Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!ucbvax!CLOUSO.CRIM.CA!philipp
From: philipp@CLOUSO.CRIM.CA ("Philip A. Prindeville")
Newsgroups: comp.protocols.tcp-ip
Subject: IP options...
Message-ID: <8904250239.AA26167@juicy-juice.lcs.mit.edu>
Date: 25 Apr 89 02:39:33 GMT
Sender: daemon@ucbvax.BERKELEY.EDU
Organization: The Internet
Lines: 24

I remember hearing on this list several months ago a discussion
about the semantics of interpreting loose/strict source and
record route options, and a possible document that was to come
of this.  Does anyone have a pointer to said document?
(As I remember, it was precipitated because a few implementa-
tions didn't do The Right Thing.)

Also, I was wondering if certain IP options are implicitly
(ie. the RFC doesn't explicitly say) mutually-exclusive,
such as loose and strict source and record routing, or
timestamping and record routes...  Actually it would seem
silly to have more than one of these in a packet.  Is it
an error to have more than one of these?

Lastly (for now), if one has a timestamp area that isn't
a multiple of 4 (or 8) bytes long, and the overflow field
overflows, should the Parameter Error message point to the
length or overflow field?

Please reply directly.

Thanks,

-Philip