[JDEV] 0.9 Release Candidate

[Thomas Charron]

*Yes, I am indeed ALIVE!!*
Quoting Jeremie <jeremie at jabber.org>:
>   * etherxd is gone, all of the etherx functionality is now in libetherx
> and starts with the server (you just start jserver)

  While etherxd is gone, looking at the process, one could create an etherxd by 
creating a 'wrapper' of some sort, correct?  On that note, etherx, from what 
I'm observing, still starts, but has been absorbed completely into the library, 
and will pretty much start anyway by whatever transport happens to run first.  
Is this correct?

  I'm asking primarily for scaling concerns, as etherx pretty much now is a pth 
thread, and hence, jabber-transport will now, as the first transport started, 
be even more locked down to one processor.

  What I've ended up hacking out was creating a small 'admin' transport that 
does nothing but start etherx, and handle giving statistics on the locally 
running machine by request.  I'm using etherx and jabber for something beside's 
user to user messaging, utilizing XML namespaces for passing such things as RDF 
fragments around.

  One of the features that I've been looking to add to etherx as part of this, 
and have, in part, done, is a message queue, storing undeliverable streams to 
disk to reattempt delivery of them several times untill returning the error 
packet back.  In my case, by using a simple transport to get etherx running, I 
can ensure proper permisions, etc, are in place so that etherx would itself 
have proper read/write rights, etc, to it's espool directory, but one of the 
additional transports might not.  Perhaps in the future some sort of flag could 
be set to say 'don't start an etherx thread if something else hasn't already'?

--- 
Thomas Charron
<< Wanted: One decent sig >>
<< Preferably litle used  >>
<< and stored in garage.  ?>>