[Laszlo-dev] Any gotchas I should know about for converting old Class code to new class syntax?

[P T Withington]

Any chance of getting this into the Wiki, whereever we have Legal's  
porting notes?

On 2007-01-09, at 17:01 PST, Philip Romanik wrote:

> Hi Tucker,
>
> I was about to say I don't have anything written down, but then I  
> found some notes. I'm attaching the list. I removed the last item  
> in my notes because it is no longer relevant (it was about moving  
> $debug blocks after the class which is no longer an issue).  
> Unfortunately, the notes aren't complete but Henry can look at the  
> existing code for guidance.
>
> Phil
>
> =====================
>
> Defining a Class
>
> Old:
> ---
>
> LzNode = Class( "LzNode" , null , mvn() );
> var mvn = function (){
>     var f = function ( parent , attrs , children , instcall ){
>         ...
>     };
>
>     return f;
> }
>
> New:
> ----
>
> initialize() is the class constructor
>
> class LzNode {
>
> function initialize ( parent , attrs , children , instcall ){
>     ...
> }
>
> ...
> }
>
>
> A static initialize() function runs once for each class
>
> // From LzNode.lzs
> static function initialize (prototype) {
>
>   // Ensure you have your own private dictionaries, not one
>   // inherited from your superclass
>   for (var hash in {setters: true, getters: true, defaultattrs: true,
> __LZdelayedSetters: true, earlySetters: true} ) {
>     if (! prototype.hasOwnProperty(hash)) {
>       prototype[hash] = new LzInheritedHash(prototype[hash]);
>     }
>   }
>
> }
>
>
> Deriving from a base class:
>
> class LzView extends LzNode {
> }
>
>
> Deriving from a trait:
>
> The need for 'Instance' is only needed is there is no base class.
>
> class LzDataText extends LzDataNode, Instance {
> }
>
>
>
> Defining variables
>
> LzNode.prototype.defaultattrs = { $hasdefaultattrs : true };  
> LzNode.prototype.datapath = null; LzNode.prototype.setters ={
>     name : "setName" ,
>     id : "setID" ,
>     $events : "__LZsetEvents" ,
>     $refs : "__LZstoreRefs" ,
>     $delegates : "__LZstoreDelegates" ,
>     placement : -1 ,
>     datapath : "setDatapath",
>     $setters : -1,
>     $classrootdepth : "__LZsetClassRoot",
>     $datapath : "__LZmakeDatapath"
> }
>
>
> var defaultattrs = { $hasdefaultattrs : true };
> var datapath = null;
> var setters ={
>     name : "setName" ,
>     id : "setID" ,
>     $events : "__LZsetEvents" ,
>     $refs : "__LZstoreRefs" ,
>     $delegates : "__LZstoreDelegates" ,
>     placement : -1 ,
>     datapath : "setDatapath",
>     $setters : -1,
>     $classrootdepth : "__LZsetClassRoot",
>     $datapath : "__LZmakeDatapath"
> };
>
>
> Overriding variable values
>
> OLD:
> LzState:
> LzState.prototype.$isstate = true;
>
> New:
> LzNode.lzs:
>     var $isstate = false;
> LzState.lzs:
>     prototype.$isstate = true;
>
>
>
> Static Functions
> ----------------
>
> OLD:
>
> LzView.__LZproxypolicies = [];
>
> LzView.__LZcheckProxyPolicy= function ( url ){
>     var pol = _root.LzView.__LZproxypolicies;
>
>     for ( var i = pol.length-1; i >=0; i-- ){
>         var resp = pol[ i ] ( url );
>         if ( resp != null ) return resp;
>     }
>
>     return _root.canvas.proxied;
> }
>
>
> NEW:
>
> static var __LZproxypolicies = [];
>
> //-------------------------------------------------------------------- 
> ---------
> // @keywords private
> //-------------------------------------------------------------------- 
> ---------
> static function __LZcheckProxyPolicy ( url ){
>     var pol = LzView.__LZproxypolicies;
>
>     for ( var i = pol.length-1; i >=0; i-- ){
>         var resp = pol[ i ] ( url );
>         if ( resp != null ) return resp;
>     }
>
>     return canvas.proxied;
> }
>
>
>
> =====================
>
>
>> Not if all it does is define attributes and methods.  If it is doing
>> anything other than that, feel free to ask.  (I'm hoping PhilR has a
>> checklist that he can send you...)
>>
>> On 2007-01-09, at 09:52 PST, Henry Minsky wrote:
>>
>> > I'm converting the rpc library code which uses the old "Class"
>> > mechanism to use the new"class" syntax, just wondering if there are
>> > any
>> > gotchas I should know about, mostly asking PhilR and Tucker if
>> > there there's anything non-obvious
>> > I should watch out for.
>> >
>