Re: naming problem: ix86-pc-linux-gnulibc1 / ix86-pc-linux-gnulibc2 / ix86-pc-linux-gnu

[Andreas Jaeger <aj at suse dot de>]

Christian Lademann <lademann@zls.de> writes:

> Please enlighten me:
> 
> Is there an official statement on how to name the architecture for
> ix86-pc-linux-gnu* environments?
> 
> It seems to be a good and supported practice to configure the tool-chain with
> "--host=ix86-pc-linux-gnulibc1" to
> support libc5. It would be nice if one could also use
> "--host=ix86-pc-linux-gnulibc2" for libc6 but at least
> gcc requires a modified "configure" to handle that target correctly. Instead,
> one has to use "--host=ix86-pc-linux-gnu"
> to configure correctly for the recent glibc. But what will happen when, say,
> glibc3 is released? Will '...-linux-gnu' then

There's no discussion about a glibc 3 release in the glibc team and
therefore we do not need to care in GCC to support something that
might happen in ten years... - and even that it might be that glibc 2
and 3 are compatible in such a way that we don't need to differentiate
them at all.

> describe the new version and will '...-linux-gnulibc2' be introduced to build
> the tools for glibc2? That way all tools for glibc2
> will have to be rebuild (which might make sense that day anyway... :-). I think
> it is better to name the targets conflictfree
> right now and avoid clashes in the future, so I would like to use
> "...-linux-gnulibc2".
> 
> I know that config.guess will give a functional result for the architecture name
> (of the compiling machine) but that is
> not allways what you want.

Changing this now might breask lots of systems, there's no need to do
so.

Andreas
-- 
 Andreas Jaeger
  SuSE Labs aj@suse.de
   private aj@arthur.inka.de
    http://www.suse.de/~aj