Re: [Pyparsing] [pyparsing] RecursiveGrammarException

[Paul M. <pt...@au...>]

Haifeng -

You have 2 problems going on here:
- you are matching 'a' before trying to match 'a->b'; the leading 'a' will
match, and the trailing '->b' will then raise an error
- you have defined a left-recursive expression (to match 'a->b', you will
try to match 'a->b', but first you will match 'a->b', etc.); the
left-recursion is what validate() is complaining about

Pyparsing is not like lex/yacc or LR parsers - it works purely left-to-right
through the input text, according to the definitions in the grammar.  There
is no lookahead to try to achieve longer matches, as is done in regular
expressions.  In pyparsing expressions, the '|' is a short-circuiting
alternative operator, creating MatchFirst expressions: the first alternative
expression to match will be the "winner", even if a later alternative might
match a longer part of the input string.  You can force all alternatives to
be tested for matching (with the longest match chosen) using the '^'
operator, which creates Or expressions.

Pyparsing does not allow you to directly implement a BNF definition, since
BNF follows different rules for matching, and has different constraints for
notation. There is more description of the BNF->pyparsing process here
(http://pyparsing.wikispaces.com/message/view/home/11262229 and
http://pyparsing.wikispaces.com/message/view/home/41102255).

You use some terms like "postfix" in your grammar, but I don't see a lot of
postfixing going on.  You do describe an infix '->' operator, but this too
is somewhat confusing; in your example, you just show "a->b", but from your
implementation, it looks like you want to support unbounded repetition,
presumably something like "a->b->c->d".

To implement your grammar to parse your two examples, you would write:

ARROW = Literal('->')
identifier = Word(alphas+"_", alphanums+"_")

primary_expression = identifier
compound_expression = identifier + ARROW + identifier

# must test for compound_expression before primary_expression
expression = compound_expression | primary_expression


To add repetition to compound_expression, you could simply change it to:

compound_expression = identifier + OneOrMore(ARROW + identifier)


-- Paul



-----Original Message-----
From: Haifeng [mailto:min...@gm...] 
Sent: Sunday, August 21, 2011 2:43 AM
To: pyp...@li...
Subject: [Pyparsing] [pyparsing] RecursiveGrammarException

Hi

I am new to pyparsing. I am trying to parse just the following C
expressions:
a
a->b

My code looks like:

from pyparsing import *

ARROW = Literal('->')
identifier = Word(alphas+"_", alphanums+"_")

primary_expression = identifier

postfix_expression = Forward()
postfix_expression << ( primary_expression | (postfix_expression + ARROW +
identifier))

But it failed on postfix_expression.validate(). What did I do wrongly here?
Thanks
----------------------------------------------------------------------------
--
Get a FREE DOWNLOAD! and learn more about uberSVN rich system, 
user administration capabilities and model configuration. Take 
the hassle out of deploying and managing Subversion and the 
tools developers use with it. http://p.sf.net/sfu/wandisco-d2d-2
_______________________________________________
Pyparsing-users mailing list
Pyp...@li...
https://lists.sourceforge.net/lists/listinfo/pyparsing-users