up2date and XML-RPC GET

[Adrian Likins]

On Wed, Oct 13, 2004 at 12:03:33PM +1000, Alan Milligan wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> It seems that someone has recently decided that GET requests for up2date
> ~ make a lot of sense.
>
	recently being about ~3 year ago
 
> While this is both more efficient in terms of having to read the request
> ~ body and having to xml parse it, unfortunately it is not part of the
> XML-RPC protocol.

	Indeed, GET's are not part of the xml-rpc protocol.

> 
> According to Dave Winer's spec at http://www.xmlrpc.com/spec, and to
> everyone else, XML-RPC is a HTTP-POST protocol.
	
	yup

> 
> I appreciate that XML-RPC's file limitations already require up2date to
> return non-XML-RPC payloads, but wonton undermining of these protocols
> leads us down the path of Microsoft.
> 
	Not sure I understand this leap. up2date uses XML-RPC for
rpc style stuff, and standard http GET's for downloads. Theres
no "XML-RPC over http GET". The internal api's for xmlrpc
requests and GET's may look similar, but this is do to
some code abstraction and not a blurring of the protocols.
We use some data marshalled into a format similar to XML-RPC
in some cases, but that is not XML-RPC. 

	And for whats its worth, up2date in fedora doesn't
use XML-RPC at all. Thats only used for communcation with
Red Hat Network servers. So this may be a bit off topic
for this list. 


> Can we agree that up2date was not designed and developed exclusively for
> RedHat environments, and that as such, it's in all of our interests to
> keep it durable.
> 
	I'd like to agree with that, but realistically, up2date
was designed for Red Hat Linux/Fedora Core/Red Hat Enterprise Linux.
I've never intentionally written something to make it imcompatiable
with other systems, but I've never released any code to make it
particularlly cross platform either. Red Hat distros are the
target platform. 

> I can provide you with the necessary patch to fix this if required, but
> I'd first like to hear whether or not you'd apply it.
> 
	What is the "this" in the above statement? I'm not sure
I understand what exactly it is that needs to be fixed. 

Adrian