[Standards] stream restarts

Peter Saint-Andre stpeter at stpeter.im
Thu May 8 20:03:37 UTC 2008


On 05/07/2008 11:55 AM, Justin Karneges wrote:
> On Wednesday 07 May 2008 10:42 am, Tony Finch wrote:
>> On Tue, 6 May 2008, Alexander Gnauck wrote:
>>> * this saves us some round trips
>> No it doesn't, if you pipeline requests and responses. The spec does not
>> forbid this. For an example of extreme pipelining, see
>> http://mail.jabber.org/pipermail/standards/2008-February/017934.html
>> though in practice you probably don't want to be that aggressive /
>> optimistic about pipelining around a request that might fail. Stream
>> restarts can't fail, so they can be pipelined.
> 
> Yes, I would like to see pipelining addressed in the spec.  Right now it is 
> unclear if this technique is legal.  The spec does not forbid it, but I don't 
> think that is enough, is it?
> 
> I'm also not sure what parts of XMPP could be declared as pipelinable.  
> Certainly the opening <stream> exchange, but I don't know if the rest can be 
> done this way, given the asynchronous nature of elements exchanged in the XML 
> stream.

If we defined pipelining, we'd need to specify all that. I'm not yet
sure if that's something we want to pursue. Heck I brought up the stream
restart stuff just to see if it's something we want to pursue, and the
consensus so far seems to be that we have no consensus to so do (unlike
say the consensus for roster synchronization).

P

-- 
Peter Saint-Andre
https://stpeter.im/

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/x-pkcs7-signature
Size: 7338 bytes
Desc: S/MIME Cryptographic Signature
URL: <http://mail.jabber.org/pipermail/standards/attachments/20080508/12ebe47c/attachment.bin>


More information about the Standards mailing list