[519] in SIPB_Linux_Development
Re: runACT(), file_exists check is *backwards*
daemon@ATHENA.MIT.EDU (svalente@MIT.EDU)
Mon Mar 14 03:17:52 1994
From: svalente@MIT.EDU
Date: Mon, 14 Mar 94 03:17:39 -0500
To: eichin@MIT.EDU
Cc: linux-dev@MIT.EDU
In-Reply-To: "Mark W. Eichin"'s message of Sun, 13 Mar 94 11:49:53 EST <9403131649.AA13452@rt-11.MIT.EDU>
Mark has pointed out some some evil and incorrect hacks in the xlogin
code. I remember inserting these hacks into the code, I don't
remember exactly why. I remember that I figured I'd clean everything
up before ever really distributing xlogin.
Anyway, I've begun trying to fix them. This means that right now, the
xlogin sources are in an imcomplete state - they don't represent what
was last compiled on Keesh, and they may or may not compile at the
moment. I'll send mail again when I think I've got things sort of
working.
I think we should start using RCS on the Keesh athena source tree.
I'll check in xlogin sometime real soon now.
An other important problem with xlogin is that doesn't completely use
shadow passwords. It _does_ check the shadow password on local
accounts (like root), but it does _not_ add new users to /etc/shadow
when it adds them to /etc/passwd. This makes programs like
xscreensaver do unpredictable things. I'll try to fix this. We
should probably distribute two separate xlogin binaries - with and
without shadow password support.
Or, we could completely desupport shadow passwords. They're a good
idea, but they create a lot of problems. Does anyone have an opinion
on this?
Have a nice day.
-Sal.
