[antlr-interest] New article on StringTemplates and Treewalkers

[Terence Parr]

On Jan 12, 2006, at 7:43 AM, Andy Tripp wrote:
> I'm not following you exactly, but I think it's a reasonable  
> question. I don't buy that I could, for example,
> use StringTemplate to output C# instead of Java, without major  
> changes to my code.

I believe you have the controller and view entangled so it would be  
hard.  I have separated things with ST.  All declarative systems seem  
to have pattern->pattern which necessarily tangles things; you'd have  
to cut-paste all of your this->that patterns to change the that.   
icky code duplication I'd argue. ;)

ANTLR v3 is doing very well at separation so far I feel.  I have  
people working on or potentially will work on targets as follows:

Java: uncle T
C++: Ric
C#: Michael / Kunle??
C: Jim Idle
Objective-C: Kay Roepke
Python: Marq Kole?, Wolfgang Hafelinger?
Ruby: Martin Traverso;
LISP: somebody is exploring this at the moment
Parrot vm compiler: Bernhard Schmalhofer
PhP: Ryan King

Note that I had bytecodes generated for the cyclic DFA for a while.   
Also note that C and Obj-C targets are very far along as far as I can  
tell.  C has a lot more work to do than Java, trust me.  So far no  
one has asked for anything to be done to code generator minus some  
attributes I forgot to push in.

> I don't think
> ANTLR could output MS or Haskell (or Lisp) without major changes.  
> ST is not fullfilling the
> "separate your view from your translation" promise in any cases  
> other than the most trivial, contrived cases

While I've not done Lisp yet, I'm pretty sure my experience and  
intuition should count for something when I say I'm certain I could  
make this work.  What do you base your assertion on?  It *seems* like  
it would be hard?  Because the order of the way you define things is  
different?  Until I find a construct that I couldn't generate with my  
code generator in multiple back-ends, I must assume it simply  
*appears* to be very difficult. No offense.  :)

If I knew lisp well, i'd just go build it ;)

> (or when the output languages are so similar as C++/Java/C# and the  
> domain is so limited as in the case
> of all the code that ANTLR spits out).

This limitation you speak of is speculation until either we get a  
lisp or other weird target or you provide an example that cannot be  
done to save us the trouble. ;)

>> With some constraints, I think this is possible.
>>
>
> OK then, what is the Java equivalent of "memset (111, 222, 333);"?
> I suppose that's a constraint :)

Depends on how natural you want the translation as I think you  
correctly pointed out ;)

Ter