[3580] in testers
Re: rlogin
daemon@ATHENA.MIT.EDU (Greg Hudson)
Sat Jun 13 01:09:48 1998
To: Joseph Sokol-Margolis <seph@MIT.EDU>
Cc: testers@MIT.EDU
In-Reply-To: Your message of "Sat, 13 Jun 1998 00:52:24 EDT."
<199806130452.AAA31895@EviSceRAte.MIT.EDU>
Date: Sat, 13 Jun 1998 01:09:36 EDT
From: Greg Hudson <ghudson@MIT.EDU>
> 0:48 smoke-screen(129):~% rlogin -a -x -l root casio
> Segmentation fault
Yeah, the "-a" option is pretty clearly broken (it passes a NULL value
of locuser to kcmd(), which completely fails to handle that case).
Should be easy to fix. You win logging into nowhere simply because
the krb5 login attempt failed, and the krb4 kcmd function doesn't make
use of locuser.
(I'm not sure what you want the "-a" option for.)
> 0:48 smoke-screen(131):~% rlogin -a -x -l root nowhere
> Couldn't authenticate to server: Bad response (during sendauth
> exchange)
> Trying krb4 rlogin...
> EVISCERATE.MIT.EDU: Connection refused
> 0:49 smoke-screen(132):~% rlogin -a -x nowhere -l root
> [works]
This one confuses me. Both forms of the command work fine when I try
to connect to cj, and there's nothing in the argument parsing which
suggests a dependency on order of arguments. Is this reliable?
