Perl script gives a blank output file - regex

Question

I'm a total noob at Perl, trying to learn some new code for a specific project. In short, I'm making a script (on osx) that is to search all xml-files in a folder and censor specific numbers. I know a one-liner could have helped, but the amount of files will be pretty huge (thousands of files), and would happen regularly so a script to do it would be nicer. And besides, there is the learning to script part :)
I've managed to open my files, make the regex work on every line on the original for my specific needs and generate a writable tempfile for my new information. This is where things stop working. I've tried to copy the new file over the old file after the loop, but I end up with a blank(!) file. I suspected there to be an error with the temp-file, but that looks perfect. I even tried, as a noobs way out, to reverse the process line by line from the temp back to the original file after changing the open mode (read) on them, but that ALSO gave an empty file.
And now my head is sort of empty. Any help would be appreciated :)
#!/usr/bin/perl
use strict;
use warnings;
use File::Copy;
chdir "/perltest/test"; #debugsafety
#file
my $workingfiles = "*.XML";
my #files = glob("$workingfiles");
#process files
my $old;
my $tmpfile;
foreach my $file (#files) {
print "$file \n";
open ($old, "<", $file) or die "No file";
open ($tmpfile, ">", 'temp.tmp') or die;
while(my $line = <$old> ) {
my $subz = $line;
$subz =~ s/([[:upper:]]{2}[[:digit:]]{6})|([[:upper:]]{1}[[:digit:]]{7})|(?:(?<![[:digit:]])[[:digit:]]{8}(?![[:digit:]])|([[:upper:]]{2}[[:digit:]]{5}[AB]))/**CENS**/g;
print $subz;
print $tmpfile $subz;
}
print "Start copying.\n";
open (my $old, ">", $file) or die "No file";
open (my $tmpfile, "<", 'temp.tmp') or die;
#copy $tmpfile, $old or die "Couldn't copy";
my $y = 0; #debug
while (my $line = <$tmpfile> ) {
print $y++; #debug
my $subz = $line;
print $subz;
print $old $subz;
}
}
print "Complete.\n";
exit;

Answer 1

You re-open your file handles before closing them. I'm an Oracle DBA masquerading as a perl developer, so I can't give the why behind it. But I know if you close your file handles, your script should work as is.
close ($old); # add this line
close ($tmpfile); # add this line
print "Start copying.\n";
It would then be good practice to close them again when you are done "copying" back to them.

Answer 2

Explicitly close the filehandle when you're done writing to it. Things will still be buffered until you do that.
Also would make more sense to
rename($file, "$file.old");
rename("temp.tmp", $file);
rather than looping through the file (or using File::Copy::copy) to make a backup copy of it.
Lastly, for simple edits I could suggest making an effort to get comfortable with doing it on the command line so you don't need scratch your head and wonder "now what did I do in that script last time?". It can be a big timesaver in the long run.
perl -p -i.bak -e 's/pattern/text/;' files*
is the general form.