Bug in lexer_v3

[Matthias Andree]

Evgeny Kotsuba <evgen at shatura.laser.ru> writes:

> Hi,
>
> I have just foud bug that seems is present  throw as min v 0.15.8

>
> If mail is in form
> [any header lines]
> Content-Type: message/partial;
> [some header lines]
>
> Date: blabla
> From: blabla
>
> say  simply:
> ------1.msg:  -8<---------------
> Content-Type: message/partial;
>
> Date: blabla
> From: blabla
> -------8<---------------

Yes. RFC-2046 section 5.2.2 (message/partial fragmentation/reassembly)
is not yet implemented. I'll put it on TODO.

Bogofilter currently does tag headers for embedded MIME headers in
message/rfc822 either, and I do not think it should in the latter
case.

For message/partial, it's a bit more complex, see the RFC.  Basically,
we should take all headers from the first part's outer header except
Content-*, Subject, Message-ID, Encrypted, MIME-Version which are taken
from the first part's inner header. The inner and outer headers of
subsequent parts are discarded.

I am a bit displeased about the many places that we store state in
lexer_v3.l. Anything except the MIME stack should move into flex
"BEGIN" states.

-- 
Matthias Andree