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

Hello Monks, I'm a bit baffled by a disappearing string. Maybe someone can offer some insight. I have a a multi-line string that MAY have embedded full line comments from the start of the string. I want to remove the comments. Here is a sample:

my $string = "--This is a comment\n--This Comment continues here\nNOT a comment";

My desired output is: $string = "NOT a comment"

But when I run my code, I get $string = "--This Comment continues here" 

#!/usr/bin/perl -w
#use strict;

my $count = 1;
my $string = "--This is a comment\n--This Comment continues here\nNOT 
+a comment";

while( $string =~ m/^--.*\n(.*)/ ) {
    #Debug Output
    print "ORIGINAL $count: $string\n****\n";    

    $string = $1;

    #Debug Output
    print "\n+++\nMODIFIED $count: $string\n";    
    $count++;
}

print "DONE $count: $string\n";    

exit;

__OUTPUT__
+++
MODIFIED 1: --This Comment continues here
DONE 2: --This Comment continues here

C:\temp>perl strip.pl
ORIGINAL 1: --This is a comment
--This Comment continues here
NOT a comment
****

+++
MODIFIED 1: --This Comment continues here
DONE 2: --This Comment continues here

C:\temp>perl strip.pl
ORIGINAL 1: --This is a comment
--This Comment continues here
NOT a comment
****

+++
MODIFIED 1: --This Comment continues here
DONE 2: --This Comment continues here

[download]

Replies are listed 'Best First'.
Re: Removing nested comments with regexes
by Thelonius (Priest) on May 18, 2003 at 21:03 UTC
    The problem is that the dot "." does not match "\n". So after the first match, $1 = "--This Comment continues here" without the "\n" and following line. If you use the /s modifier, dot will match "\n", but then you have to be careful of matching too much, that is, the "^--.*\n" would then match many lines. There are several solutions to this, such as /^--.*?\n(.*)/s or /^--[^\n]*\n(.*)/s, but it might be simpler to do just this:
    while ($string =~ s/^--.*\n$//m) {
  print "string = $string\nDeleted $1\ncount=$count\n";
  $count++;
}

    [download]
[reply]
[d/l]
[select]

      However, your code will stop working if you consider

      my $string = "--This is a comment\nNot a comment\n" .
    "--This Comment continues here\nNOT a comment";

      [download]

      Comments do not necessarily occur at the beginning of the (rest of the) string. (At least that's how I understood the question.) I propose to simply do 

      my $string = "...";

$string =~ s/(^|\n)--[^\n]+//g;

print "No comments: $string\n";

      [download]

      If you're perfectionistic, you will notice that this will leave a leading newline if the string starts with a comment. You can fix that by using 

      my $string = "...";

$string =~ s/\n--[^\n]+//g;
$string =~ s/^--[^\n]+\n//;

print "No comments: $string\n";

      [download]
[reply]
[d/l]
[select]
      Thanks for the substitution idea!

      I removed the "$" and the from the regex since it wouldn't match the data correctly. Also in your example there is no longer a grouping on $1. Here is the fixed code: 

      my $count = 1;
my $string = "--This is a comment\n--This Comment continues here\nNOT 
+a comment;\n";

print "Original: $string\n";

while ($string =~ s/^--.*\n//m) {
  print "string = $string\ncount=$count\n";
  $count++;
}
print "Final: $string\n";
exit;

      [download]
[reply]
[d/l]
Re: Removing nested comments with regexes
by graff (Chancellor) on May 19, 2003 at 01:54 UTC
    Since the problem involves whole lines, you could split the multi-line string at line breaks, and use grep to ignore the lines you don't want: 
    @wanted_lines = grep !/^--/, split( /\n/, $string );

    [download]
[reply]
[d/l]

Back to Seekers of Perl Wisdom