimes have packages listed as BD-Uninstallable which aren’t, but it’s rare.) bye, //mirabilos (hat: x32 user and sorta porter, former m68k porter) -- [...] if maybe ext3fs wasn't a better pick, or jfs, or maybe reiserfs, oh but what about xfs, and if only i had waited until reiser4 was ready... in the be- ginning, there was ffs, and in the middle, there was ffs, and at the end, there was still ffs, and the sys admins knew it was good. :) -- Ted Unangst über *fs