On 10/25/07, Michael Sparer <[EMAIL PROTECTED]> wrote:
>
>
> Hi Xavier,
>
> thanks for your effort but please don't take too much time investigating
> my
> issues :-)
>
> I'm using wicket 1.3.0-beta3 and wicketstuff push... well since it doesn't
> show the svn version numbers the version of monday this week.
This is very similar to what I use (with a new version of wicketstuff-push I
pushed on tuesday I think, which fixes a problem with backward button). But
if you use the dojo based implementation, I can't speak much about it since
Vincent did all the job.
Yesterday I
> had a more thorough look into what's exactly happening i.e. how the push
> project interacts with the wicketstuff dojo project and integrates cometd.
> There I found out that in order to communicate with wicket there's an
> extra
> round trip to the server taking place. So the channel sends a message to
> the
> client-javascript (including the data to be shown on page) which then
> makes
> an ajax call back to the server where the data gets rendered into xml and
> is
> then shown on the page. I wondered if that was intentional and if yes,
> what
> purpose it serves.
>From my understanding (but I didn't develop this) the problem is that it's
not straightforward to render a wicket component outside a web request, so
the additional cycle is there to call the rendering in a "normal" web
request: comet is only used to detect there is something to refresh, then
the refresh is done with a usual wicket ajax call. But I may be wrong...
For my project I've changed the java-script function which starts the AJAX
> call sothat it renderes the message to proper javascript and directly
> inserts it in the document's DOM without making an additional call. This
> however, makes the server site less flexible as only a javascript string
> can
> be processed on client site - but that's ok for my app.
Indeed if you don't need component rendering reducing the cycles is better.
I had plan to work on a comet based implementation of IPushService some time
ago, but I've never found the time... If I do one day I'll try to reduce the
cycles to as less as possible, and I may also have to review some component
refresing concepts, like refreshing a RepeatingView when only one line has
been added. The only easy way to do this for the moment is to add the
component to the target request, but then the whole component is rendered
and sent to the client. Having something finer would be nice. But that's
only ideas, and I'm really lacking of time.
Xavier
Xavier Hanin wrote:
> >
> > Michael,
> >
> > As I said I don't use the ChannelService myself, but I used to have
> > similar
> > troubles with the PushService, which I think I finally fixed, but it
> > wasn't
> > easy. Since the fix is in TimerChannelBehavior (which is used by both
> the
> > push and channel timer based implementation, it should work for both,
> but
> > I
> > haven't tested. Can I ask which version of wicket and wicketstuff-push
> > you're using, and also which implementation of IChannelService? If you
> > want
> > to investigate in the code, the core of the fix relies on the redirect
> and
> > setRedirectId in TimerChannelBehavior.
> >
> > Xavier
> >
> > On 10/24/07, Michael Sparer <[EMAIL PROTECTED]> wrote:
> >>
> >>
> >> Hi Xavier,
> >>
> >> first of all thank you for the effort you've done for the push-project.
> >> Up
> >> to now it really provided the features I need for my app. But now I'm
> >> kind
> >> of stuck. My problem is: I have two panels on one page. Both panel
> >> subscribe
> >> themselves as ChannelListener for the same channel. But unfortunately
> >> whenever an event invoking the listeners occurred, the onEvent method
> >> gets
> >> executed twice on both listeners (i.e. four times). It occurrs only if
> >> you
> >> have two panels i.e. two listeners on the same page.
> >>
> >> IMO the error is somewhere in the javascript where a function gets
> >> executed
> >> twice...but I haven't found it so far. So do you have any clue?
> >>
> >> thanks
> >>
> >> Michael
> >>
> >> Xavier Hanin wrote:
> >> >
> >> > On 10/23/07, Michael Sparer <[EMAIL PROTECTED]> wrote:
> >> >>
> >> >>
> >> >> Hi Xavier,
> >> >>
> >> >> thanks for your reply - well yes, that's certainly right. I've just
> >> lost
> >> >> sight of it. But I got another question: how stable is the
> >> >> wicketstuff-push
> >> >> project or what sections are stable and which have to be improved
> for
> >> use
> >> >> in
> >> >> a professional webapp?
> >> >
> >> >
> >> > AFAIK wicketstuff-push is developed by vincent demay and myself.
> >> Vincent
> >> > worked on the cometd channel implementation, and I've worked mostly
> on
> >> the
> >> > timer based IPushService implementation (where events do not come
> from
> >> the
> >> > client at all). I don't know how stable is the channel implementation
> >> > since
> >> > I don't use it myself, maybe Vincent could give more details (not
> sure
> >> if
> >> > it's used in production or not). For the IPushService implementation,
> >> it
> >> > still requires more testing and improvements, because the problem is
> >> not
> >> > easy to solve with a timer based polling implementation: we have to
> >> detect
> >> > clients disconnection, and also handle back button where pages go
> back
> >> to
> >> > life from the a serialized form. And we also need to work on a comet
> >> based
> >> > implementation of this push service. So I think there's still need
> for
> >> > improvement and bug fixing in this section.
