gpgme: pending not being cleaned up after keylist retrieval?

Marcus Brinkmann Marcus.Brinkmann at
Wed Dec 5 11:49:01 CET 2001

On Tue, Dec 04, 2001 at 05:08:10PM -0800, Charles Duffy wrote:
> In gpgme 0.2.3, gpgme_op_keylist_start exits succesfully with
> c->pending=1. gpgme_keylist_next never cleans this up, even after
> returtning with GPGME_EOF.
> This is preventing me from doing a gpgme_op_sign (which has a
> fail_on_pending_request() call) after using the keylist functions, which
> is rather problematic.

Thanks for the report.

> Suggestions? (Please cc me in any reply)

Well, at EOF it should set pending to 0, this is a good suggestion.
However, if you just look for a certain key you don't even want
to bother to read the remaining keys, and I think that we could provide a
new function to abort/cancel any pending request.  This would be generic
(you could also use it to cancel a started sign operation etc).

gpgme_cancel is unfortunately taken for some other purpose.
I am thinking about gpgme_op_cancel (confusing?) or gpgme_op_abort
or gpgme_abort.  Comments welcome.

I don't see how thi could be solved any other way.


More information about the Gnupg-devel mailing list