Greetings, O monks;

I am fairly new to Perl, and this is my first submitted question. I seek wisdom on a problem that looks extremely simple, but has been causing me a lot of grief.

I wish to parse a string to find all-caps words of at least 3 characters. My example code looks like this:

my $finalline = "3.2   COMPLIANCE WITH LAWS AND REGULATIONS. While the
+ Product is in its possession or under its, or its sub-contractor's c
+ontrol, THEDUDES shall comply, or ensure that its subcontractors comp
+ly, with all applicable federal, state and local statutory and regula
+tory requirements regarding the manufacture, if applicable, packaging
+, handling transportation and storage of the Product.";
@words = ($finalline =~ /([A-Z][A-Z][A-Z]+)/g);
print ("@words \n");
[download]

This returns:

COMPLIANCE WITH LAWS AND REGULATIONS THEDUDES
[download]

...as desired.

However, when I try to plug this into my larger program, the code behaves oddly. Specifically, $finalline has the desired content, but @words usually ends up empty. The especially odd thing is that if I remove the "/g" above, @words will retrieve the first capitalized word, but if I put it back in, it will retrieve zero capitalized words(!) I have also tried adding "/gc" instead, without success. The code does more or less what it's supposed to apart from this bug (and a couple of others).

Here is my code. As far as I can tell, the relevant bits (lines 150-153, marked as "#PROBLEM CODE!!!" below) are exactly the same as in my example, but do not work correctly. Any help would be greatly appreciated.

(Addendum: I have since gone back and turned on "use strict", and declared all my variables ahead of time. This has not solved the problem.)

#!/usr/bin/perl -w

use diagnostics;
use Spreadsheet::WriteExcel::Big;
use HTML::Restrict;
use File::Slurp;
#use Win32::Word::Writer;
use RTF::Writer;

opendir(DIR2, "contract/new");
my @files2 = readdir(DIR2);
closedir(DIR2);

foreach $file2 (@files2) {
 print "$file2\n";
 open (FH2, "contract/new/$file2");
 $newfile2=$file2;
 $newfile2 =~ s/\.html//g;
 my @filelines2 = <FH2>;
 chomp @filelines2;
 my $masterplan = join(' ', @filelines2);
 $hr2 = HTML::Restrict->new();
 $masterplan = $hr2->process($masterplan);
 my @masterarray = split('DUMMY',$masterplan);
 my $caps = 0;
 my $traps = 0;
 my $escape = 0;
 my @words = "";
[download]

foreach $finalline (@masterarray){
     $escape = 0;
     $finalline =~ s/\&.{3,5}\;/ /g;
     $finalline =~ s/\s+/ /g;
     while ($finalline =~ /(\w\w+\b)/gc){
        $linelength++
     }
        if (($finalline =~ /^\s*\d|^\s*Section|^\s*Article|^\s*[A-Z]\.
+|witness whereof/i) && ($holder == $rtf)){
        while ($escape == 0){
            if (($finalline =~ /[A-Z][A-Z][A-Z]/) && ($caps == 0) && (
+$traps == 0)){
               print(">>>$finalline<<</n");
               @words = ($finalline =~ /([A-Z][A-Z][A-Z]+)/g);  
#PROBLEM CODE!!!                      
               print("@words");                                  
#PROBLEM CODE!!!                                       
               $lineholder = $finalline;                         
#PROBLEM CODE!!!                                            
               $finalline = join(' ', @words);                   
#PROBLEM CODE!!!                                    
               $caps++;
}
#etc.
[download]

Update: Removed extraneous code; added new line (print (">>>$finalline<<</n");)

Update 2: smls's solution works. Thanks smls!

Update 3: Restored problem code (see below).