[antlr-interest] antlr 3.0

[Terence Parr]

Hi Bob,

Interesting suggestion.  However, I think it's probably best to focus on 
getting 2.x more "stable" and try out a few more ideas.  I'd like the 
spec for 3.0 to be written completely after we've seen all the features 
in action ;)   Further, I'll probably have to guide the construction of 
the 3.0 spec; we could  designate syntax committee, error handling 
committee, predicate committee, runtime committeee, where's my damn car 
keys committee, etc... [you can replace "committee" with "rabble" if you 
want ;)].

I'm not sure completely sure where/when the "book" fits into all of 
this.  I have been assuming that the most useful path to the community 
is to fix up 2.x to include, say Loring's rewrites for trees and my 
output grammar stuff, then finish the book based upon 2.x and then do 
3.0 spec and then code it.  If we hold off on the book until 3.0 
version, you'll never see the book ;)  Besides, the book will mostly be 
concepts that will apply to any new version.

My main concern with 3.0 is a better implementation and more rigorous 
testing.  Probably will have hoisting and full LL(k) back in there like 
PCCTS had.   There may be some concept and syntax changes.  Mostly 
though I'd like to clean up the code (and have up-to-date doc) so people 
can understand the analysis and build new code generators etc...  The 
2.x implementation is a bit muddy. ;)

Ter

On Thursday, December 6, 2001, at 08:27  PM, bob mcwhirter wrote:

>> I'll have to start a betting pool then. :-?
>
> So, bet on a date that antlr-3.0 will be released by?
>
> Tangentially...
>
> With my day job, I'll sometimes get stretches of Copious Amounts of
> Free Time.
>
> I love antlr.  I'd love antlr-3.0.
>
> I've hand-written several LL(k) parsers based upon Ter's methods,
> mostly because of short-comings with Antlr (notable, size and
> incompatibility with WebLogic's internal antlr.jar and Classloader
> isolation bugs).
>
> I'd certainly never want to compete with Ter, since he's the bonafide
> parsing $DIETY.
>
> Instead of us waiting for Ter to get the ball rolling on 3.0, maybe
> let's pick up the slack (only with Ter's blessing, of course), and
> start work on 3.0.
>
> We all know the hard part isn't writing the code.  We can be bantering
> around ideas, and floating designs.
>
> Thoughts, anyone?
>
> 	-bob
>
>
>
>
> Your use of Yahoo! Groups is subject to 
> http://docs.yahoo.com/info/terms/
>
>
--
Chief Scientist & Co-founder, http://www.jguru.com
Creator, ANTLR Parser Generator: http://www.antlr.org


 

Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/