[antlr-interest] ANTLR 3.0 tree construction proposal

[Terence Parr]

On Jan 31, 2005, at 1:41 PM, Martin Probst wrote:
> Hi,
>
>> 	What is my point?  Tree construction syntax is already a barrier to
>> overcome to reading Antlr grammars.
>
> It might be too late or just insignificant and certainly somewhat
> offtopic but here is my users perspective: I find the ANTLR style of
> defining tree construction _very_ strange. IMHO it's close to code
> obfuscation with all these ^, !, :, =, # and [].

Well, the ^ and ! should be very clear, right?  You are merely 
specifying what is a subtree root and what is to be ignored.  It's the 
interaction with #(...) etc... in 2.x that is confusing I think.

> I would really prefer it if ANTLR would stick to some clear keywords 
> and
> maybe find a way of distinguishing ANTLR actions (e.g. where you do
> $setType and similar things) and actions in the target language.

I'm trying to avoid having any special stuff in actions. :)  The 
rewrites help a lot.  I don't anticipate needing actions to build trees 
at the moment.

> The readability of tree construction syntax is really very poor at the
> moment.

Thanks for the feedback!

Do you think the new rewrite stuff would be easy to understand? Some of 
surely would.  For example, build a flat tree with elements reordered:

r : A B => B A ;

Will that help do you think or is it just as bad?  I really want to get 
this right!

What are your thoughts on Loring's counter-proposal?

Ter
--
CS Professor & Grad Director, University of San Francisco
Creator, ANTLR Parser Generator, http://www.antlr.org
Cofounder, http://www.jguru.com