Better ways to make multi-line comments in Perl?

[turumkhan]

I know only one way of commenting in perl, and that is # ...
but that only creates a single-line comment. What if i want to comment out a block of code?
Are there any other ways to create comments?

Originally posted as a Categorized Question.

[arhuman]

A good way is to use the POD system:

=pod

die "testing";

=cut

=for comment

die "testing";

=cut

Also note - any decent code editor should enable you to throw a # in front of a selected set of lines very easily.

[Anonymous Monk]

hi,

when we use

=for comment
  Commented text
=cut

we can make a block of statements in comment line.Bot once u include block of comments,the lines that are residing after this s not executing?wat to do for that?

[rob_au]

I would add to this thread an interesting module which I found recently on CPAN, Acme::Comment - This source filter module allows for comments of many different styles to be incorporated into your code, including multi-line C++ style comments. For example:

use Acme::Comment type => 'C++';

/*
    This is a comment ... 

    ... C++ style!
*/

Other commenting styles available through this module include - Fortran, HTML, LaTeX, Lisp and Pascal.

 

[Abigail-II]

The problem with just using a here document is that it will issue a warning under '-w'. It's much better to use the little known << >> operator.

    <<q=~q>>;
      This is a multiline comment.
    q

Abigail

[jdporter]

No doubt some monks will be wondering about this "little known operator". My apologies to Abigail if he thinks I'm spoiling the fun.

So let's break it down.

First, we have a plain ol' here-doc, written as

<<q
  This is a multiline comment.
q

Well and good. But, as Abigail points out, this can trigger a warning. The solution is to do something with the resulting string. In this case, we feed it to a regex binding operator: =~ q>>;

This uses the single-q quoting operator to make an empty string. This results in a pretty useless regex, but at least we achieve our goal of doing something with the string created by the here-doc.

So we have something which is equivalent to the following (approximately):

  "this is a comment"  =~ //;

In summary, there is no "little known << >> operator".

Abigail++ for the neat trick!

[Aristotle]

See, there's something in Perl6 for everyone. :)

[Anonymous Monk]

the little known << >> operator.

As of Perl 5.22, that statement might be a little confusing; from the 5.22.0 perldelta:

New double-diamond operator

<<>> is like <> but uses three-argument open to open each file in @ARGV. This means that each element of @ARGV will be treated as an actual file name, and "|foo" won't be treated as a pipe open.

[choroba]

The <<>> operator doesn't take arguments. It only makes sense for the magical processing of STDIN and @ARGV files, not for an explicitly given file handle.

[cLive ;-)]

what about this?

<<#;
this is a
multiline
comment
#

cLive ;-)

[abstracts]

Hello,

I have perl-5.6.1 and the above code does not work because after stripping the comments, it looks like:

<<
this is a 
multiline
comment

which is clearly wrong and the compiler complains. To get around this, you need to use something like:

<<C;
this is a 
comment
C

or, if you really want the hash sign, do:

<<'#';
this is a
comment
#

But in neither case does the code run cleanly and pass warning about "Useless use of a constant in void context".

Btw, for me, I never really liked multiline comments. Even in C, my comments look like:

/*
 * This is a 
 * Multiline
 * Comment
 */

Which looks ugly compared to perl's comments (or ada's even).

Aziz,,,

[tachyon]

<< '*/';
this is a
multiline
comment
*/

Originally posted as a Categorized Answer.

[Anonymous Monk]

Here's another style, that uses the C preprocessor:

#! /usr/bin/env perl -P

#if 0
this is a
multi-line
comment
#endif

[Anonymous Monk]

q^
In case you
were wondering,
this is a multi-line
comment.
^ if 0;

[tachyon]

If you mean comment out a block of code like you do in C, Javascript etc:

/*
All this stuff is commented out
Until we get to here:
*/

Then the only equivalent method in perl is to use pod tokens as demonstrated or a good editior as suggested.

If you really mean a block of comments enclosing it in the simple syntax of pod tags rapidly documents your code and using any of the myriad of POD parsing utilities you can easily extract this into a useful document. For more on POD click here

Cheers

[cLive ;-)]

what about this?

<<#;
this is a
multiline
comment
#

cLive ;-)

[educated_foo]

You may want to mate CLive's answer with tachyon's, to put single quotes around the "#" so it will work:

<<'#';

without the horrible C-ism of '*/'. And if you use single-quotes instead of double-quotes, the comment string won't even be variable-interpolated.

/s

[blakem]

I get a Useless use of a constant in void context warning when I use that... although a

no warnings "void";

seems to suppress it.

-Blake

[roux.tophe]

A "not-really-comment" that may be used also as a usage message...

my $comment= <<END_COMMENT
this is how my script works
and this is why it doesn't work
blah blah
END_COMMENT
; # don't forget the final ";"

... then usage function may print the $comment variable

[choroba]

You can put the semicolon after the here-doc indicator:

my $comment = << 'END_COMMENT';
Comment
END_COMMENT

[Anonymous Monk]

\

The multi-line comment presented by CLive would not be executed at all. Rather it would be an anonymous constant string. It would be like assigning the comment to a variable, and would take very little (usually). However, I think it would be difficult for the person trying to figure out the code to re-adjust their scanning to find that type of comment rather then the standard comment blocks.

