[bug-mailutils] Parsing 'Received:' header into components

[Robby Villegas]

I want to (try valiantly to) decompose the value of a Received line
into useful components.  I was about to embark on writing a parser
using the functions in parse822.h, but I figure someone else must have
attacked the Received line and written the code.  I don't want to
reinvent this wheel!


Take the following Received header as an example:

Received: from somewhere.com (mailhub.somewhere.com [10.20.30.40])
        by home1.somewhere.com (8.12.11/8.12.11) with ESMTP id j2CKbxYX012911
        for <address@hidden>; Sat, 12 Mar 2005 14:37:59 -0600

I think my top level groups will be prefixed by these five words:

  {from, by, via, with, for}

since they seem to give overall routing information.
The sixth element will be date/time.


Within each group, I want to parse the usual pieces as much as I can. 
To illustrate, here are three of those group broken into nested lists:

with ESMTP id j2CKbxYX012911 ->
   {ESMTP, id, j2CKbxYX012911}

Sat, 12 Mar 2005 14:37:59 -0600 ->
   { {Sat, 12, Mar, 2005}, {{14, 37, 59}, -0600} }

from somewhere.com (mailhub.somewhere.com [10.20.30.40]) ->
   { from, {somewhere, com}, {{mailhub, somewhere, com}, {10, 20, 30, 40}} }


I gather from a smidgen of reading and some spam I've looked at, that
parsing this field is troublesome.  The structure may not be as
regular as I outlined above.  Especially when it comes from spammers. 
But if I can destructure 95% of cases with a small number of
heuristics, that would be very useful.

Do you know of existing code to do this, or should I embark on it?

Thanks,
Robby Villegas


P.S.  Here's a less standard piece in a comment which is interesting:

(version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NOT); ->
   {{version, TLSv1/SSLv3}, {cipher, DHE-RSA-AES256-SHA}, {bits, 256},
{verify, NOT}}