On Tue, 2004-05-18 at 14:01 -0400, Matthew Miller wrote: > On Mon, May 17, 2004 at 10:47:42PM -0400, Jeremy Katz wrote: > > On Sun, 2004-05-16 at 23:09 -0400, Matthew Miller wrote: > > Yep, this is with anaconda-10.0-5? I had late breaking changes fixing > > this stuff all up... > > Well, I rebuilt today against a complete FC2-final tree, and it works fine > now. So something in there with the mixed packages was doing something > weird. Dunno what. Oh well. :) Okay. Sticking my fingers in my ears and going "la la la la la la" > But, in the midst of doing that, I learned something interesting: if you > leave out glibc-*.i386.rpm (and just have the i686 package), everything will > seem to work, except that you won't get any glibc at all in the second stage > image, which will of course manifest itself as "/usr/bin/anaconda: file not > found" (or something like that), which will be very confusing, as you keep > saying "what are you talking about? it's _right there_". There's not a ton that I can really do about that. The error is basically what you get back from the kernel. I guess that buildinstall/upd-instroot could error out if "essential" packages are missing. Accepting patches :) > > And actually, are you on i386 or x86_64? It seems this worked on the > > final FC2 i386 tree but not x86_64. Probably for a reason that would be > > obvious to me if it wasn't nearly 11 pm > > i386 -- x86_64 is my agenda for July. So I'm not thinking about it yet. Heh, fair enough Jeremy
