Re: ppp & pam (was: Re: ppp's ip-{up,down} and possible utilization of 'run-parts')
[ Brokenly-long lines wrapped ]
Philip Hands <phil@hands.com> writes:
> ppp is needed for doing an install from the internet via a dialup
> link. PAM is not needed until you want people to log into the
> system, so libpam is a waste of space on the install disks.
>
> I'm not certain it's worth the effort either, since libpam is only
> 21k and binary is almost exactly the same size (112 bytes bigger)
> --- opinions ?
libpam0g is 103k; libpam0g-util is 625k; libpwdb0g is a further 135k.
That's 863k. My complaints about pam and ppp are this:
o When I filed the bug you were linking a libc6 application with a
libc5 library. That's not good.
There is now a libc6-based pam in Incoming, once this gets processed,
this objection, obviously disappears.
o I've heard[1] that the pam support in ppp is badly broken.
I don't know if this is still the case with 2.3.2, I don't do ppp and
can't test it. If someone who can test it does so and says it's
okay, then this objection goes away too.
o By linking ppp with pam you are dragging libpam0g, libpam0g-util and
libpwdb0g into base.
This is fine, *as long as* it's been discussed and agreed first, I
don't like 3 shared library packages being silently dragged into
base. If we're going to do PAM for 2.0, then this will have to be
done anyway. But have really got the time to PAM-ify critical
applications like the shadow suite and have them working and debugged
and have 2.0 released before the next millennium?
> BTW does libpam0 need to be recompiled for libc6 before I can use it
> in ppp ?
Yes; linking with both libc5-based and libc6-based libraries is plain
Evil. (At least, I sincerely hope no one is going to dispute that)
[1] According to Michael Johnson on the pam list, which,
unfortunately, doesn't seem to be archived.
--
James
--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
debian-devel-request@lists.debian.org .
Trouble? e-mail to templin@bucknell.edu .
Reply to: