This is the mail archive of the
libstdc++@gcc.gnu.org
mailing list for the libstdc++ project.
Re: std_limits.h os_defines.h
- To: Mark Mitchell <mark at codesourcery dot com>
- Subject: Re: std_limits.h os_defines.h
- From: Benjamin Kosnik <bkoz at redhat dot com>
- Date: Thu, 9 Aug 2001 15:37:05 -0700 (PDT)
- cc: David Edelsohn <dje at watson dot ibm dot com>, "gcc-patches at gcc dot gnu dot org" <gcc-patches at gcc dot gnu dot org>, "libstdc++ at gcc dot gnu dot org" <libstdc++ at gcc dot gnu dot org>
Ummm.... whoops.
I'm going to stand back on this and let Gaby and Mark handle it. I'd
prefer it if discussions on what is going on were public, but hey. C'est
la vie. Please make sure that what every you come up with allows
newlib-based cross compilers to work (ie, test on x86-linux cross
powerpc-eabism)
> I think the scheme that we are planning on using is robust enough
> for now and it solves a very important problem: making cross-compilation,
> and multilibs, work correctly. Since I thought you agreed, I decided
> to go this way in 3.0.1.
... of course, cross compilation works, as it uses limits_generic.h. The
problem seems to be native multilibs where the ISA changes, such as AIX.
> Doing something better in 3.1 is great.
Yup.
> Do you have time to look at the problem Gerald reported with colons
> in path names that is apparently triggered by the staging headers
> work? That is one of the few regressions from 3.0, and we definitely
> need a solution.
I'll look at it.
-benjamin