Subject: | |
From: | |
Reply To: | |
Date: | Wed, 10 Jan 2007 13:19:27 -0600 |
Content-Type: | TEXT/PLAIN |
Parts/Attachments: |
|
|
On Wed, 10 Jan 2007, Troy Dawson wrote:
> Glenn Morris wrote:
> > Glenn Morris wrote:
> >
> >> Maybe I misunderstand how this is supposed to work. Because the same
> >> thing also seems to be true for libjpeg-devel, libpng-devel,
> >> libtiff-devel, etc...
> >
> > ... in other words, I guess I'm supposed to be able to _run_ i386
> > binaries on x86_64, but not _compile_ them. (Though it does seem odd
> > to me that the libfoo.so symlinks get relegated to the devel
> > packages).
> >
> > Apologies for working this out on the list!
>
> Hi Glenn,
> You came to the correct conclusion before I was able to respond.
> The ultimate goal of x86_64 is to have everything run 64 bit.
> But that's not realistic, because there are still lots of programs that
> only run 32 bit.
> So whenever possible, everything is x86_64 (64 bit).
> If that isn't possible, we put in the 32 bit version, and all it's
Note that the "we" is mostly TUV.
> supporting 32 bit dependencies.
> For the development rpm's, we only put in the 64 bit devel rpm's, unless
> it just has to be 32 bit. It's quite rare that we put in 32 bit devel
> rpm's. Mainly because we're looking at the ultimate goal, of everything
> being 64 bit.
> People can get the 32 bit devel rpm's from the i386 tree if they want.
Note that trying to get some 32bit products to build on x86_64 is not
easy. We build 32bit products on x86_64 by using a i386 chroot area.
In RHEL5b2 TUV has added alot more of the i386 -devel rpms for the x86_64
release.
-Connie Sieh
>
> Troy
> --
> __________________________________________________
> Troy Dawson [log in to unmask] (630)840-6468
> Fermilab ComputingDivision/CSS CSI Group
> __________________________________________________
>
|
|
|