[5607] in Kerberos
Re: krb5-b4 telnet on solaris 2.4
daemon@ATHENA.MIT.EDU (David Parter)
Sat Aug 5 00:47:27 1995
To: Sam Hartman <hartmans@MIT.EDU>
Cc: "Nathan J. Dohm" <dohm+@cmu.edu>, kerberos@MIT.EDU
In-Reply-To: Your message of "Fri, 04 Aug 1995 12:27:24 EDT."
<199508041627.MAA19735@tertius.mit.edu>
Date: Fri, 04 Aug 1995 23:31:58 -0500
From: David Parter <dparter@cs.wisc.edu>
> Of course, these are guesses; it may just be completely
> unrelated to anything any of us think of.
In one of my earlier debugging sessions, I thought the problem was
related to spurious "urgent" data indications from select() in
process_rings(). Then it seemed to not be that at all....
I was right the first time. It seems that solaris 2.4 has a bug such
that recv() will cause false indications of urgent data, but read()
won't. change the recv() in process_rings to a read(), and the problem
goes away. This is SUN bug #1185700 (which I found on sunsolve AFTER I
had found the problem, but they couldn't find when I called to report it
earlier... oh well).
--david
