lexer, tokens, and content-types

[David Relson]

At 01:40 PM 12/8/02, Matthias Andree wrote:

>David Relson <relson at osagesoftware.com> writes:
>
> > they belong there.  With the addition of a content-type module, we'll
> > have a structure better suited for handling mime, base64, uuencode, etc,
> > etc.
>
>I've had a very short glance at EPS and while flawfinder and RATS were
>rather unspectacular (haven't audited the code though), though I spotted
>one bug already:
>
>EPS lacks support for the "binary" transfer encoding as per RFC-2047,
>and lists a "raw" transfer encoding that is unlisted on
>http://iana.netnod.se/assignments/transfer-encodings, so if at all, it
>would have been x-raw. Not sure where they picked that "raw" up. Maybe a
>thinko.
>
>Other than that, it uses a nice iterator-style approach to parse the
>mail, which may turn out useful, but it eats boundaries.
>
>We'll have to look more closely to figure if EPS is suitable for
>bogofilter.
>
>I'm adding eps at inter7.com to the To: header to let them know there's a
>bug in content.c.

Matthias,

I've been reading their docs and looking at their sample programs.  Their 
approach seems reasonable.  It sounds as though you've gotten deeper into 
the code than I have, so far.

As I'm reading, I'm also proofreading and correcting.  I'll send them a 
corrective patch when I'm done.

If we do go ahead with eps, I'm sure we'll find other details to fix.  'Tis 
also known as "life in the fast lane" :-)

David