Re[5]: [cedet-semantic] Using Semantic to parse XPATH

[Eric M. L. <er...@si...>]

>>> Alex Schroeder <al...@gn...> seems to think that:
>Thanks a lot for your help!
  [ ... ]
>> %token NAME symbol "^[a-zA-Z_][a-zA-Z_0-9.-_]*$"
>>
>> Name: NAME ( $1 )
>>   ;
>
>Cool.
>
>Can you explain the difference between the following two, however?
>
>%token ANCESTOR-AXIS	        "ancestor"
>%token NAME 	    symbol 	"^[a-zA-Z_][a-zA-Z_0-9.-_]*$"
>
>As far as I understand the documentation, the second form applies only
>when the entire token has the symbol syntax class.  That means I
>should change all my FOO-AXIS tokens to the second form, right?
  [ ... ]

Hi,

  There is some small magic here.  When you specify a type for the
  token (symbol or punctuation) the string is a regular expression.
  When you exclude the type, it creates a new token primitive with
  the name specified, and it must match exactly with an implied ^ and
  $.

>I would like to run this very simple test:
>
>(with-temp-buffer
>  (insert "child::para")
>  (xpath-default-setup)
>  (bovinate t))
>
>However, I get the following error:
>
>Debugger entered--Lisp error: (wrong-type-argument listp [[((default . error) (25 . 4) 
>... [ugly crud deleted] ...
>  assq(bovine-toplevel [[(... ... ...) (... ...) (...) (...) (... ... ... ... ... ... .
>  (cdr (assq nonterminal table))                                                       
>  (let ((matchlist ...) (starting-stream stream) (nt-loop t) nt-popup nt-stack s lse lt
>  semantic-bovinate-nonterminal(((CHILD-AXIS 1 . 6) (punctuation 6 . 7) (punctuation 7 
>... [etc.] ...
>
>What does that mean?
>
>My BNF file contains the following line:
>
>%start         LocationPath
>
>Thus, I get suspicious when I see the bovine-toplevel thing in the
>backtrace.  The error seems to be somewhere else, however: TABLE is
>not a list but some weird vector instead.  Perhaps that is correct,
>because xpath-tables has exactly this value:

Unfortunately, I am not yet very savvy in the wisent parser's details.
Perhaps David Ponce can offer some suggestions.

As for `bovine-toplevel', the %start token is just replaced in the
actual language definition.  That should probably be fixed up
sometime.

Eric

-- 
          Eric Ludlam:                 za...@gn..., er...@si...
   Home: www.ultranet.com/~zappo            Siege: www.siege-engine.com
Emacs: http://cedet.sourceforge.net               GNU: www.gnu.org