On 27 June 2011 04:41, David W. Hodgins <[email protected]> wrote: > On Sun, 26 Jun 2011 21:41:00 -0400, Ahmad Samir <[email protected]> > wrote: > >> On 27 June 2011 03:11, David W. Hodgins <[email protected]> wrote: >>> >>> On my Mageia 1 install ... >>> >>> urpmi --auto-select --resume --noclean >>> To satisfy dependencies, the following packages are going to be >>> installed: >>> Package Version Release Arch >>> (medium "Core Release") >>> libxine1 1.1.19 5.mga1 i586 >>> (medium "Tainted Release") >>> libvlc-devel 1.1.9 4.mga1.taint> i586 > >> vlc-1.1.9-4.mga1.tainted existed in the tainted/release repo before >> Mageia 1 was ever released, so it's not an update. > > Ok, got it. I had installed vlc before enabling tainted. > >> Tainted packages have a higher EVR than core ones >> (vlc-1.1.9-4.mga1.i586 vs. vlc-1.1.9-4.mga1.tainted.i586), hence urpmi >> proposed to replace vlc from core by vlc from tainted. > > With libxine1, though it's strange. The log shows I ran > urpmi: called with: -a --fuzzy xine- > which installed libxine1-1.1.19-4.mga1.tainted.i586. > > When I ran the auto-select it installed > libxine1-1.1.19-5.mga1.i586 from Core, replacing the tainted version. > > That makes sense as 19-5 is greater than 19-4. I don't understand > is why the first urpmi command didn't select the 19-5 version, while > the update did.
Indeed, ideally the release of both of core and tainted are kept in sync, this is a glitch, i.e. someone submitted to core/release and forgot to submit to tainted/release, or he submitted to tainted but the build failed... etc. Please open a bug report about that one. > > Also why they aren't both 19-5. > -- Ahmad Samir
