Re: Breaking /emul/ia32-linux for squeeze
On Thu, Mar 12, 2009 at 04:54:51AM +1100, Aníbal Monsalve Salazar wrote:
> Do you have a date for the glibc change?
I was hoping for pretty soon after a thorough discussion.
On Wed, Mar 11, 2009 at 09:12:31PM +0100, Kurt Roeckx wrote:
> /usr/lib32 isn't exactly FHS either, but it's better than
> /emul/ia32-linux
>
> Will this also change for ia64? As far as I know, there that path
> is hardcoded in the kernel.
Can you be more specific? I'm not aware of this.
On Wed, Mar 11, 2009 at 05:36:26PM -0400, Michael S. Gilbert wrote:
> Is this necessary? There are already softlinks set up:
> /usr/lib32->/emul/ia32-linux/usr/lib and /lib32->/emul/ia32-linux/lib.
It's not necessary any more than conforming to standards is necessary.
We could just as easily move everytihng to /Library, but that would
upset people just as much as /emul does.
> If debian is going to go down this road, then I think plain old lib
> should also get removed in the process (e.g. put everything into either
> lib32 or lib64; temporarily provide a lib softlink until a full
> transition happens). This could make a 64 to 128 (or other bitness)
> transition and associated backwards-compatibility easier, perhaps...
That is a far more disruptive change than what I had intended.
On Wed, Mar 11, 2009 at 11:16:47PM +0100, Tollef Fog Heen wrote:
> Could we pretty please use the multiarch paths here if we start moving
> stuff around? We're going to need to patch gcc/binutils if we're to
> compile stuff against those paths anyway.
By multiarch paths, you mean /usr/lib/i386-linux/ in this case?
I'm fine with doing that and changing both /usr/lib32 and /emul/ia32-linux
to be symlinks thither for the time being.
Reply to: