This is the mail archive of the
libc-alpha@sourceware.cygnus.com
mailing list for the glibc project.
Re: [Craig Metz <cmetz@inner.net>] libc/1184: Latest glibc + toolchaincan't really build 386 lib
- To: Andreas Jaeger <aj@arthur.rhein-neckar.de>
- Subject: Re: [Craig Metz <cmetz@inner.net>] libc/1184: Latest glibc + toolchaincan't really build 386 lib
- From: Chris Seawood <cls@seawood.org>
- Date: Thu, 1 Jul 1999 04:36:40 -0500 (EST)
- cc: libc-alpha Mailinglist <libc-alpha@sourceware.cygnus.com>, Craig Metz <cmetz@inner.net>
On 1 Jul 1999, Andreas Jaeger wrote:
> We've received the append bug report. Any ideas what's broken here?
You want to use -march=i386 rather than -m386. -m386 only sets -mcpu=i386
which, iirc, only deals with optimization levels and -march determines
whether or not the cpu architecure specific opcodes are used. Last I
checked (early 1.0.x), egcs seemed to default these values to be whatever
cpu/arch it is built on. However, the egcs 1.1.2 that comes with RH6.0
appears to use mcpu=i386 & march=i386 so that might have changed.
- cls