[antlr-interest] Antlr noobie, nondeterminism abounds

[WesSantee]

Greetings all,

I'm just getting started in antlr (and language recognition in
general).  I've been looking through some of the examples, but after
trying my hand at my own grammar, I'm getting lexical nondeterminism
warnings everywhere, and I don't know how to get around them.

Here's an example:

1) Create a lexer rule representing exactly four decimal digits.
2) Create a lexer rule representing any 32-bit unsigned decimal value.

Since #2 is a superset of #1, I get nondeterminism warnings.  And no
wonder.  The problem is, I can't think of a way to disambiguate one
from the other.  Bumping k is only going to help in the case where the
number is >4 digits.

Here's another example:

1) Create a lexer rule representing ASCII characters 1..7F (call it
CHAR).
2) Create a lexer rule representing ASCII characters 1..FF (call it
CHAR8).
3) Create a lexer rule represeting everything in CHAR *except* '\r'
and '\n'.

Here again, CHAR8 is a superset of CHAR, and #3 is even crazier.

It may be that only people suffering from lexical muddy thinking would
think to implement rules like this, and that may very well be my
problem. :)  I'm trying to translate an ABNF formal syntax (the IMAP
protocol, RFC 3501 to be specific) into an antlr grammar, and so far
straight translation is not working very well.  

If you have solutions to the examples above, or recommendations on how
to cure lexical muddy thinking disease, I'm open to suggestions!

Cheers,
-Wes




 
Yahoo! Groups Links

<*> To visit your group on the web, go to:
     http://groups.yahoo.com/group/antlr-interest/

<*> To unsubscribe from this group, send an email to:
     antlr-interest-unsubscribe at yahoogroups.com

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