I don't even think it's a syntax really. It looks the way it does for the
same reason that your left eye is wired to your right brain (b/c god
created it that way ;)
We had been discussing moving more of the instantiation logic into the
runtime, but even in a JS runtime you need the to compiler to extract
method bodies and also to decorate constraints with dependencies. (The
latter is possible but impractical with regular expressions in the
runtime.)
Jim and I have been talking about changing the way that replication works
(see his RFC from a few months ago on declared replication.) Ideally we
would move these classes into LZX, but to do see we would need to
formalize aspects of the compiler pass. Ideally, we'd do something like
what you suggest (though SWF compatible) but add a place where the
compiler can pass additional attributes.
We may need to do this before legals.
A
On Apr 28, P T Withington wrote:
> Does anyone else think that this syntax is barfalicious? Could we have a
> better way?
>
> Is the arglist meant to be the JSON representation of the XML that you would
> have used statically?
>
> With the advent of e4x, could we just use the XML? Something like:
>
> canvas.appendChild(canvas.createElement(
> <{new_class} name="first" width="100" height="100" bgcolor="blue>
> <view name="second" bgcolor="{red}" x="100" width="20" height="20 />
> </{new_class}>));
>
> ?
>
> On 2006-04-28, at 00:33 EDT, Adam Wolff wrote:
>
> > I didn't test, but I think this should be:
> > new new_class(canvas, {name: "first", width: 100,
> > height: 100, bgcolor: blue},
> > [ { name : "LzView" ,
> > attrs: {name: "second", bgcolor: red, x: 100,
> > width: 20, height: 20} ], true );
> >
> >
> > A
> >
> >
> > On Apr 27, Max Carlson wrote:
> >
> > > Hi,
> > >
> > > I can't get this code example to work correctly to demonstrate how to
> > > instantiate a new class. The third argument is suppose to be for the
> > > children of this instantiated class, but how do you set the parent view
> > > of the children. Right now, the view "second" is appearing at the top
> > > level. This occurs naturally enough because I have supplied "null" as
> > > its parent argument, but canvas.first produces a "first property not
> > > found" error. The small red square should appear inside the blue square
> > > and be its child
> > >
> > > <canvas>
> > > <class name="new_class" width="100" height="100"/>
> > > <handler name="oninit">
> > > var parentview =
> > > new new_class(canvas, {name: "first", width: 100,
> > > height: 100, bgcolor: blue},
> > > [ new LzView(null, {name: "second", bgcolor: red, x: 100, width: 20,
> > > height: 20})], true);
> > > </handler>
> > > <handler name="onaddsubview" args="view">
> > > if (view.name <http://view.name> == null) return;
> > > Debug.write("view = " + view.name <http://view.name>);
> > > </handler>
> > > </canvas>
> > >
> > > _______________________________________________
> > > Laszlo-user mailing list
> > > [email protected]
> > > http://www.openlaszlo.org/mailman/listinfo/laszlo-user
> > >
> > _______________________________________________
> > Laszlo-user mailing list
> > [email protected]
> > http://www.openlaszlo.org/mailman/listinfo/laszlo-user
>
_______________________________________________
Laszlo-user mailing list
[email protected]
http://www.openlaszlo.org/mailman/listinfo/laszlo-user
