This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: egcs-1.1.1, mips-sgi-irix6.2
- To: mrs at wrs dot com (Mike Stump)
- Subject: Re: egcs-1.1.1, mips-sgi-irix6.2
- From: Joe Buck <jbuck at Synopsys dot COM>
- Date: Mon, 7 Dec 98 16:19:30 PST
- Cc: matthias dot abraham at wolfson dot oxford dot ac dot uk, oliva at dcc dot unicamp dot br, egcs at cygnus dot com
> > From: Alexandre Oliva <oliva@dcc.unicamp.br>
> > Date: 06 Dec 1998 01:42:10 -0200
>
> > AFAIK, libm was only required in the old days of libg++, because
> > some of the libg++ functionality needed functions from libm, and its
> > automatic inclusion could be dropped by now.
>
> Nope, that isn't why we do it. Check the libstdc++ usage of undefined
> externs and what libm.a provides, I think you'll find some.
libm is required by the complex number classes as well as elsewhere.
It needs to stay. libstdc++ calls libm functions.
Even if that weren't true, it seems absurd to ask users to provide an
extra flag for one small part of the standard library but give them the
rest (including much more complicated functionality) by default. The
requirement for typing -lm for C is, as far as I'm concerned, a historical
mistake.