This is the mail archive of the
libc-hacker@cygnus.com
mailing list for the glibc project.
Re: glibc2.1 [offtopic]
- To: zack@rabi.columbia.edu (Zack Weinberg)
- Subject: Re: glibc2.1 [offtopic]
- From: hjl@lucon.org (H.J. Lu)
- Date: Thu, 11 Feb 1999 08:06:57 -0800 (PST)
- Cc: libc-hacker@cygnus.com, egcs@egcs.cygnus.com
> >
> >Well, it is on the mainline now.
>
> Are we talking about the same patch? The one I don't like is the one
No. I was talking about
Sat Jun 27 19:06:04 1998 H.J. Lu (hjl@gnu.org)
* configure (gxx_include_dir): Changed to
'${prefix}/include/g++'-${libstdcxx_interface}.
* config.if: New to determine the interfaces.
.....
>
> >> you created it to fix. Instead it creates more binary
> >> incompatibilities.
> >
> >It works for me. You never gave me a convincing example to show
> >it is broken. Maybe we have different opinions on what "broken"
> >means in this context.
>
> "It works for me" != "it works for everyone".
>
> You localized a bunch of symbols that have been in libgcc since GCC1
> and will never go away. Things like _muldi3. Those symbols are
> re-exported by libc. We can't take them out without breaking every
> binary that needs them.
It is up to libc to export it, not libgcc. It still works for me.
_muldi3 is global even with my libgcc.map patch.
--
H.J. Lu (hjl@gnu.org)