[Laszlo-dev] Bug with Events?

[P T Withington]

On 2007-01-16, at 00:13 EST, Tyler Pitchford wrote:

> When using an outside actionscript object (in this example a
> SharedObject) I'm getting some odd results when I try to attach
> handlers to events on the object.
>
> If I use the code:

For the purpose of exposition, I'll refer to  the instance these  
appear in as `yourInstance`.

>  <event name="onSync"/>
>  <handler name="oninit">
>  var t = this;

Here you are creating your own closure over `this` (which is  
`yourInstance`), so that calling the `onSync` method on the `_so`  
object correctly calls the `_onStatus` method of `yourInstance`.

>  this._so.onSync = function(evtObj) {
>          t._onStatus(evtObj);
>      };
>  </handler>
>
> and in the attaching class:
>
>  this._onSyncDel = new LzDelegate(this, "syncHandler", this._so,  
> "onSync");
>
> everything works fine and this._so.onSync = an LzEvent object. If,
> however, I use this:
>
>  <event name="onSync"/>
>  <handler name="oninit">

Here, no closure is created.  You are assigning the naked method  
`_onStatus` to the `onSync` method of the `_so` object.   
`yourInstance` is not captured.  Inevitably, when `_so.onSync` is  
invoked `this` in `_onStatus` is `_so`, not `yourInstance` as you  
might have expected.

You could call this a bug in Javascript, since other languages  
automatically close over `this` when capturing a method.  This bug is  
proposed to be fixed in Javascript 2, but in Javascript 1, you either  
need to create your own closure, or you need to use the LZX delegate  
system (which is really just a manual closure).

>    this._so.onSync = this._onStatus;
>  </handler>
>
> and
>
> the same in the attaching class this._so.onSync = a Function object
> pointing to the evet exposings _onStatus method. Then if I change the
> code to:
>
>  <event name="onSync"/>
>  <handler name="oninit">
>    this._onSyncDel = new LzDelegate(this, "_onStatus", this._so,  
> "onSync");
>  </handler>
>
> The event simply doesn't fire. While it took me awhile to figure out
> the fix I demonstrated in the first example, shouldn't at least the
> first two function the same? It appears calling this._so.onSync =
> this._onStatus is actually overwriting this.onSync instead of
> this._so.onSync, any ideas?
>
> Cheers,
>  Tyler