[PATCH] pam_smbpass for CVS HEAD

[Alexander Bokovoy]

On Tue, Dec 11, 2001 at 01:54:19PM -0600, Steve Langasek wrote:
> On Tue, Dec 11, 2001 at 08:54:30PM +0200, Alexander Bokovoy wrote:
> > Attached patch makes pam_smbpass compileable again in CVS HEAD.
> > I suspect that nobody uses pam_smbpass 'cause some of these problems
> > exist in SAMBA_2_2 branch as well for a while... But fixing it is
> > important as well.
> 
> Well, or perhaps no one that uses pam_smbpass spends a lot of time in
> the HEAD branch. :)  I don't do a lot of constant tweaking w/
> pam_smbpass, because very little changes in that code except to keep it
> up-to-date with the current Samba branch for the purpose of /compiling/;
> for /running/, you can just as well use pam_smbpass compiled against
> Samba 2.0.7, and unless you need something cutting-edge like LDAP or tdb
> storage, it'll interoperate fine with whatever version of Samba is on
> the server.
This module (pam_smbpass) is included into my own compile set test which I
run against each Samba build because we are still supplying it in ALT
Linux's Samba RPMS. 

> BTW, in your patch I noticed that you've removed the xstrdup()
> function... I believe in the SAMBA_2_2 branch, this function has been
> renamed, not removed -- if the xstrdup() function that's common to Samba
> is less paranoid than the one in pam_smbpass, this would seem
> appropriate to do for CVS HEAD, as well.
In CVS HEAD there were two xstrdup() definitions -- one global which all
Samba code used and one in pam_smbpass/support.c. They obviously
conflicted in the same namespace during linkage phase. If you want to keep
pam_smbpass with its own xstrdup() version then it is better to rename in
(prefixing by something like smbpass_ for example). I saw that renaming in
SAMBA_2_2, actually I even sent a patch to do so a few months ago so may
be it was result of my action. :) 

-- 
/ Alexander Bokovoy
$ cat /proc/identity >~/.signature
  `Senior software developer and analyst for SaM-Solutions Ltd.`
---
Nov 21 20:58:58 alconost kernel: VFS: Busy inodes after unmount. 
		    Self-destruct in 5 seconds.  Have a nice day...