a few GPGME issues

[Werner Koch]

On Tue, 23 Apr 2002 00:02:31 +0200, Marcus Brinkmann said:

> We can think about rectifying this when we change/extend the interface to
> supply the individual data in a useful format (userid etc).

I did the passphrase callback has a temporary hack to until we can use
the gpg-agent.  It later turned out that it will be needed for
symmetric-only encryption, but in that case a user ID is not required.
It does only make sense to use this callabck in an automated
environment, so the only thing which is required is a caller supplied
pointer to be passed transparently to the callback.

  Werner