Porting GPA to GPGME
Marcus Brinkmann
Marcus.Brinkmann@ruhr-uni-bochum.de
Sun Aug 11 03:02:02 2002
On Sat, Aug 10, 2002 at 09:15:37PM +0200, Miguel Coca wrote:
> I'm porting GPA to use GPGME.
Sounds good :)
> This is giving me problems:
>
> * The signatures page doesn't work at all. I can't figure out how to
> list the signatures on a key with gpgme.
This is currently unsupported. Seems we need to do some work here.
> I thought trust items would
> help with that, but I can't find out what they actually do :-)
> (every key that I've tried gives me an empty list of them) Any
> suggestions?
Possibly. The trust item interface is experimental, but the main problem
right now is that it is also unsupported by GnuPG. There was an
implementation for --list-trust-path in 1.81.2.20, but it went away when
the trustdb code was revamped in 1.81.2.21. Werner, was that a temporary
thing or what do you envision for listing the signatures on a key?
> * Exporting secret keys does not work, and I don't think it can be
> done with gpgme.
It's arguably a bad thing to give out the secret key :) Sometimes it might
not even be possible, like if it is stored on a smart card or otherwise
managed by gpg-agent.
> Again, any suggestions? This also affects making
> backups of secret keys.
Maybe Werner can make a comment on if it should be allowed to attempt that
in GPGME.
> * Deleting keys should work as soon as that is fixed in gpgme
> (according to the gnupg-devel archives, that's a known bug). Does
> anyone know when that will be solved?
You need this change in GnuPG HEAD, will be in 1.2 I think:
* delkey.c (do_delete_key): Always allow to delete a key in batch mode
when specified by fingerprint. Suggested by Enzo Michelangeli.
> Mostly the code is keeping it's original structure. While gpgme has some
> features that could be used to improve GPA, I think it's better to have
> something that works first, and then improve on it.
I also expect that we have to make some changes/improvements to GPGME.
> To assist during the migration, I've built a dummy gpapa with all the
> definitions and functions, but whose functions do nothing. That way we can
> compile the program even when it's a mix of both API's without touching the
> real gpapa, which probably has some code that will be needed later on. I'll
> commit them with the rest.
If the CVS version represents something I can compile and run, I would make
sure that I always follow your changes and set up a test environment, so I
can easier reproduce eventual problems with GPGME etc. Just drop us a note
when you feel happy about the state of the repository for this type of
testing.
Thanks,
Marcus
--
`Rhubarb is no Egyptian god.' GNU http://www.gnu.org marcus@gnu.org
Marcus Brinkmann The Hurd http://www.gnu.org/software/hurd/
Marcus.Brinkmann@ruhr-uni-bochum.de
http://www.marcus-brinkmann.de/