As long as you're doing that ...
$ perl -e '$s="JustAnotherPerlHacker\n"; $s=~s/(?<=.)(?=[A-Z])/ /g; pr
+int $s'
Just Another Perl Hacker
By using zero-width assertions on both sides, we don't need to actually match anything. We're matching, literally, zero characters - we're matching simply a location with no length associated. Note, however, that I would make it just a bit different for a bit more flexibility:
$s=~s/(?<=\S)(?=[[:upper:]])/ /g
Match the location between a non-space and a capital letter. Using a character class gives us flexibility for other languages.
This may also be useful for splitting:
$ perl -e '$s="JustAnotherPerlHacker\n"; @a = split /(?<=\S)(?=[[:uppe
+r:]])/, $s; print "@a"'
Just Another Perl Hacker
Since split is splitting on nothing, there is nothing actually removed.
