From: pete@... (Peter J. Kernan)
Date: 2001-08-24T02:28:24+09:00
Subject: [ruby-talk:20192] Re: [q] question about ruby regex

On Thu, 23 Aug 2001 12:38:40 +0900, Dave Thomas <Dave@PragmaticProgrammer.com> 
  was ready with:
||"丞 旦 枢" <cwyang@aratech.co.kr> writes:
||
||> Hi.
||> 
||> I've been trying to match C-style comment in ruby regex.
||> 
||> my wrong approach is as follows.
||> 
||>       case str
||>       when /\/\*[^(\*\/)]*\*\//
||
||you could try:
||
||        when %r{/\*.*?\*/}m
||

 that will also grab non-comments, such as  printf("/* not a comment */\n");

 i once wrote a C comment grabber in perl which i think is correct (passed all
 of the tests i could think of), but havn't bothered to convert it to ruby re's 
 yet. here it is:

my $even_slashes = qr{(?<!\\)(?:\\\\)*};
my $not_dq = qr{[^"\n\\]*(?:\\.[^"\n\\]*)*};    #not " nor escaped "
my $not_q = qr{[^'\n\\]*(?:\\.[^'\n\\]*)*};     #not ' nor escaped '
my $dq_string= qr{${even_slashes}"$not_dq"};    #double-quoted string 
my $q_string= qr{${even_slashes}'$not_q'};      #single-quoted string
my $comment = qr{(/\*.*?\*/)}s;                 #comments

my $re = qr{(?:$dq_string|                      #gobble non-comments first
                $q_string|                      #
                $comment)+                      #
           }x;
local $/;
while (<>) {
  print "$_\n" for grep {defined} /$re/g;
}


  Pete
-- 
screwbean: a shrub or small tree found in the US Desert Southwest