Thanks for the further info. I ended up going with a similar case to your second suggestion (m/\G(\\\$\w+|\$\w+)/gc, though I didn't repeat the \w+ since there was other stuff in the regex after that) and it seems to be working fine.

The s/// solution seems like it'd be quite expensive, requiring many, many memcopy's. As I'm using this tokenizer fairly heavily, I don't think that's really a good idea. Though, of course, maybe Perl's smarter than I'm giving it credit for. ;-)

Hopefully pretty soon I'll be able to post the code I'm working on here for some peer review, as well as a link to a "showcase" page. It's getting pretty close to finished and I really think it's pretty cool. I'm looking forward to getting everyone's opinion on my code and the ideas behind it. I've really appreciated your help here, MeowChow, and I hope that you'll give it a look when I finally get it all posted. =)

For now, I think I should really go to bed. It's 6am here and I have to get up in 6.5 hours to get ready for work (so that I can go in and work on this project some more! :-). Have a good night/day/whatever, everyone and thanks again!

bbfu
Seasons don't fear The Reaper.
Nor do the wind, the sun, and the rain.
We can be like they are.