-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 According to Gary V. Vaughan on 10/27/2008 11:49 PM: > The current test code inside strerror.m4 says: > > AC_LANG_PROGRAM( > [[#include <string.h> > ]], > [[return !*strerror (-2);]]) > > Which is fine for Solaris 9 and 10 which return "Unknown Error", > but causes a SIGSEGV on older Solaris machines
This is intentional. The gnulib module for strerror is _supposed_ to replace the Solaris 8 strerror, because of this very bug. Use of the strerror module entitles you to write code that can correctly assume that strerror never returns NULL, even on out-of-range input. > The easy fix would be to change both instances of 'strerror(-2)' > to 'strerror(1)' in strerror.m4, but that is still prone to the > same failure if 1 is not a valid errno. Maybe forcing an errno > setting error by opening a non-existent file is more prudent? No, that would be breaking the module's promise of always returning a valid string. > > 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. That might be a bug - if the gnulib strerror module is not in use, then the #define strerror rpl_strerror should not be in place. But since it looks like gnutls has been subsequently fixed to avoid strerror altogether, I'm not sure if this still indicates any problems. - -- Don't work too hard, make some time for fun as well! Eric Blake [EMAIL PROTECTED] -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.9 (Cygwin) Comment: Public key at home.comcast.net/~ericblake/eblake.gpg Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iEYEARECAAYFAkkG+y0ACgkQ84KuGfSFAYCK1ACfaSkOzljDOiH3nwamnXKPdl0V cZkAoMbXi64Q9cvQckwVqmNIi5neJeHW =ejdr -----END PGP SIGNATURE-----
