On Sat, 2007-09-22 at 19:44 -0400, seth vidal wrote: > On Sat, 2007-09-22 at 19:42 -0400, Michael Wiktowy wrote: > > On 9/22/07, seth vidal <skvidal@xxxxxxxxxxxxxxxxx> wrote: > > > If you want to avoid this behavior in the future add: exclude=*.?86 to > > > your yum.conf. That doesn't work when you want _some_ i386 packages which you specify explicitly. Or on ppc64/sparc64 where you do want the 64-bit kernel but all userspace can be 32-bit. It's just a fairly dodgy workaround. > > Is this something that anaconda should add to all x86_64 installations > > or is this just a workaround that might hide bugs and prevent Steve > > from being the canary in the coal mine? > > There isn't a bug here.Yum is doing exactly what it says to do. "it says to do"? I'm not quite sure of your meaning there -- I think you're really saying "yum is doing exactly what yum does". Which is true, but not really relevant. I consider yum's behaviour buggy, as do many others. Arguing over whether to call bug #235756 a bug or an RFE doesn't really buy us anything. It's just bug #235756. All bug reports are really requests for enhancement at some level, aren't they? If I say 'yum install foo', I just want _one_ version of foo installed -- the 32-bit PPC version that I actually intend to use. I don't want a pointless 64-bit version installed, and I don't want an i386 version installed for qemu to use under /usr/qemu-i386 either. If I want those, I'll ask for them. Neither do I want yum to search for ssh sessions where I am logged in as root to other boxes, and install 'foo' there too. Although 'install foo' is, when you think about it, a fairly vague command and can be wilfully misinterpreted to mean any of the above, it really _ought_ to be interpreted with a little bit of common sense. By doing _any_ but the first of the operations above, yum would be exceeding its mandate; doing something I didn't _really_ ask it to do. It's supposed to be a useful piece of software, for $DEITY's sake -- not a genie in a bottle trying to distort my wishes. It would be really nice if we had a proper fix for this misbehaviour before F8. -- dwmw2 -- fedora-devel-list mailing list fedora-devel-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/fedora-devel-list
