Re: Verifying inline gpg signatures

[Simon Josefsson <jas@extundo.com>]

Jesper Harder <harder@myrealbox.com> writes:

> Jesper Harder <harder@myrealbox.com> writes:
>
> [We fail to detect inline PGP with format=flowed]
>
>> Simon Josefsson <jas@extundo.com> writes:
>>
>>> Perhaps mm-uu-dissect should be run on all text/plain parts?  Or all
>>> text/.* parts?  Opinions?
>>
>> It's a step in the right direction ... but not enough.  The problem is
>> that text/plain parts with a format parameter haven't been QP decoded
>> at this stage, so uu-dissect will fail.
>
> But I suppose making it work for non-QP encoded articles is better than
> nothing?

Unless it breaks something...

> I first assumed that it wouldn't be useful because Gnus always QP
> encodes inline PGP, so it didn't work when I tested it with a message
> from Gnus.  But the Mozilla plugin which probably generates all of the
> problematic messages doesn't do it for us-ascii, so it works in that
> case.
>
> Should we do it?

IMHO the ideal seems to be to run the mm-uu stuff both before and
after MIME processing.  Perhaps even iteratively until no more actions
can be performed.  Of course, it would have to be careful to not
infloop or some other silly thing.  This way we could detect all UU
things (shar archives, gnats, diff, binhex, yenc pgp, etc) no matter
how deep.  Consider a postscript file binhex encoded and PGP encrypted
and then PGP signed.