> >> >
> >> > BTW, I've just checked in an improvement about this problem of page
> >> > deserialization which required an API change for the IPushService. If
> >> you
> >> > plan to use it, do not forget to do an svn update.
> >> >
> >> > Xavier
> >> >
> >> > thanks in advance
> >> >>
> >> >> Michael
> >> >>
> >> >> Xavier Hanin wrote:
> >> >> >
> >> >> > On 10/22/07, Michael Sparer <[EMAIL PROTECTED]> wrote:
> >> >> >>
> >> >> >>
> >> >> >> Hi,
> >> >> >>
> >> >> >> today I wondered if there's a wicket-approach for pushing
> messages
> >> >> from
> >> >> >> the
> >> >> >> server to the client (also called reverse ajax or pushlets). Well
> >> yes,
> >> >> >> there
> >> >> >> is one called wicketstuff push. I looked at its examples and soon
> >> >> managed
> >> >> >> to
> >> >> >> implement my own pushing stuff. In the examples (chat-example)
> >> >> however,
> >> >> >> the
> >> >> >> ChannelService that delivers the messages to its listeners, is
> >> stored
> >> >> in
> >> >> >> the
> >> >> >> application. I.e. there is only one channel for the whole
> >> application.
> >> >> >> Now imagine I want to create an additional private-chat-room for
> >> two
> >> >> >> users
> >> >> >> --> I'd need a ChannelService only for two users. So my question
> >> is:
> >> >> >> what's
> >> >> >> the wicket way to pass and store such an object which is
> applicable
> >> >> only
> >> >> >> for
> >> >> >> two users (sessions)?
> >> >> >
> >> >> >
> >> >> > You can use an application wide ChannelService and use the channel
> >> name
> >> >> > ("chat/message" in the example) to isolate messages by
> >> >> private-chat-room.
> >> >> >
> >> >> > Xavier
> >> >> >
> >> >> > Thanks in advance
> >> >> >>
> >> >> >> Michael
> >> >> >>
> >> >> >> --
> >> >> >> View this message in context:
> >> >> >>
> >> >>
> >>
> http://www.nabble.com/wicketstuff-push-and-sharing-an-IChannelService-tf4671463.html#a13345278
> >> >> >> Sent from the Wicket - User mailing list archive at Nabble.com.
> >> >> >>
> >> >> >>
> >> >> >>
> >> ---------------------------------------------------------------------
> >> >> >> To unsubscribe, e-mail: [EMAIL PROTECTED]
> >> >> >> For additional commands, e-mail: [EMAIL PROTECTED]
> >> >> >>
> >> >> >>
> >> >> >
> >> >> >
> >> >> > --
> >> >> > Xavier Hanin - Independent Java Consultant
> >> >> > http://xhab.blogspot.com/
> >> >> > http://ant.apache.org/ivy/
> >> >> > http://www.xoocode.org/
> >> >> >
> >> >> >
> >> >>
> >> >> --
> >> >> View this message in context:
> >> >>
> >>
> http://www.nabble.com/wicketstuff-push-and-sharing-an-IChannelService-tf4671463.html#a13358479
> >> >> Sent from the Wicket - User mailing list archive at Nabble.com.
> >> >>
> >> >>
> >> >>
> ---------------------------------------------------------------------
> >> >> To unsubscribe, e-mail: [EMAIL PROTECTED]
> >> >> For additional commands, e-mail: [EMAIL PROTECTED]
> >> >>
> >> >>
> >> >
> >> >
> >> > --
> >> > Xavier Hanin - Independent Java Consultant
> >> > http://xhab.blogspot.com/
> >> > http://ant.apache.org/ivy/
> >> > http://www.xoocode.org/
> >> >
> >> >
> >>
> >> --
> >> View this message in context:
> >>
> http://www.nabble.com/wicketstuff-push-and-sharing-an-IChannelService-tf4671463.html#a13381208
> >> Sent from the Wicket - User mailing list archive at Nabble.com.
> >>
> >>
> >> ---------------------------------------------------------------------
> >> To unsubscribe, e-mail: [EMAIL PROTECTED]
> >> For additional commands, e-mail: [EMAIL PROTECTED]
> >>
> >>
> >
> >
> > --
> > Xavier Hanin - Independent Java Consultant
> > http://xhab.blogspot.com/
> > http://ant.apache.org/ivy/
> > http://www.xoocode.org/
> >
> >
>
> --
> View this message in context:
> http://www.nabble.com/wicketstuff-push-and-sharing-an-IChannelService-tf4671463.html#a13401455
> Sent from the Wicket - User mailing list archive at Nabble.com.
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]
>
>
--
Xavier Hanin - Independent Java Consultant
http://xhab.blogspot.com/
http://ant.apache.org/ivy/
http://www.xoocode.org/
