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

I need to insert a string only once after matching the comments. This should be trivial but maybe I'm too tired to think it through. :( There are a few hundred files and they contain different syntaxes for the comments. I found a similar question at this node but the solution would match multiple times. My code currently updates the first line of the file and I want it to update after all the current comments. Thanks for any tips! 
open (IN, 'file.txt');
my $out = '';
my $newComment = 0;
while (<IN>) 
{
    if(($newComment < 1) && ($_ ne /^\#|^--|^\/\/^\s*\/\*/|^\s*\*|\*\/
+$/))
   {
     $out .= "#This file was updated with the fix"; 
     $out .= "\n";
     $newComment++; 
   }   

    $out .= $_;
}
open (OUT, '>output.txt');
print OUT $out or die "Can't print to OUT: $!";
close OUT;
close IN;

__DATA__
/*
* This is the new function
 * It has the following parameters:
 *
 * 1. Filename 
*  2. URL
*/  
# All the following are valid comment styles:
# hello perl style
-- hello sql style
/* hello c style
 * continue c style
   end c style */
// c++ c style
#Revision date: 2005-10-10
--start of file
<--  here is where the new comment should go.

configurationtype=search
url=http://www.google.com


--End of file

[download]

Replies are listed 'Best First'.
Re: inserting string only once after matching a pattern
by ikegami (Patriarch) on Nov 06, 2006 at 06:31 UTC
    Many problems:
    • You have an unescaped / in the regexp. Maybe you should use m{...} instead of /.../.
    • There's a | missing in the regexp.
    • The regexp would be more readable if you used the x option and added whitespace. 
      if (!
   m{
      foo
   |
      bar
   |
      baz
   }x
) {
   ...
}

      [download]
    • $_ ne /.../ should be $_ !~ /.../, !($_ =~ /.../) or just !/.../.
    • You read the file a line at a time, yet you have no flag telling you if you're in the middle of a multi-line /*...*/ comment.
    • Why do you accumulate the text in $out instead of printing as you go along?
[reply]
[d/l]
[select]
      Thank You!

      I used the m{} syntax and fixed the regexp issues. I'm actually posting on a different computer from the code so the regexp issues were mistyping it.

      I'll study the different not equals soon. I didn't realize the the ne was the wrong syntax. I actually had (!~) initially but I read on a webpage that it's better to use "ne".

      At first I wasn't sure why I would need flag to determine if I'm in a multi-line /*...*/ comment. The syntax used in the files, always has a * as the first visible character in multi-line comments. Although you're right that it's not required. That's a nice catch. I haven't programmed in C in years.

      The only reason for accumulating the text is that the write commands are all together instead of spread out. The actual logic of the write operations take up a page of text and it's nice to treat the write operation as a single block. Since the files are small, avg. is 7k, and the largest are no more than 100k, memory is not a problem. Is this considered bad form in perl?

[reply]

        I actually had (!~) initially but I read on a webpage that it's better to use "ne".

        When comparing against a constant strings, it's more efficient to avoid regexps. 

        $var eq 'abc'
$var ne 'abc'
$var1 eq $var2
$var1 ne $var2

        [download]
        are faster and simpler than 
        $var =~ /^abc\z/
$var !~ /^abc\z/
$var1 =~ /^\Q$var2\z/
$var1 !~ /^\Q$var2\z/

        [download]

        I'll study the different not equals soon.

        $var ne /regexp/
        means
        $var ne ($_ =~ /regexp/)

        • When matching against a regexp, use a match operator (=~ or !~).
        • When comparing strings, use a string comparison operator (eq, ne, lt, gt, le, ge or cmp).
        • When comparing numbers, use a numerical comparison operator (==, !=, <, >, <=, >= or <=>).
[reply]
[d/l]
[select]
Re: inserting string only once after matching a pattern
by artist (Parson) on Nov 06, 2006 at 11:38 UTC
    You might want to use Regexp::Common to extract comments from the code.
    --Artist
[reply]

Back to Seekers of Perl Wisdom