Others already pointed out the scalar context trick. There are also non-destructive methods available to count the occurence of a certain character. You can find this in perlop as well. Depending on what you need exactly, you can use different solutions:

$txt = "This is a string";
$count = $txt =~ tr/ //; #counts all spaces
$count = () = $txt =~ m/\s/g;   #counts all whitespace characters
$count = () = $txt =~ m/\s+/g;  #counts whitespace occurences (default
+ [split] behavior)
[download]

It's probably a good idea to check out 7 Stages of Regex Users as well. Oh yeah, Don't be confused by the =~. It's not only an assignment, it's also a function that returns something, just like when it acts on $_. See perldoc, illustrated by:

$count = tr/ // for ($txt);
[download]

Does exactly the same, but now the match is done on $_. $_ is assigned to by for. Don't assign to $_ yourself (or use local). A complete different approach still is the use of index, which may be handy under certain circumstances.

Hope this helps,

Jeroen
"We are not alone"(FZ)
Update: I must have been sleeping.... read merlyn's reply... code is working now.