stripping xfs.o module stops it working on Linux

[Assar Westerlund]

"Brandon S. Allbery KF8NH" <allbery at ece.cmu.edu> writes:
> On that subject, Nat Lanza gave me some modified patches -- he claimed mine 
> didn't work for him because they passed NULL/0 to k_afslog_uid().  But in 
> every version of kth-krb I've used, k_afslog_uid() uses defaults... I'm 
> guessing he was using the local -lkrbafs code and it doesn't do defaults (I 
> haven't looked at or used it; unlike CS, we've never needed it).  Thus, a 
> new aklog patch will be forthcoming.

As far as I can tell, all versions of k_afslog_uid have accepted cell
== NULL and realm == NULL.  On the other hand, I know that the krbafs
code is rather old and hasn't been updated by its lazy maintainter.
(I guess the real answer is to try to make a separate distribution out
of libkafs (but that's work - best left to somebody else) or make the
world see the true light and only run the code I say they should :-)

/assar