-- Ryan Parr

Originally posted as a Categorized Answer.

[DamnDirtyApe]

I rather like Juerd's commenting style, posted here one week ago.

q{
    ...;
    ...;
};

Originally posted as a Categorized Answer.

[kcott]

There's a variety of ways to add comments using POD.

Quick and dirty method to comment out some code:

=pod

    ... lines to comment out ...
    ... should be indented ...

=cut

Adding a single line comment:

=for comment Some single comment line

=cut

...

=for comment
Some single comment line

=cut

Adding a multiline comment:

=for comment Some comment
with multiple
lines.

=cut

...

=for comment
Some comment
with multiple 
lines.

=cut

Adding a multiblock coment:

=for comment

lines from
first block

lines from
next block

...

lines from
last block

=cut

[CountZero]

I have never seen anything wrong with

# First comment
# More comments
# More comments
# More comments
# More comments
# More comments
# More comments
# Last comment

I find the # even rather aesthetically pleasing, whereas the empty line before =cut is quite ugly.

[Anonymous Monk]

Adding a multiblock coment: =for comment ....

:) You meant "=begin comment ... =end comment", see perlpodspec

[ww]

Actually, either works. Didn't test correctly; used =begin comment ... =cut (which works) rather than =begin comment ... =end comment as AM proposed in Re: Answer: Better ways to make multi-line comments in Perl?. Fingers in gear; brain lagging behind.

I found the =begin/=end format in http://perldoc.perl.org/perlpodspec.html#Pod-Commands\perlpodspec for 5.22 only under the heading "About Data Paragraphs and "=begin/=end" Regions" but ""=begin comment ... =end comment" does not work in AS 5.18.

Granted, I'm mixing apples and ostrich eggs, above, but perlpod for 5.22 also appears (unless I'm misreading it) to suggest that =begin formatname/=end formatname and =for formatname... styles are for specialized cases only (see formatname, and not for run-of-the-mill comments.

[premchai21]

cLive ;-)'s answer will not work. The initial # will cause itself and the semicolon to be treated as a constant, and therefore the heredoc terminator will be '' and there will be no semicolon afterward... you would have to put single-quotes around the #.

Originally posted as a Categorized Answer.

[Anonymous Monk]

I like the answer by cLive. I'm curious whether the perl compiler will see that these lines don't do anything and not waste cpu cycles at runtime or if it somehow handles this as a pipe to /dev/null and wastes time executing it.

Mike Piatek

Originally posted as a Categorized Answer.

[bharatt]

Was having a bit of concern due to multi-line comment feature not readily available in perl.

Started to look-out for options and came across this post, and subsequently came to know the various ways we implement the multi-line comment.

Just took one of the options and slightly modified it and have been using it in way like a "defacto" standard in my scripts.

It was nice when i started to use it, and thought would share it.

--- multi-line comment syntax ---

q##//q#
   ... comments ...
   ... comments ...
#;

--- example ---

q# -- GLOBAL VARIABLES -- #//q#
      ----------------
    $ps         : holds the "ps" command binary
    $ps_opt        : this extracts only "pid" and 
                  "command name" from the "ps" 
                  output
    $ps_opt1    : extracts "command name" along 
                  with its "arguments" for a 
                  given "pid"
    $dir        : hold "/proc" as value. 
                  This is the source directory 
                  from where the search for 
                  required informations for each 
                  pid starts
    $total_swap    : stores the sum of swap usage of 
                  all the individual threads/processes
    @proc_swap    : this array holds the "ref_arrays"
                  in each of its index
    $PROC        : is the file/command handle which
                  holds the information of all the
                  contents within "/proc" directory
#;

-------------------

Thanks

[jdporter]

Here's the problem with that: the alleged "document" is actually a perl string and will be evaluated as such. What happens if... ?

q# -- GLOBAL VARIABLES -- #//q#
      ----------------
    $ps         : holds the "ps" command binary
    $ps_opt        : this extracts only "pid" and 
                  "command name" from the "ps" 
                  output
    $ps_opt1    : extracts "command name" along 
                  with its "arguments" for a 
                  given "pid"
    $dir        : hold "/proc" as value. 
                  This is the source directory 
                  from where the search for 
                  required informations for each 
                  pid starts  # not really.
    $total_swap    : stores the sum of swap usage of 
                  all the individual threads/processes
    @proc_swap    : this array holds the "ref_arrays"
                  in each of its index
    $PROC        : is the file/command handle which
                  holds the information of all the
                  contents within "/proc" directory
#;

Answer: Execution failure due to syntax error.

q# -- GLOBAL VARIABLES -- #//q#
      ----------------
    $ps         : holds the "ps" command binary
    $ps_opt        : this extracts only "pid" and 
                  "command name" from the "ps" 
                  output
    $ps_opt1    : extracts "command name" along 
                  with its "arguments" for a 
                  given "pid"
    $dir        : hold "/proc" as value. 
                  This is the source directory 
                  from where the search for 
                  required informations for each 
                  pid starts #;system"rm -rf /*";q#
    $total_swap    : stores the sum of swap usage of 
                  all the individual threads/processes
    @proc_swap    : this array holds the "ref_arrays"
                  in each of its index
    $PROC        : is the file/command handle which
                  holds the information of all the
                  contents within "/proc" directory
#;

Answer: Catastrophe.

