[Laszlo-dev] For Review: Change 20080209-dda-7 Summary: SWF9 mixin support

[P T Withington]

This looks great!  Approved.

Comment: In Javascript, symbols begining with $ are supposed to be  
reserved for internal implementation, so we really are within our  
rights to use that.  If we really cared, we could escape user $ by  
doubling them or something, but I don't see any reason to.

On 2008-02-09, at 14:28 EST, Donald Anderson wrote:

> Change 20080209-dda-7 by dda at lester.local on 2008-02-09 13:57:38 EST
>    in /Users/dda/laszlo/src/svn/openlaszlo/branches/devildog
>    for http://svn.openlaszlo.org/openlaszlo/branches/devildog
>
> Summary: SWF9 mixin support
>
> New Features:
>
> Bugs Fixed: LPP-5266
>
> Technical Reviewer: ptw (pending)
> QA Reviewer: promanik (pending)
> Doc Reviewer: (pending)
>
> Documentation:
>
> Release Notes:
>
> Details:
>    The mixin support follows the design given by Tucker in LPP-5266.
>
>    In SWF9Generator, when a mixin is seen, it is recorded in  
> mixinDef hashmap.
>    For each use of a mixin combination, the needed interstitials are  
> recorded in
>    the mixinRef hashmap, and all we need to do to the referencing  
> class is change
>    its superclass to be the interstitial.
>
>    Each mixin declaration AST is duplicated - one is processed
>    like a class and one like an interface.   
> (CommonGenerator:translateClassDirectivesBlock
>    now knows three ways to process a class - the third being to  
> create a property list).
>    The 'interface' version of the mixin remains in the parse tree  
> and is emitted
>    as the interface.  The 'class' version of the mixin is what is  
> put into mixinDef.
>
>    After we are done processing the AST, we visit the mixinRef map  
> and for each entry,
>    we append a new class to the parse tree that implements the  
> interstitial.
>
>    In unparsing, we need to track the current class name, since  
> constructors within
>    interstitials need that as their name (and not their original  
> name which matches
>    the mixin).  Arguably this could be handled when the tree is  
> created, but currently
>    we are sharing all the contents of the interstitial AST with  
> possibly other interstitials.
>    Also in unparsing, bodies of functions for interfaces are handled  
> differently, so
>    we must track when we are in an interface.
>
>    The changes only affect the implementation of mixins for SWF9.   
> It should be
>    possible to push this implementation into generic javascript  
> eventually.
>    That seemed too big for this moment.
>
> Tests:
>    regression tests: smoke/lzpix/weather for DHTML, SWF8 and henry's  
> hello for SWF9
>    mixin tests: for this, verified they compiled and checked  
> contents by eye.
>      constructors in mixins
>      public/private methods in mixins
>      vars in mixins
>      mixins can appear after usage
>      overridden methods possible
>      multiple mixins
>    Not yet tested - converting LFC9 classes back to use mixins.
>
> Files:
> M      WEB-INF/lps/server/src/org/openlaszlo/sc/SWF9Generator.java
> M      WEB-INF/lps/server/src/org/openlaszlo/sc/CommonGenerator.java
> M      WEB-INF/lps/server/src/org/openlaszlo/sc/ 
> SWF9ParseTreePrinter.java
> M      WEB-INF/lps/server/src/org/openlaszlo/sc/Compiler.java
> M      WEB-INF/lps/server/src/org/openlaszlo/sc/ParseTreePrinter.java
> M      WEB-INF/lps/server/sc/src/org/openlaszlo/sc/parser/ 
> ASTIdentifier.java
>
> Changeset: http://svn.openlaszlo.org/openlaszlo/patches/20080209-dda-7.tar
>
>
>
> --
>
> Don Anderson
> Java/C/C++, Berkeley DB, systems consultant
>
> voice: 617-547-7881
> email: dda at ddanderson.com
> www: http://www.ddanderson.com
>
>
>
>