Iliad - updating elements

> Le lundi 13 juillet 2009 à 14:50 +0200, Stefan Schmiedl a écrit :
> > Before I dive in and learn to do it manually, is there an idiomatic
> > way of updating a given element (periodically)?
> >
> > I am building the supervisor page for the OnlineTester, which
> > contains something like a "dashboard" showing the completion rate
> > of each test. This is currently implemented as widget although
> > it does not need to be, now that I think about it. Anyways, its
> > contents should be refreshed once every minute.
>
> You should probably use one of the periodical updater jQuery plugins.
> http://plugins.jquery.com/search/node/updater

> On Mon, 13 Jul 2009 15:05:24 +0200
> Nicolas Petton <[hidden email]> wrote:
>
> > Le lundi 13 juillet 2009 à 14:50 +0200, Stefan Schmiedl a écrit :
> > > Before I dive in and learn to do it manually, is there an idiomatic
> > > way of updating a given element (periodically)?
> > >
> > > I am building the supervisor page for the OnlineTester, which
> > > contains something like a "dashboard" showing the completion rate
> > > of each test. This is currently implemented as widget although
> > > it does not need to be, now that I think about it. Anyways, its
> > > contents should be refreshed once every minute.
> >
> > You should probably use one of the periodical updater jQuery plugins.
> > http://plugins.jquery.com/search/node/updater
>
> Smart solutions for a bigger problem :-)
>
> > There is no integration in Iliad for now, so you will have to write
> > some Javascript code ;)
>
> Here's what I have now, but consider cheating (a bit):
>
>   contents [
>     <category: 'building'>
>     ^ [ :e | |id|
>       id := self session nextId printString.
>       e build: self statusWidget.
>       e anchor id: id; text: 'Refresh'; action: [ self statusWidget markDirty ].
>       e script: 'window.setInterval( function() { jQuery( "a#' , id , '" ).click(); }, 60*1000 );'
>     ]
>   ]
>
> Question #1:
> Is there a reason why "session nextId" is a number instead of a string?

> > Question #2:
> > The solution above nicely logs the expected XHR requests in firebug.
> > When I tried
> >             Iliad.evaluteAnchorAction( "a#' , id , '" );
> > instead, no such requests were logged. Do I actually go and read
> > the Javascript books on the shelf behind me or is there something
> > else I'm not aware of?
>
> Iliad.evaluateAnchorAction(anchor) expects a DOM element, not an id.
> So Iliad.evaluateAnchorAction(jQuery(#someId)) should work.

> On Mon, 13 Jul 2009 20:07:52 +0200
> Nicolas Petton <[hidden email]> wrote:
>
> > Hmm, yes. As far as I remember, numbers are needed to know in which
> > order actions need to be executed, in a form for instance.
> >
> > But I don't think you need Session>>nextId, it is mostly used for
> > actions registration. You could use Iliad.Id instead.
>
> Channeling my youngest kid:
>
> butbutbut ... you *told* me I could use *either* of them ... sniff

>
> > > Question #2:
> > > The solution above nicely logs the expected XHR requests in firebug.
> > > When I tried
> > >             Iliad.evaluteAnchorAction( "a#' , id , '" );
> > > instead, no such requests were logged. Do I actually go and read
> > > the Javascript books on the shelf behind me or is there something
> > > else I'm not aware of?
> >
> > Iliad.evaluateAnchorAction(anchor) expects a DOM element, not an id.
> > So Iliad.evaluateAnchorAction(jQuery(#someId)) should work.
>
> ... and it does, great!
>
> This means I'll skip the general javascript and give the jQuery docs
> a thorough reading instead. Looking at the implementation of
> evaluateAnchorAction led me to my erroneous belief.
>
>     evaluateAnchorAction: function(anchor) {
>         var actionUrl = jQuery(anchor).attr('href');
>         this.evaluateAction(actionUrl);
>     },
>
> hmmm.... final question: How would I do this without putting an anchor
> into my page? I guess I'd have to do some work to gather the expected
> session tokens and whatever else is required?

> On Mon, 13 Jul 2009 22:40:05 +0200
> Nicolas Petton <[hidden email]> wrote:
>
> > Yes, you can. It's just that they are made for different purposes.
>
> So I went along and replaced "self session nextId printString" with
> "Iliad.Id new: 8". Imagine my surprise when I found the following:
>
> <script>
> $(document).ready( function() { $("#Iliad.Id (97 104 50 57 114 113 112 97 )").focus(); } );
> </script>

> >> Yes, you can. It's just that they are made for different purposes.
> >
> > So I went along and replaced "self session nextId printString" with
> > "Iliad.Id new: 8". Imagine my surprise when I found the following:
> >
> > <script>
> > $(document).ready( function() { $("#Iliad.Id (97 104 50 57 114 113 112 97 )").focus(); } );
> > </script>
> >
> > Ok, I'll do the right thing and subclass it .-)
>
> No, what you should do is override Iliad.Id>>#displayOn: (or #printOn:).

> On Tue, 14 Jul 2009 01:44:53 +0200
> Paolo Bonzini <[hidden email]> wrote:
>
> > >> Yes, you can. It's just that they are made for different purposes.
> > >
> > > So I went along and replaced "self session nextId printString" with
> > > "Iliad.Id new: 8". Imagine my surprise when I found the following:
> > >
> > > <script>
> > > $(document).ready( function() { $("#Iliad.Id (97 104 50 57 114 113
> > > 112 97 )").focus(); } ); </script>
> > >
> > > Ok, I'll do the right thing and subclass it .-)
> >
> > No, what you should do is override Iliad.Id>>#displayOn: (or
> > #printOn:).
>
> Nah ... that would be fiddling in foreign code ... I don't do that ;->
>
> The subclass stays, because 8 random characters for in-page ids
> "should be enough for everyone".