Re: GDM and alternate authentication methods

["Brandon S. Allbery KF8NH" <allbery ece cmu edu>]

Marius Andreiana wrote:

> Ben Ford wrote:
> > I propose that we make GDM modular, in a sense, it can be the graphical
> > equivilant of PAM.  
> I thinks this is  the right approach. PAM should be used for auth, GDM
> only a gui 
> for it. PAM manual tells about the auth types you listed too.

That's nice.  Is there any intention of supporting systems that don't have 
PAM?  (I'm thinking of AIX 4.3.2 and Digital Unix 4.0D/4.0F.)

Not to mention that PAM+Kerberos+AFS is still a crapshoot, thanks to way too 
many program authors who either don't bother to read the manual or figure 
that because local auth still works if they cut corners, it's perfectly 
alright to cut those corners.  We *still* don't use PAM here because of 
this, even though there has been a general improvement at least on the Linux 
side of things.

In short:  will GDM be usable without PAM?

--
brandon s. allbery     [os/2][linux][solaris][japh]    allbery kf8nh apk net
system administrator        [WAY too many hats]          allbery ece cmu edu
electrical and computer engineering                                    KF8NH
carnegie mellon university      ["better check the oblivious first" -ke6sls]