[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: current glibc vs debian sid
From: |
Akim Demaille |
Subject: |
Re: current glibc vs debian sid |
Date: |
22 Oct 2002 14:38:33 +0200 |
User-agent: |
Gnus/5.0808 (Gnus v5.8.8) XEmacs/21.4 (Honest Recruiter) |
| Anyway, I think the better question is why should the behavior have changed
| in this way instead of being compatible with how it has always been. It
| has always been blindingly obvious to me what the right thing to do is:
| Set build with config.guess unless --build is specified;
| set host=build unless --host is specified;
| set target=host unless --target is specified;
| if target_alias != host_alias then decide you are building a cross compiler;
| if `config.sub host` != `config.sub build` then decide you are
cross-compiling.
This is exactly what happened! This is why many people wanted to
clean this area. Also, there were too many broken environments that
were considered as cross-compiling environments by configure, for it
used to think that `if I can't run what I compile, I'm
cross-compiling'. In the end, too many novices were caught, experts
were confused by the double interface to specify the
build/host/targets (via options or as argument, or both!), and even
the GCS were wrong (or confused) in their definition of
build/host/target.
We did what we could to preserve some form of backward compatibility,
but the scheme was broken and to be changed.
Re: current glibc vs debian sid, Roland McGrath, 2002/10/18
- Re: current glibc vs debian sid,
Akim Demaille <=