[Bug 107] samba-3.0.1

Alexander E. Patrakov semzx at newmail.ru
Fri Jan 2 05:34:32 PST 2004

On Friday 02 January 2004 17:05, Greg Schafer wrote:
> All I know for sure is.. the autoconf macro in the bug
> report was clearly reversed and therefore wrong. It is now right.

Do I understand correctly that the autoconf macro usage was wrong and also the 
test program returned the inverted return codes, and that's why the end 
result was correct in my case before you reported the bug?

Also, do I understand correctly that your bug report is about the presumed 
case when the test program fails to compile at all, but the autoconf check 
returns that memory keytabs are supported?

> If there
> are other bugs in that autoconf macro checking code then that is a
> different issue and worthy of a different bug report.
> I just had a quick scan of samba bugzilla and saw this:
> https://bugzilla.samba.org/show_bug.cgi?id=912
> At a quick glance it looks identical to your problem. It at least has the
> autoconf macro round the right way. Maybe it'll fix your problem!

Thank you very much, it is exactly the same problem. But I don't see the fix 
in the CVS yet. Do you agree that the following is correct (the exit codes 
are changed to the opposite ones relative to SAMBA 3.0.1)? Remember that 
non-zero return value from krb5_kt_resolve indicates an error.

    krb5_context context;
    krb5_keytab keytab;

    if (krb5_kt_resolve(context, "MEMORY:", &keytab))

  if test x"$samba_cv_HAVE_MEMORY_KEYTAB" = x"yes"; then
               [Whether in-memory keytabs are supported])

> Reminds me of an old Linus quote:
> '"regression testing"? What's that? If it compiles, it is good, if it
> boots up it is perfect'
> :-)

Windows XP is perfect, but not good :-)

Alexander E. Patrakov

More information about the blfs-book mailing list