Re: False negative in ben?
On 06/04/12 16:16, Stéphane Glondu wrote: Le 06/04/2012 13:00, Joachim Breitner a écrit : I regularly check http://release.debian.org/transitions/html/haskell.html, it is a great tool. But I am confused why it says haskell-sfml-audio, -openal and -alut is bad on armel armhf mips mipsel ppc s390 s390x. According to the parameters, this means that some of the packages of the source are uninstallable, but edos-debcheck or apt-get install in a chroot work flawlessly. Ben considers only the latest version of arch:all packages, whereas dak waits for the source package to be built on $arch before making the new versions of its arch:all packages available on $arch. and this is also true for all binary packages (arch:any too) and source packages… in ben. I'm not sure exactly what the desired behaviour is in this case... Britney does the same. So I'm not sure we should modify this behavior. It seems a sane filter when considering the migration problem. My 2cents, -- Mehdi -- To UNSUBSCRIBE, email to debian-release-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/4f86fb93.6000...@debian.org
Re: False negative in ben?
Le 06/04/2012 13:00, Joachim Breitner a écrit : I regularly check http://release.debian.org/transitions/html/haskell.html, it is a great tool. But I am confused why it says haskell-sfml-audio, -openal and -alut is bad on armel armhf mips mipsel ppc s390 s390x. According to the parameters, this means that some of the packages of the source are uninstallable, but edos-debcheck or apt-get install in a chroot work flawlessly. Ben considers only the latest version of arch:all packages, whereas dak waits for the source package to be built on $arch before making the new versions of its arch:all packages available on $arch. For example, in the case of -openal, libopenal1 (1:1.13-6) is indeed not installable on armel (from Ben's point of view): it depends on libopenal-data (1:1.13-6), which is arch:all and non-existent since only version 1:1.14-1 is considered. I'm not sure exactly what the desired behaviour is in this case... Cheers, -- Stéphane -- To UNSUBSCRIBE, email to debian-release-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/4f7efab3.8090...@debian.org
