Test anxiety
--#--
Before we spot the flaw, let's do a little trick. Let's build a test 
harness for our regex. This is a sensible thing to do <i>whenever</i>
we're trying to build a regular expression matcher.<br>
<br>
<code>
#!/usr/bin/perl

open(FH, $ARGV[0]);
@strings = <FH>;
close FH;

foreach $string_to_test(@strings) {
	chomp($string_to_test);
	if ($string_to_test =~ m/^\d{0,3},?(\d{3},?)*\d{3}(\.\d{2})?$/) {
		print "$string_to_test:  PASSED\n";	
	}
	else {print "$string_to_test:  FAILED\n";}
}
</code>
<br>
Using this script, we can feed our regex candidate a list of strings.
This way, we can systematically alter the patterns we put in the 
string, copy and paste, etc.  In a few seconds you can build a list to 
test your regex in various ways, and do it systematically rather
than randomly guessing what will make it pass or fail.<br>
<br>
Here's another trick: Put all the strings you think will <i>pass</i> 
into one list and file, and the ones you think will fail in another. That 
makes it easier to spot the exceptions that prove you wrong.<br>
<br>
So, what's the flaw?<br>