[Laszlo-dev] Detecting classes
P T Withington
ptw at pobox.com
Mon Jan 18 05:50:09 PST 2010
On 2010-01-18, at 04:34, Rami Ojares / AMG wrote:
> Below is the debug output.
> At the bottom is the result of clicking one of the "<Pane>" objects in the debugger.
> So from that I draw the following conclusions.
>
> The argument c (the child) is an object with attributes
> - attrs, children (might be missing), class
>
> attrs seems to be an object (curly braces)
> children seems to be an array (brackets)
> class seems to be something special ( less-than, greater-than) I would assume this is class (sorry but I don't know exactly what is the distinction between class and object in javascript).
I need to write up more about how the debugger represents objects. Basically, it tries to print a literal representation that is isomorphic to the actual object. Javascript is a prototype-based language, and we build classes and instances on top of the prototyping mechanism. There is no literal representation of an object that has a prototype different from Object.prototype, so the debugger represents such objects in the format <<"type"#"unique id"| "mnemonic name">>. It will also use this format if an object that does have a literal representation is abbreviated due to being longer than Debug.printLength, or deeper than Debug.printDepth, or contains circular references. Finally, if you use the format "%#w", the debugger will only use the literal representation if the literal representation will be `==` to the actual object -- any other object will be represented using the detailed <<>> format.
> So far so good.
> The childClass seems to be always something of the form <...> (ie. class)
> Like it was supposed to.
>
> Then I execute the crucial line
> childClass.prototype is lz.Pane
>
> But it always returns false...damn.
I think I see the problem. You must be using swf9/10? And in that runtime, the test I described will not work. Hm. In fact, I'm not sure how you can ask if one Class is a subclass of another Class in AS3. I'll have to research that!
More information about the Laszlo-dev
mailing list