Re: bootstrap/3250: gcc

[law@redhat.com]

The following reply was made to PR bootstrap/3250; it has been noted by GNATS.

From: law@redhat.com
To: Craig Rodrigues <rodrigc@mediaone.net>
Cc: jzago@ifaedi.insa-lyon.fr, gcc-gnats@gcc.gnu.org, rodrigc@gcc.gnu.org,
        gcc-bugs@gcc.gnu.org
Subject: Re: bootstrap/3250: gcc 
Date: Mon, 21 Jan 2002 09:02:16 -0700

  In message <20020118191617.A8286@mediaone.net>, Craig Rodrigues writes:
  > Now, PR 3250 reports the same error, but using
  > gcc 2.95.3 on Solaris 2.5.1:
  > http://gcc.gnu.org/cgi-bin/gnatsweb.pl?database=gcc&pr=3250&cmd=view
  > 
  > steve@smc.vnet.net rejects my e-mail domain, so I can't get feedback. :(
  > 
  > What should I do with this PR?\
 Unsure -- I scanned it, but didn't find the details I was looking for; 
 specifically we'd need both the gcc.log output and maybe more details for
 the Solaris system (the gcc.log appears to be for hpux).
 
 It's also likely we'd have to look at the header files on the Solaris
 machine where this build is failing -- at least in the HPUX case the
 root of the problem is an annoying mis-match between va_list provided
 by the system header files and the one provided by GCC.
 
 This causes gcc-2.95.x era compilers to issue a warning (gcc-3.0.x compilers
 will ignore the minor mis-match).  The warning causes the configure process
 to believe that locale.h is unusable.  Thus HAVE_LOCALE_H is not defined
 and thus it isn't included by toplev.c.  However configure did fine  the
 setlocale function, so toplev.c has the call to setlocale which contains
 the LC_ALL as an argument.
 
 jeff