On May 28, 2010, at 5:41 PM, Darryl L. Pierce wrote:
On Fri, May 28, 2010 at 11:24:46AM +0200, Andrew Beekhof wrote:
>> const char* function(const char *);
>
> Not sure about this method signature.
> It would mean each function has to do its own packing and unpacking which might lead
to excessive duplication.
> It would also cause problems for API calls like "get the paste buffer"
which could easily contain images (and thus zero'd bytes which would confuse
strlen()).
Yeah, I was trying to keep things simple. Perhaps better would be to
pass around just a void * and let each API know for itself what to
extract. My main goal there was just to avoid having the Multiplexer dig
into the XML and try to map values and types to specific method
arguments.
It should at least parse the string into an xmlNode/xmlDoc for the API call to work with
it easily.
So I'd not be passing around void*.
> Since the input is xml, perhaps xmlNodePtr is a better I/O vector.
>
> Other than that, the patch is fine.
> I'd go ahead and commit it and we can continue discussing the method signature.
Agreed.
This is pushed now.
--
Darryl L. Pierce, Sr. Software Engineer @ Red Hat, Inc.
Delivering value year after year.
Red Hat ranks #1 in value among software vendors.
http://www.redhat.com/promo/vendor/