Re: [Wicket-user] Server Push (COMET / Reverse Ajax)

2006-08-26 Thread Lorin Metzger
Lorin Metzger wrote: Eelco Hillenius wrote: The problem that I had was that it seemed like all other requests for that session were also blocked even ones that had nothing to do with that behavior. You would have to make sure the request for a push behavior is

Re: [Wicket-user] Server Push (COMET / Reverse Ajax)

2006-08-26 Thread Igor Vaynberg
the problem with the approach is that a behavior is tired to a component that is tied to a page. any thread accessing the page needs to block because you dont want two or more theads working on the page.i think you need to take a bit of a different approach have a threadsafe queue of push events

Re: [Wicket-user] Server Push (COMET / Reverse Ajax)

2006-08-25 Thread Lorin Metzger
Eelco Hillenius wrote: Hi, COMET is not implemented as a core feature at this time. Do you have a specific use case where you would like to use it for? That might be a good start to investigate the options. One use case would be, a application with a small group of users. The users

Re: [Wicket-user] Server Push (COMET / Reverse Ajax)

2006-08-25 Thread Igor Vaynberg
i might be totally missing something - but dont you need a special servlet container that maintains open socket separate from threads by using nio? like jetty 6? and then once you have that you need to use that servlet container's special api to do this? with the current architecture if you have a

Re: [Wicket-user] Server Push (COMET / Reverse Ajax)

2006-08-25 Thread Eelco Hillenius
On 8/25/06, Igor Vaynberg [EMAIL PROTECTED] wrote: i might be totally missing something - but dont you need a special servlet container that maintains open socket separate from threads by using nio? like jetty 6? and then once you have that you need to use that servlet container's special api

Re: [Wicket-user] Server Push (COMET / Reverse Ajax)

2006-08-25 Thread Janne Hietamäki
On 25.8.2006, at 19.43, Eelco Hillenius wrote: with the current architecture if you have a hundred clients doing comet you will have a hundred threads - not so good. Yeah. What do you think Lorin? Is there something we might have missed from e.g. DWR? An alternative for your use case

Re: [Wicket-user] Server Push (COMET / Reverse Ajax)

2006-08-25 Thread Lorin Metzger
Eelco Hillenius wrote: On 8/25/06, Igor Vaynberg [EMAIL PROTECTED] wrote: i might be totally missing something - but dont you need a special servlet container that maintains open socket separate from threads by using nio? like jetty 6? and then once you have that you need to use

Re: [Wicket-user] Server Push (COMET / Reverse Ajax)

2006-08-25 Thread Eelco Hillenius
I like the idea of the AjaxPushBehavior mentioned in the issue tracker posting. At one point I tried extending AbstractDefaultAjaxBehavior, and adding an addOnLoadModifier() which asynchrously called back on page load and then I blocked that thread, with the intention of waking up that

Re: [Wicket-user] Server Push (COMET / Reverse Ajax)

2006-08-25 Thread Lorin Metzger
Eelco Hillenius wrote: The problem that I had was that it seemed like all other requests for that session were also blocked even ones that had nothing to do with that behavior. You would have to make sure the request for a push behavior is handled by it's own

[Wicket-user] Server Push (COMET / Reverse Ajax)

2006-08-24 Thread Lorin Metzger
Hi, I was searching the mailing lists and noticed a few comments about Reverse Ajax and COMET. Mainly jokes about who would ask about it first :) . Has anybody made any serious attempts to add DWR like reverse-ajax support to wicket? If not does anybody have any suggestions on what the

Re: [Wicket-user] Server Push (COMET / Reverse Ajax)

2006-08-24 Thread Eelco Hillenius
Hi, COMET is not implemented as a core feature at this time. I think there is a feature request open for such support with a specific implementation for Jetty. And maybe some users have created something COMET like. Do you have a specific use case where you would like to use it for? That might