Re: strerror.m4 test fails incorrectly on solaris 8 and earlier [and bonus gnutls bug]

[Bruno Haible]

Gary V. Vaughan wrote:
> The easy fix would be to change both instances of 'strerror(-2)'
> to 'strerror(1)' in strerror.m4

The 'strerror(-2)' is there on purpose: We don't want a strerror
function that crashes. The comment in strerror.m4 says it:

    if test $gl_cv_func_working_strerror = no; then
      dnl The system's strerror() fails to return a string for out-of-range
      dnl integers. Replace it.
      REPLACE_STRERROR=1

and the test verifies it:

  str = strerror (-3);
  ASSERT (str);
  ASSERT (*str);

> I found this while linking against gnutls, which uses the string
> module, which then uses '#define strerror rpl_strerror' in its lgpl
> library when strerror is not found.  That, in turn, leads to an
> undefined rpl_strerror symbol in libgnutls.so.

Unless you provide details why libgnutls.so contains no rpl_strerror
function, or at least a "how to reproduce" recipe, I'd guess it's a problem
in gnutls' build infrastructure - and, as Simon said, probably already fixed.

Bruno