gpgme 0.3.8: 3 questions, 2 bugs & 1 request

[Werner Koch]

On Wed, 24 Jul 2002 21:52:15 +0200, Stéphane Corthésy said:

> - What are the IO callbacks used for? In which cases should I use them?

This can be used to integrate gpgme into the big select loop most GUI
frameworks use, so that you can use gpgme asynchronously.

This needs more documentation.

> - What is the GpgmeDataEncoding used for? In which cases should I use it?

The backend, either gpg or gpgsm, usually tries to figure the encoding
of the input data itself.  This is not 100% reliable so gpgsm has a
feature to allow to specifiy the input encoding.  It is useful if you
have just the base64 encoding which can't be detected automatically.

> - What are the following new key attributes, GPGME_ATTR_SERIAL,
> GPGME_ATTR_ISSUER, GPGME_ATTR_CHAINID, GPGME_ATTR_ERRTOK? Are they
> used only for S/MIME keys?

Yes. ERRTOK is a way to get more detailed error information but it is
not yet used with the gpg backend.  CHAINID is used with S/MIME to
build the certification chain; it is currently used by Mutt and
Kmail's certificate display code.  ISSUER and SERIAL are the
respecitive X.509 attributes.

Eventually there will be example code on how to use gpgme in a MUA.

> Now, two bugs:
> - The new function gpgme_op_keylist_ext_start() doesn't behave the
> same way as gpgme_op_keylist_start() when we pass it only one

Marcus will look at this.

> Finally, here's a request: I need to be able to set the --charset
> option when performing an operation; currently, gpgme doesn't set it,

The GPGME interface takes all strings as UTF8, so any conversion
should be done by the caller.  However we might have missed to pass
--charset utf-8 to gpg.  Marcus can you please check this?


Shalom-Salam,

   Werner