2.6.x goals?

[Simon Josefsson]

Daniel Kahn Gillmor <dkg-debian.org at fifthhorseman.net> writes:

> On Wed 2008-06-18 12:07:13 -0400, Simon Josefsson wrote:
>
>> Btw, my goals for 2.5.x is to integrate Nikos' crypto.h stuff, and
>> to re-indent the code.  As far as I know, there are no other tasks
>> that are pending, or are there?  Anything major to be included
>> should be made available on a separate branch for review relatively
>> soon, or it will have to wait for 2.7.x.
>
> I have three OpenPGP-related goals i'd like to see met for 2.6:
>
>  * the ability to deal with passphrase-encrypted OpenPGP keys.
>
>  * the ability to verify a specific UserID in an OpenPGP certificate
>    without reference to any other UserIDs in the certificate.
>
>  * the ability to create new OpenPGP keys (e.g. with certtool) and to
>    add UserIDs and subkeys.

Sounds fine to me.

> I don't know if these are already part of the MPI rewrite that you
> mention: are the details for that documented someplace?

Nope, the MPI stuff is the gnutls_with_ext_mpi branch:

http://git.savannah.gnu.org/gitweb/?p=gnutls.git;a=shortlog;h=refs/heads/gnutls_with_ext_mpi

It allows modular replacement of libgcrypt.

> I'd be happy to contribute toward this as well, if i can find the time
> and can figure it out.  I don't want to step on any toes, though, and
> i don't want to create a lot of extra work if there are larger-scale
> rewrites of the relevant sections going on concurrently.  Any thoughts
> on how i should approach this?

The best way to get started is to create a branch and add your functions
to that branch and ask me to merge it when you think it is stable.  I'm
not aware of anything large-scale work being done in this area.  The
only one doing major work on the OpenPGP stuff is Nikos, so he can tell
for sure.

All development discussion happens on this list, so if there is nothing
on a particular topic here it means nobody has done anything beyond
possibly thinking about it.

> I'd also be interested in developing the more-compact OpenPGP function
> space that was just discussed on this list.  Guidance as to how to do
> that cleanly would be appreciated.

I think the simplest would be if you propose the API you are thinking
of, and ask whether people like it better.  It may be useful for you to
review the GPGME stuff and decide whether you think it is possible to do
closer integration or compatibility with it or not.  Nikos may want to
chime in here too.

/Simon