Re: Problematic linking between glibc and shared libgcc

[Mark Kettenis <kettenis at wins dot uva dot nl>]

   From: Richard Henderson <rth at redhat dot com> 
   Date: Mon, 19 Feb 2001 12:06:34 -0800 

   On Mon, Feb 19, 2001 at 11:58:33AM -0800, Ulrich Drepper wrote:
   > (unless you keep track of the version numbers for each platform
   > individually which is a maintainer's nightmare).

   It is tracked indiviually for every platform.

Which obviously is the only right approach.  But who's in charge of
tracking things?  AFAIK there isn't even an official maintainer for
i?86-*-linux*, which probably is the platform where

   > I've said all this before and the solution gcc people came up
   > with is ignoring the problem.

   I've not been ignoring this problem.  But I'm going to start
   ignoring you if you keep spouting off about this.  We're not
   the complete idiots you want to make us out to be.

With all respect, most people on the GCC team are excellent compiler
engineers but hardly experts on (dynamic) linkers and shared
libraries.  Up to now you've been implenting this shared libgcc almost
singehandedly.  Some things still need to be sorted out
(libgcc-std.ver still contains some question marks, should collect2 be
made a little cleverer about when to choose the shared libgcc and when
not?), but I haven't seem much input from the other GCC maintainers on
these issues.  And the GCC release manager has repeatedly stated that
he doesn't want to special case any systems (when in fact you're
already doing that).  That makes people, including me, worry about how
the GCC team as a whole is committed to, and capable of, maintaining a
fully backwards compatible ABI for the shared libgcc on those systems
where GCC is the system compiler.

Mark