bkiers / pcreparser Goto Github PK

$ java -cp antlr-3.3-complete.jar org.antlr.Tool PCRE.g
Error: Could not find or load main class org.antlr.Tool

$ ls
antlr-3.3-complete.jar PCRE.gr
$ pwd
/home/charly/gitrepo/PCREParser/src/main/antlr3/pcreparser

Iam not sure what iam I missing

At this point only named groups with letters can be processed. And as far as PCRE standard is concerned, digits can be there as well.

name
: letters -> NAME[$letters.text]
;

P.s. Great project btw, helped a lot!

"(?<hello_world>.+)"

That should be a valid capture group, but instead we see errors like this:

line 1:4 missing GreaterThan at '_'

The fix should be as simple as adding underscore to allowed characters in the NAME type.

P.S. Thank you for this awesome AST. It made my life a hundred times easier!

I've found what appears to be an error with parsing named captures.

Using the online parser I entered the following regular expression:

(?<named_cap>hello)

It produces the following AST:

'- ALTERNATIVE
   '- ELEMENT
      '- NAMED_CAPTURING_GROUP_PERL
         |- NAME='named'
         '- ALTERNATIVE
            |- ELEMENT
            |  '- LITERAL='_'
            |- ELEMENT
            |  '- LITERAL='c'
            |- ELEMENT
            |  '- LITERAL='a'
            |- ELEMENT
            |  '- LITERAL='p'
            |- ELEMENT
            |  '- LITERAL='>'
            |- ELEMENT
            |  '- LITERAL='h'
            |- ELEMENT
            |  '- LITERAL='e'
            |- ELEMENT
            |  '- LITERAL='l'
            |- ELEMENT
            |  '- LITERAL='l'
            '- ELEMENT
               '- LITERAL='o'

As you can see, instead of identifying a named capturing group with the name named_cap, it identifies a group with the name named, and inteprets _cap> (including, oddly, the trailing >) as literals.

According pcrepattern(3), underscores are valid in named capture names:

In PCRE, a subpattern can be named in one of three ways: (?<name>...) or (?'name'...) as in Perl, or (?P<name>...) as in Python. ... Names consist of up to 32 alphanumeric characters and underscores, but must start with a non-digit.

Am I correct in thinking this is a bug, or have I misunderstood something?