[antlr-interest] ANTLR 3.0 beta timing

[Terence Parr]

On Jun 10, 2006, at 2:30 PM, Brian Cox wrote:

> On Friday 09 June 2006 20:11, Terence Parr wrote:
>> Any comments on v3 timing or the idea of lulu.com?
>
> Sounds good to me!  My only concern so far has been the reliance on  
> ANTLR
> 2.7.6 for using ANTLR v3, but if you can now or will be able to  
> soon build
> ANTLR v3 with itself I'd imagine that little problem will soon be  
> history.

Yeah, relying on v2 is a drag, but ST relies on it also and so BOTH  
tools need to be updated.  Note that this is ONLY for the tool itself  
and not the generated code.  The generated code is v3 clean.

> To be honest, I started using ANTLR v3 some weeks ago after much  
> frustration
> with ANTLR 2.7.6.

heh heh heh

> For some reason, I couldn't get an ANTLR 2 grammar of my
> own making to be accepted by ANTLR 2, let alone getting it to build  
> AST
> trees.  With ANTLR v3, it's been relatively easy!

Now that is an endorsement!

> Initially, the "early
> access" nature of v3 scared me away, but I'm glad I finally took  
> the plunge
> and am eagerly awaiting the beta releases and, eventually, the  
> final release
> of 3.0 and whatever feature enhancements await us for v3.1.

:)

The early access is my way of saying "the world will change out from  
underneath you maybe".  However, it has proven rather serviceable.  I  
found it very effective building my start on Mantra (http:// 
www.linguamantra.org) and I found my first actual bug in the analysis  
engine in over a year...I messed something up a while ago and it just  
surfaced.  I was able to fix in 30 minutes and the resulting code is  
simpler and more efficient. :)

Trust me, v3 works.  It's just the fringes which are a little shaky.   
I'm now in bug fix mode and it will go beta w/o a real manual.  Then  
I work on the manual, but 3.0 final will definitely appear before  
school starts.

I'm trying to sneak incremental parsing into 3.0, but it might not  
make it.  Prashant Deva, of ANTLRStudio fame, and I are trying to  
formalize his strategy from ANTLRStudio but I'm completely focused on  
getting 3.0 out and to finish the journal paper with Jean Bovet on  
ANTLRWorks.

Terence