bug-inetutils
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [bug-inetutils] On r-commands and iruserok.


From: Mats Erik Andersson
Subject: Re: [bug-inetutils] On r-commands and iruserok.
Date: Fri, 30 Dec 2011 18:37:05 +0100
User-agent: Mutt/1.5.18 (2008-05-17)

fredag den 30 december 2011 klockan 10:17 skrev Alfred M. Szmidt detta:
>    At the moment I am struggling to get the format string right for rlogind
>    when it calls login(1) on Solaris. The available string is outdated, but
>    is is hard to get it right. On Solaris rshd functions properly, but rlogind
>    produces an immediate logout until I can find the correct string.
> 
> What about a different approach, can we skip the this step or doing it
> differently?  I.e., why do we call login?  Can we do the same using
> normal library calls?

We must let the system implement any PAM policies it so desires,
therefore we still are confined to login(1) as superintendent.
Even more so, once the kerberized code enters the build.
The execle(3) invocation is now plausible for Solaris, but
perhaps my PAM setting is incomplete, thus preventing my success.

> I'm heading of for the train up to Stockholm today, and was planning
> on making the release on the train before new year.  If there is
> something that must must must get into this release, please say and
> push the fix.

I have pushed the changes I need and have ready. Together with the
previous change in "rcp" (OpenSolaris segfaults on format "%qd")
we now build full r-tools sets on GNU/Linux, GNU/kFreeBSD, OpenBSD,
NetBSD, and GNU/OPenSolaris; probably also for FreeBSD and OpenIndiana,
but I lack time to test this.

The gang of five (rcp-rsh-rlogin-rshd-rlogind) mostly work with known
issues for NetBSD, GNU/OpenSolaris, and GNU/kFreeBSD. (I have tested
three generations today!)

> One release per year is something we must be able to do, and damned be
> I if I don't do one this year!

Make your release now. I am sufficiently happy with the porting that
was achieved for NetBSD and OpenSolaris in the last few weeks. We will
encounter issues related to the present state of the r-tools once we
examine the Kerberos code during next iteration. They are heavily
intertwined, so it is better to hope that some independent soul
happens to test and comment or complain.

Best regards,
  Mats



reply via email to

[Prev in Thread] Current Thread [Next in Thread]