[Laszlo-dev] Syntactic sugar

[André Bargull]

If we had (native) setters in all runtimes, things would be much easier. 
I've you add an alias for "setAttribute", I'd like to remind you on the 
current limitation of the "setAttribute" inline compiler:
> // FIXME: [2002-12-03 ptw] This substitution is not correct
> // because it does not verify that the method being inlined is
> // actually LzNode.setAttribute.
So any new alias (let's say "set(..)") will be restricted to the same 
limitation.

And there is another issue which comes to my mind: in my opinion, it's 
not only the longer method name ("setAttribute" instead of "setX"), but 
also the whole function call with attribute-name in quotes + value. Just 
as an idea, what about adding a new syntax? (I do know adding a 
non-ECMAScript syntax may be harmful..)
For example to set the "x"-attribute to "10" on a view "v":
  set v.x = 10;




> I'd like to replace:
>
>    v.setAttribute('x', ...);
>
> with:
>
>    v.x = ...;
>
> someday.  Having an alias `set` for `setAttribute` would be a fine  
> interim step.
>
> On 2008-12-03, at 00:45EST, Max Carlson wrote:
>
> >/ I've had a number of developers tell me they miss the shorthand of  
> />/ setX(...) - 4.2 now requires setAttribute('x', ...).  This is good  
> />/ for consistency, but it's bad for backward compatibility and makes  
> />/ code more verbose.  setAttribute('foo', ...) requires a lot of extra  
> />/ typing...
> />/
> />/ What do you think about having the compiler expand the old setX/Y()  
> />/ methods into the corresponding setAttribute('x/y')?  We could add a  
> />/ terse alias for setAttribute() - set(), or maybe s()...
> />/
> />/ Thoughts?
> />/ -- 
> />/ Regards,
> />/ Max Carlson
> />/ OpenLaszlo.org
> /