This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: AIX binaries
- To: David Edelsohn <dje at watson dot ibm dot com>
- Subject: Re: AIX binaries
- From: Gabriel Dos Reis <Gabriel dot Dos-Reis at cmla dot ens-cachan dot fr>
- Date: 10 Jul 2001 01:04:45 +0200
- Cc: Phil Edwards <pedwards at disaster dot jaj dot com>, Gabriel Dos Reis <Gabriel dot Dos-Reis at cmla dot ens-cachan dot fr>, Matt_Conway at i2 dot com, gcc at gcc dot gnu dot org
- Organization: CMLA, ENS Cachan -- CNRS UMR 8536 (France)
- References: <200107092243.SAA25092@makai.watson.ibm.com>
David Edelsohn <dje@watson.ibm.com> writes:
| My patch re-invokes mknumeric_limits with the xcompiling flag (or
| whatever it was renamed) if the first attempt to run it fails. If the
| second attempt with xcompiling fails, the developer is SOL and we're back
| to the current behavior.
Hmmm, that sounds a bit contorsious. I would prefer we pass a flag
telling whether to generate <limits> for native build or a cross (in
which case, it would just have to copy a file somewhere).
| Personally, I would rather have include/bits/std_limits.h
| pre-computed in the config/{cpu,os} directory like atomicity.h. For
| instance, mknumeric_limits assumes that it can ascertain the FP rounding
| mode which the user can change on AIX.
C++ requires the rounding mode to be compile-time constants.
-- Gaby