On 1/25/09 2:12 PM, Jack Moffitt wrote:
>> assumptions about the client implementations. Still, I am not sure if this
>> behavior is the best for the requests outside the defined window, e.g.
>> (lastRid + 10) while 'requests'==2.
> 
> RIDs outside of the window should result in an immediate 404 just like
> the spec says.  The window value we generally work with in our clients
> is 5 or so.  There is room in that window to resequence a few stray
> requests.

It would be great to capture this kind of information at
<http://xmpp.org/tech/bosh.shtml>. In fact I think we need things like
/tech/bosh/overview.shtml and /tech/bosh/implementations.shtml and
/tech/bosh/tutorial.shtml and /tech/bosh/developerguide.shtml -- any
volunteers to help work on such pages? I'm happy to give out SVN access.

>> If these errors usually happen, why don't we consider changing the XEP and
>> making client side cache-resend obligatory?
> 
> I assume it's not required because certain devices may not be able to
> cache because of memory requirements.

Right. Many "features" of BOSH are there to handle constrained client
environments.

>> How does your implementation perform with other hold values? The server may
>> dictate another hold value at the session creation response.
> 
> Strophe doesn't allow arbitrary hold values right now (mostly because
> it doesn't make much sense in a browser), but the logic could easily
> be applied to hold > 2.

See above about a developer guide.

Peter

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

Attachment: smime.p7s
Description: S/MIME Cryptographic Signature

Reply via email to