Re: Weak symbols in libresolv?

[Roland McGrath]

Please take that up on bug-glibc.  It is not a Hurd-specific issue.  There
are already other platforms for which it is a practical concern with this
very case (any platform in libc/shlib-versions that starts at GLIBC_2.2 or
later--currently s390x, cris, x86_64, ia64, sparc64, sh, hppa on Linux).

> Roland McGrath <roland@frob.com> writes:
> 
> > resolv.h #define's those symbols to the __ names, so anything
> > compiled with the current resolv.h ought to use the __ names
> > directly.
> 
> I'm sure you're aware that such name mangling makes a standard autoconf
> tests like
> 
>   AC_CHECK_LIB(resolv, res_query)
> 
> fail, as the configure script will try to link with the symbol without
> including resolv.h. The test must be rewritten as
> 
>   AC_CHECK_LIB(resolv, __res_query)
> 
> (Unless autoconf has been hacked since I had the same problem when
> upgrading to GMP-3.0, which uses similar name mangling defines like
> #define mpz_foo __gmpz_foo).
> 
> /Niels