Re: patch for pam authentication in cvs

[Kevin Wang]

 From Brian Murphy
> >Brian Murphy writes:
> >>Otherwise I can't see why one of these patches is not included
> >>in the standard CVS they are no worse than the standard
> >>system password checking mechanism.
> >
> >We try not to throw things into CVS just because someone thinks it would
> >be a good idea.  When lots of people tell us they are successfully using
> >a patch and would like it to be included in the standard distribution,
> >then we consider it.  (Of course, if some *developer* thinks it's a good
> >idea, the process does get considerably sortened.)  Now that we have an
> >"experimental" branch, there's a better chance of things getting
> >accepted there.
> >
> Well it just so happens that PAM in particular is mentioned on the 
> wishlist of
> future enhancements (cvsclient.info) and also mentioned in the current 
> password
> checking code as the right fix to a hack on HP/UX to get the encrypted 
> password.

Also, virtually all modern OSes use PAM, and is much preferable than
a separate database of usernames.  Especially when you want to be
able to control access to cvs via ldap, windows domain, or some other
non-unix-/etc/passwd method.

Chalk me up for one more supporting vote for getting SOME sort of pam
support in the base product.  I don't care where, how, or when; that's
your job!

   - Kevin