Alternation in Effective Perl Programming Example

[Carl-Joseph]

On page 72, the book Effective Perl Programming shows some code to remove C-style comments from a string. I've reproduced the code below.

local $_;

$_ = "
This_Is_Code = 0; /* This is a comment */
/*
 * Can it handle a mult-line comment, i.e.
 * one that goes on for more that one 
 * line? Yes, it can.
*/
This_Is_Code++;
";
                     
for (split m!("(:?\\\W|.)*?"|/\*|\*/)!) {
   if ($in_comment) {
      $in_comment = 0 if $_ eq "*/"
   }  else {
      if ( $_ eq "/*" ) {
         $in_comment = 1;
         print " ";
      } else {
         print;
      }
   }
}

I have a question about the regular expression that's used in the call to split(). The expression (:?\\\W|.) seems to say match either a non-word character, or any character. Why have both patterns in the alternation when the dot includes non-word characters?

Thanks,

Carl-Joseph

[chromatic]

It's a little more than that. Should it be (?:\\\W|.) instead?

If so, as written, it would be interpreted as non-capturing parenthesis, matching either a backslash and a nonword character or a single character.

Otherwise, it's (an optional colon followed by a backslash and a non-word character) or (a single character).

[a]

Appears to be (looking at Effective perl) a typo, as the chapter leans heavily on (?: just before the example. You should mention it to merlyn as its not in the errata (neither is the misspelled 'Eart' on page 64 ;-)

a

[a]

And, to the original question: I think the idea is to 'inch along'

  for (split m!("(?:\\\W|.)*?"|/\*|\*/)!) {

will return the pieces ending w/ an escaped non-word char, a single char or the comment begin/end markers. Hmm, an escaped non-word char or a single char in quotes? No, as many \\\w|. as found between quotes. I guess so:

biff = "not a comment /*";
/* marker for not a comment */

won't start $in_comment too early.

a