Incognito has asked for the wisdom of the Perl Monks concerning the following question:

When parsing through a properly formatted JavaScript file (containing just functions), we are interested in returning the entire string of a specific function, say foo() in this snippet: 
foo (a, b, c) {
  if (c) { a++; } else { b--; }
  print "hi";
}
fum (d, e, f, g, h) {
  if (d) { e++; } else { f = g + h; }
  print "ho";
}

[download]
we want to match the entire first function... or actually we just want the ability to write a regex that matches any (not too complex is okay) valid function...

Replies are listed 'Best First'.
Re: How do I match/extract a function from a JavaScript file?
by boo_radley (Parson) on Oct 06, 2001 at 01:16 UTC
    be sure to show your work.
    :P
    eh? wot? oh, alright. You probably want to look at Text::Balanced for any serious answer, otherwise perlre can assist you.
[reply]
Re: How do I match/extract a function from a JavaScript file?
by ForgotPasswordAgain (Vicar) on Sep 30, 2007 at 17:53 UTC

    You need to escape comments and strings, too. Try this one:

    bum () {
  // :}
  if (d) { e++; } else { f = g + h; }
      print "ho \" :\}";
  /* if (hi) {
  */
     if (ho) {
         print 'hee \\\' :}';
     }
}

    [download]
[reply]
[d/l]
      ok..here we go with this .. 
      #!/usr/bin/perl
use strict;
use re 'eval';
my $Jscript = join (" ",<DATA>);
# Extract function foo from Jscript
my $content = _extract("foo",$Jscript);

sub _extract {
    my ($funct,$Jscript) = @_;

    #regexes to match all but starting positions of strings and commen
+ts 
    my $double_quote_str = qr!(\\.|[^"\\])*"!;
    my $single_quote_str = qr!(\\.|[^'\\])*'!;
    my $comment1 = qr![^\n\r]*!;
    my $comment2 = qr!.*?\*/!s;    
    
    #now match complete strings or comments..whatever matches first wi
+ll 'eat' others
    my $escape_strings = qr!
        
        (   
            '(?{$start='sing_quot'})|
             "(?{$start='dbl_quot'})|
            //(?{$start='comment1'})|
            /\*(?{$start='comment2'})
        )
        (?(?{$start eq 'sing_quot'})$single_quote_str)
        (?(?{$start eq 'dbl_quot'})$double_quote_str)
        (?(?{$start eq 'comment1'})$comment1)
        (?(?{$start eq 'comment2'})$comment2)
        
    !x;

    my ($round,$curly);
    
    # match balanced nested round braces ..escape strings and comments
    $round = qr!
    \((
        $escape_strings  
        |
        [^()]|(??{$round})
      )*
    \)
    !x;
    
    # match balanced nested curly braces ..escape strings and comments
    $curly = qr!
    { (
        $escape_strings
        |
        [^{}]|(??{$curly})
      )*
    }
    !x;
     
    if ($Jscript =~ m!(\bfunction\s+$funct\s*$round\s*$curly)!) {
        print "Matched :\n",$1;
    } 
}

__DATA__

//.. function foo goes here .......

      [download]
[reply]
[d/l]
Re: How do I match/extract a function from a JavaScript file?
by Shakya (Initiate) on Sep 30, 2007 at 17:35 UTC
    Following definitely works in "not so complex" scenarios 
    #!/usr/bin/perl
use strict;

my $Jscript = join (" ",<DATA>);
# Extract function foo from Jscript
my $content = _extract("foo",$Jscript);

sub _extract {
    my ($funct,$Jscript) = @_;

    my ($round,$curly);
    $round = qr!\(([^()]*|(??{$round}))*\)!;
    $curly = qr!\{([^{}]*|(??{$curly}))*\}!;

    if ($Jscript =~ m!($funct\s*$round\s*$curly)!ms) {
        print "Matched :\n",$1;
    } 
}

__DATA__

foo (a, b, c) {
  if (c) { a++; } else { b--; }
  print "hi";
}
fum (d, e, f, g, h) {
  if (d) { e++; } else { f = g + h; }
  print "ho";
}

    [download]
    Output is : 
    Matched :
foo (a, b, c) {
   if (c) { a++; } else { b--; }
   print "hi";
}

    [download]
    On more explanation on matching nested parentheses, you can refer Friedl's Mastering Regular Expressions, page 330.
[reply]
[d/l]
[select]
        sub _extract {
    my ($funct,$Jscript) = @_;

    my ($round,$curly);
    $round = qr!\(([^()]*|(??{$round}))*\)!;
    $curly = qr!\{([^{}]*|(??{$curly}))*\}!;

    if ($Jscript =~ m!($funct\s*$round\s*$curly)!ms) {
        print "Matched :\n",$1;
    } 
  }

      [download]
       err, I can't see where you've matched the function keyword or the function operator anywhere in your code. That would seem to be a prerequisite to parsing out functions from ECMAScript source.

        foo (a, b, c) {
    if (c) { a++; } else { b--; }
    print "hi";
  }
  fum (d, e, f, g, h) {
    if (d) { e++; } else { f = g + h; }
    print "ho";
  }

      [download]
      That's not valid EMCAScript.

      -David.

[reply]
[d/l]
[select]

Back to Seekers of Perl Wisdom