gpgme 0.3.3 questions

[Stephane Corthesy]

> > - Which GpgmeData types can be read/written?
>
> All.

How is it possible to write GpgmeData initialized with  
gpgme_data_new_with_read_cb(), with gpgme_data_new_from_file(), with  
gpgme_data_new_from_filepart() ?
According to gpgme_data_new_from_mem() signature, buffer is  
non-writable (const char *), thus this type should be non-writable if  
copy is 0.


> > What's the point of returning 0, without being at EOF?
>
> It might choose to interrupt the read operation instead of waiting for
> more data.

I still don't understand why it would do this. Would it return an  
error code?
If I ask for length, how can I sure that it returned me the real  
length (up to EOF)?

/**
 * gpgme_data_rewind:
 * @dh: the data object
 *
 * Prepare the data object in a way, that a gpgme_data_read() does start
 * at the beginning of the data.  This has to be done for all types
 * of data objects.
 *
 * Return value: An error code or 0 on success
 **/

Do we really need to call rewind before ever trying to read a data?  
If we just inited the data, we don't, do we?

In GpgmeRecipients, would it be possible to provide a function which  
would return the validity assigned to a name contained in the  
GpgmeRecipients instance?

BTW, here's another bug: it it not possible to retrieve fingerprints  
for subkeys

Stephane