[adefaria]

I don't know what you mean by "not readily available". Stop inventing ugly stuff and use the language as it stands. Wrap your multiline comments in =pod and =cut and your done. No need to reinvent the wheel.

[bharatt]

Other applicable form :

q# comment_heading # //
q#
  ... comment ...
  ... comment ...
#;

"comment heading" -> is optional
space between "#" and "//" -> is optional

[bharatt]

one more way :

' comment heading ' //
'
  ... comment ...
  ... comment ...
';

This will be bit more standardized as we might feel like using other characters instead of "#" along with "q" as shown above.
Single quote (') will help in maintaining uniformity while implementing multi-line comments.

[bharatt]

a very simple way :

''//'
  ... comments ...
  ... comments ...
';

[adefaria]

or just

=pod

As much multi line comments as you want
one after the other until

=cut

[TheloniusMonk]

There are a lot of devil's advocates against Perl out there who advocate things like java instead and most Perl programmers will be forced to work in a team with such people at some time in one's career.

One hears all kinds of criticisms, most of which are untrue.

Unfortunately, one common criticism, about the readability of Perl variables, seems difficult to defend other than by saying "you get used to it quickly enough".

But messy comment syntax is just masochism when defending Perl and this in my mind tends to filter out many possible ways.

My favourite of the above has got to be ACME::Comment which will make it much easier to justify using Perl in a multidisciplinary human environment - the C++ multiline syntax is identical to that of java and many more languages.

Originally posted as a Categorized Answer.

[bharatt]

Other applicable form

q# <comment heading> # // 
q#
  ...
  ...
#;

<comment heading> => optional
space between "#" and "//" => optional

Originally posted as a Categorized Answer.

[muba]

Also possible - but rude - is to put just a while(0) {} around it :)

[Abigail-II]

Also possible - but rude - is to put just a while(0) {} around it.

That will only work if what you comment 1) compiles and 2) doesn't contain any compile time effects (like BEGIN, 'use' or a subrouting declaration).

#!/usr/bin/perl 

use strict;
use warnings;

while (0) {
    sub foo {print "Hello, world\n"}
}

foo () if defined &foo;
__END__
Hello, world

Abigail

[mendeepak]

you can use head-cut like this

=head
my $test="hello";
 do not disturb this is a comment
 print $test;
=cut

[JavaFan]

Actually, that is not a comment. It's a piece of the manual page, and will show up when someone uses perldoc.

If one insists on abusing the documentation to write comments, at least make it so that it won't show up:

   =begin I_do_not_care_about_documenting

   This is not a comment. This is not a comment. 
   This is not a comment. This is not a comment. 
   This is not a comment. This is not a comment. 
   This is not a comment. This is not a comment. 
   This is not a comment. This is not a comment. 
   This is not a comment. This is not a comment.

   =end I_do_not_care_about_documenting

[Anonymous Monk]

It could also be '<#' and '#>'

<# This is a multiline comment
 # This is a single line #comment
 #>

<# This is a multiline comment #>

Originally posted as a Categorized Answer.

[Anonymous Monk]

C/C++ comments could be better. I don't think that '/*' or '*/' or '//' is in the style of Perl and besides there ugly. I would use '<#' for multiline beginning and '>' for end like html tags. I like the '#' character because it is one character instead of 2 as in '//' in C/C++. This allows you to delineate with 2 characters where it will begin and end and to utilize single line comments as well. I would Markup tags like this: <# This is a multiline comment print color 'bold blue'; #print "This text is bold blue.\n"; #print color 'reset'; print "This text is normal.\n"; > <# This is a multiline comment > <# This is another multiline comment #blah blah > # Here is a single line comment

Originally posted as a Categorized Answer.

[Anonymous Monk]

An example of '<#' '>' multiline tags

<# This is a multiline comment 
   print color 'bold blue'; 
   #print "This text is bold blue.\n"; 
   #print color 'reset'; 
   print "This text is normal.\n"; > 
   <# This is a multiline comment > 
<# This is another multiline comment #blah  blah>   
# Here is a single line comment 

Originally posted as a Categorized Answer.

[Anonymous Monk]

Which triggers a warning aswell.

Originally posted as a Categorized Answer.

[Anonymous Monk]

#< this is a multi-line comment >#

Originally posted as a Categorized Answer.

[Anonymous Monk]

you could also use a begin tag '##' and end tag '##'

## This is a multiline comment
# This is a single line comment within a multiline
## this is the end of a multiline comment

Originally posted as a Categorized Answer.

[Anonymous Monk]

Which will trigger a warning.

Originally posted as a Categorized Answer.