Parse::RecDescent Grammar Fun

[ichimunki]

Below is my code. I am happy with my grammar except that if a token is surrounded by punctuation then the token gets slurped into it. I want the third line of $text to result in ... <punct: ..><link: spaced><punct: ..> ..., not in ... <punct: ..[[><word: spaced><punct: ]]> ... (as happens currently). Anybody know how I can fix it? Also, if it looks like my grammar is screwy, or there are other things I can do to improve this, let me know-- I am just getting my arms around the basics of this very cool module, so pointers will be appreciated. Thanks!

#!/usr/bin/perl -w
use strict;
use Parse::RecDescent;

my $grammar = join( '', <DATA> );

my $parser = Parse::RecDescent->new( $grammar ) or
    die "Error: Bad grammar\n";

#while(<STDIN>){ $text .= $_; }
my $text =<< "SUB_STDIN";
A nicely [[spaced]] link.
A poorly[[spaced]]link.
Another poorly..[[spaced]]..link.
SUB_STDIN

my $results = $parser->startrule( $text ) or
    die "Error: Bad text\n";

__DATA__

startrule: <skip:''> bit(s) 
bit: eol | word | space | token | punct

eol: /\n[ \t]*/ {print "<newline>\n" }
space: /[ \t]+/ {print "< >" }
word: /[\w\']+/ {print "<word: $item[1]>" }
punct: /[^\w\s]+/ {print "<punct: $item[1]>" }

token: link
link: /\[\[(.+?)\]\]/ {print "<link: $item[1]>" }

[hsmyers]

Mostly your immediate problem is that you told the grammar that [ and ] are 'punct' so it happily eats the delimiter for 'link'! Try:

punct: 
    /[^\w\s\[\]]+/ 
    {print "<punct: $item[1]>" }

Note the inclusion of [ and ] (escaped) in the definition of what 'punct' is 'not'.

--hsm

"Never try to teach a pig to sing...it wastes your time and it annoys the pig."

[ichimunki]

Ah, but they are punct if they appear as [ or ] (singles). Also, they qualify as punct if they appear as pairs out of order: ]] stuff [[. They would also be punct if you just had [[ and no matching ]]. They would also be punct if they appeared as [[]]-- you can't have a link to absolutely nothing. :)

So far I'm solving the problem by simply forcing punct to be a single character instead: /[^\w\s]/. That way the punct rule doesn't eat up my string before it can run the token test on relevant pieces, since token is checked before punct. What I was kind of hoping for was a way to use a negative lookahead of some sort... or a way to build an extra rule layer that would somehow account for this.

[hsmyers]

This may not be all that robust, but it is better than square one...

#!/usr/bin/perl -w
use strict;
use Parse::RecDescent;

my $grammar = join( '', <DATA> );

my $parser = Parse::RecDescent->new( $grammar ) or
    die "Error: Bad grammar\n";

my $text =<< "SUB_STDIN";
A nicely [[spaced]] link.
A poorly[[spaced]]link.
Another poorly..[[spaced]]..link.
A harder problem [spaced] link.
Yet harder still..]]spaced[[..link.
The real problem is [[spaced]] followed by ]] or [[ link.
SUB_STDIN

my $results = $parser->startrule( $text ) or
    die "Error: Bad text\n";

__DATA__

startrule: 
    <skip:''> bit(s) 
bit: 
    eol 
    | word 
    | space 
    | token 
    | punct
eol: 
    /\n[ \t]*/ 
    {print "<newline>\n" }
space: 
    /[ \t]+/ 
    {print "< >" }
word: 
    /[\w\']+/ 
    {print "<word: $item[1]>" }
punct: 
    /[^\w\s\[\]]+/ 
    {print "<punct: $item[1]>" }
    | /(?<!\[)\[(?!\[)/
    {print "<punct: $item[1]>" }
    | /(?<!\])\](?!\])/
    {print "<punct: $item[1]>" }
    | /(?<!\[\[)\]\]/
    {print "<punct: $item[1]>" }
    | /(?<!\]\])\[\[/
    {print "<punct: $item[1]>" }
token: link
link: 
    /\[\[(.+?)\]\]/ 
    {print "<link: $item[1]>" }

Notice: new test cases for bracket as 'punct'.

--hsm

"Never try to teach a pig to sing...it wastes your time and it annoys the pig."

[Abigail-II]

Your problem is the greedyness of punct. If it's ok that !@#$ will be parsed as four tokens, you should be able to get away with:

    punct: /[^\w\s]/ {print "<punct: $item[1]>" }

although I have not tested it.

Otherwise, you may want to try something like (also untested):

    punct: /(?:[^\w\s[]+|\[[^\w\s[]*|\[\[(?!.+?]]))+/ 
           {print "<punct: $item[1]>"}

Note that your link rule consumes "[[]]]" completely, as a leading [[, a ] as the content, and a trailing ]]. Similary, "[[]] foo [[]]" is consumed by the link rule completely, with "]] foo [[" as the part inside the 'link'.

Abigail

[ichimunki]

Yes, I think letting punct be single character tokens is the way to go here. That regex is probably intuitive to some, but it looks like a maintenance nightmare.

Thanks for the info on the link rule. I had completely forgotten to stop to think about what all that would match.