Re: [Evolution] gpg help

[Not Zed <notzed ximian com>]

On Mon, 2002-06-03 at 16:20, Adrian 'Dagurashibanipal' von Bidder wrote:
> On Mon, 2002-06-03 at 05:16, Jeffrey Stedfast wrote:
> > There were some bugs in the 1.0.x PGP/MIME code, these should all be
> > fixed now in the development CVS. Perhaps you came accros one of these
> > bugs?
>
> *tongue hanging out*
>
> when will this be in a release? Is it really true? All known gpg bugs
> fixed?
>
> Think I'll have to try the cvs...

I dunno about all, but some more have been fixed.  And only for
multipart/signed (i dont ever want to support inline pgp, its even more
broken).  However i ran some tests, and i could generate mails that
verified in evolution which didn't in mutt; but that was definetly a
mutt bug (it was dropping a trailing \n, but only for some
evolution-generated mails).  evolution verified everything i created in
either though.

The multipart/signed rfc's are broken, they break valid assumptions you
can make about mime (which is a _transport_ mechanism), so they could
simplify their implementation.  But there's no guarantee you can ever
get a valid signature, even if the content isn't changed.

So even if evolution was 'perfect' in its pgp/mime implementation, there
are bugs in the design of multipart/signed which can cause failure (esp
if for example any mailer blows apart mime parts and stores them
decoded, which imho is a perfectly valid thing to want to do).  But
then, I guess it depends on what you expect from a signed message. 
s/mime's encapsulated objects get around this problem, but with the side
effect that you can't read the messages without an s/mime aware mailer.

But anyway, the 1.0.x code did have a couple of serious bugs which would
mean certain types of message (basically anything ending in blank
lines?) would never verify, the 1.1.x (1.2) code fixes this and some
other issues.