Your C-style for loop runs from zero to minus one, because $#efile refers to the last index of the array @efile, which is created on the spot - empty. Your task will be much simpler if it is written using the idiomatic $_ variable.

#!/usr/bin/perl

use warnings;
use strict;

my $input_file = 'dump.vcd';

open (INPUT, '<', $input_file)
    or die $!;

while (<INPUT>) {
    s/\$var\w \d+ ([!,\#,\",\$,\%,&]) (\w) \$end/\$var $1 $2/;
    print;
}

close INPUT;
[download]

I don't believe your character class is exactly what you think it is because of all the commas and escapes. I didn't modify the s/// statement because I didn't know its intent.

There are a bunch of errors and dodgy practices there. The first thing that you need to do is add strictures:

use strict;
use warnings;
[download]

Then you need to declare all the variables you use by prefixing them with my. For example:

my $input_file = "dump.vcd";
[download]

You should always use the three parameter version of open to improve code readability and reduce possible errors of various sorts:

open INPUT, '<', "$input_file";
[download]

The while loop is reading one line at a time so you don't need a for loop (see perlsyn - foreach) to iterate over anything - $efile is not an array.

In your regex you use a character set ([!,\#,\",\$,\%,&]), but it isn't what you possibly expect. The characters included in the set are !,#"$%& which includes comma, that may surprise you. Note that none of the quoted characters in the set need to be quoted.

The complete code rewritten and modified to take itself (on my system) as the input looks like:

#!/usr/bin/perl
use strict;
use warnings;

my $input_file = "noname1.pl";

open INPUT, '<', "$input_file";

while (my $efile = <INPUT>) {
    $efile =~ s/\$var\w \d+ ([!,\#,\",\$,\%,&]) (\w) \$end/\$var $1 $2
+/;
    print $efile;
}

close INPUT;
__DATA__
$var1 1 , x $end
$varx 2 ! _ $end
$vary 3 # aa $end
[download]

prints:

#!/usr/bin/perl
use strict;
use warnings;

my $input_file = "noname1.pl";

open INPUT, '<', "$input_file";

while (my $efile = <INPUT>) {
    $efile =~ s/\$var\w \d+ ([!,\#,\",\$,\%,&]) (\w) \$end/\$var $1 $2
+/;
    print $efile;
}

close INPUT;
__DATA__
$var , x
$var ! _
$vary 3 # aa $end
[download]

Note that I added the __DATA__ (see perldata - Special Literals) section so I could include something for the regex to match.

use strict;
use warnings;
[download]

The $ for $var and for $end is quoted so they are literal matches.

---

DWIM is Perl's answer to Gödel

perl -pi -e 's/foo/bar/g' file.txt
[download]

vcat is very handy in converting a VCD file into a much simpler format for further parsing.

The other simulator vendors may have similar capabilities. It might be worth a look if you have some heavy-duty parsing, just so you avoid re-inventing the wheel, as I have done far too often with VCD:)

Update (12 jan 2012): Verilog::VCD

I think the Verilog::VCD module is just a VCD parser. Is there any module out there that can make creating VCDs easier? Thanks.

You are right... Verilog::VCD can only be used to parse VCD files. It can not create a VCD file. I am not aware of any CPAN module which can create a VCD file. This looks like an opportunity for you to contribute.

Why do you need Perl code to create a VCD file? All Verilog simulators can create VCD files.