Re: current glibc vs debian sid

[Roland McGrath <roland at gnu dot org>]

> That subject has been discussed more than once on the Autoconf mailing
> lists.  It's not clear to me that there is a GNU coding standards
> violation, but if you see it clearly then I'd appreciate a brief
> explanation.  I'm interested in the practical issues as well as any
> coding-standards violations.

I think the practical issues are obvious: What worked before does not work
now.  People expect "configure HOST" to work for a native compile, even
when specifying HOST is wholly superfluous (after all, some of us remember
when config.guess didn't exist and even a time or two where it didn't get
the right answer).  

The coding standards say, "The configure script should also take an
argument which specifies the type of system to build the program for."
This is a very reasonable basis for that expectation.  Nowhere does it say
"unless you meant a native compilation"--on the contrary, it is clearly
implied that native compilation is the normal case, since cross-compilation
scenarios are mentioned only briefly and after everything else.