Help interpreting this code

[Itatsumaki]

Howdy monks,

I came across this code in an open-source module today, and I've been trying to understand what it does. I'm refactoring the model so it works under -w and one line (marked in code with # GIVES WARNING #) gives:

\1 better written as $1 at TempFileNames.pm line 209.

I wanted to make sure I understood the whole sub (not just the reg-ex) before making any changes. The author is unavailable for a few days, so I thought I'd get some guidance on what's going on here faster. All comments are direct from the code, and aren't mine.

sub normalizedPath {
my ($path, $sep, $doSlashes, $beURLaware) = @_;
# URLawareness means that m{[a-z]+//:} will be skipped

# if ($beURLaware) then exclude the double slashed url-qualifier
if ($doSlashes)    {
   if ($beURLaware) {
      my ($type, $protocol, $urlPath) =
         ($path =~ m{(([a-z]+:)?//)?(.*)}o);
      $urlPath =~ s{/+}{/}go;
      $path = $type.$urlPath;
#     $path = ($protocol ne ''? $protocol: 'http:')."//$urlPath";
#     if ($beURLaware) {
#         my ($type, $urlPath) = ($path =~ m{([a-z]+://)?(.*)}o);
#         $urlPath =~ s{/+}{/}go, $path = $type.$urlPath;
      }
   else {
      $path=~s{/+}{/}go;
      }    #mult slashes are like single slashes
   }
   # . elimination before .. dt .. must ignore any .
   #    eliminate . components: here seems to be some bug <b>
$path=~s{(^|/)\./}{\1}og;  #g option is safe here (dt limited scope)
# Line below gives warning
while ($path=~s{(^|/)[^/]*/\.\./}{\1}o) {} #<A><b> suceeding '..' prev
+ent g option
# Line above gives Warning
return $path;
}

Any thoughts on rewriting this? I'm not quite happy to just change the \1 to $1, without understanding what's going on.

-Tats