Re: [Evolution-hackers] S/MIME support

[Jeffrey Stedfast <fejj ximian com>]

On Thu, 2003-07-03 at 11:41, Not Zed wrote:
> Hi Bernard,
> 
> > I'm interested in adding support for S/MIME to Evolution and have seen a
> > number of requests/mentions of this feature.
> > 
> > Is anyone currently working on it?  If not has anyone tried and given up (war
> > stories welcome!)?  Does anyone have any suggestions on how I can get started
> > on such a task?  Where could I find the relevant documentation on how to get
> > started with hacking evolution?
> 
> We're planning it for 1.6 i think.  Jeff will be working on it, he'll
> probably reply when he wakes up :)
> 
> He's already done some (nontrivial) work on it in the past, but i'm not
> sure whether it managed to do anything or was just experimentation.
> 
> There is no documentation, apart from the source.  There is a basic
> object and class structure for high-level
> encrypt/decrypt/signing/verification, and an implementation for gpg, but
> thats about it.  We want to keep similar api's and a common abstraction
> at least for these high-level operations for different backends (for
> plugins?).
> 
> Source to look at (in an evolution tarball or on cvs.gnome.org in the
> evolution dir) ...
> 
> Stuff in camel/*
>   camel-cipher-context.[ch]
> 	abstract high-level class for use by the mime classes
>   camel-gpg-context.[ch]
> 	implementation for gpg/pgp
>   camel-multipart-signed.[ch]
> 	handle reading/writing multipart/signed type (rfc3156)
> 	ignore most of my whiney comments :)

(abstractly defined in rfc1847 - applies to both S/MIME and PGP/MIME)

>   camel-multipart-encrypted.[ch]
> 	handle reading/writing multipart/encrypted type

unfortunately not helpful for S/MIME, since it does encryption a
different way :-(

> 
> this stuff is there but doesn't work/isn't used
>   camel-cms-context.[ch]
> 	similar to cipher-context for s/mime

right, altho ideally what should happen is that cms-context and
cipher-context get merged.

A few things:

CamelCipherValidity will have to become some sort of chain if it is
going to be used by the S/MIME code (since, as far as I understand, the
Mozilla nss verification code uses a chain that might not be easy to
flatten. or, even if it can be flattened - should it?)

>   camel-smime-context.[ch]
> 	similar to gpg context

the basics are there, nothing tested tho. I don't even completely recall
where exactly I left off.

>   camel-pkcs7-context.[ch]
> 	pkcs7 for cipher-context implementation?

this is garbage. When I originally started implementing S/MIME, the APIs
were different in nss.

> 
> There's also some utility stuff, but its mostly for identifying/etc. 
> The frontend code just consists of calling *multipart* methods with the
> right context.
> 
> The plan (i think a requirement?) is to use Mozilla's libNSS to
> implement it.

right.

-- 
Jeffrey Stedfast
Evolution Hacker - Ximian, Inc.
fejj ximian com  - www.ximian.com