Yes, bit ops are available in Perl, but I don't find them a natural way to express this particular problem. It'd've taken a lot more effort to figure out your code if I hadn't used them extensively in assembler, back in the day. Just about every use of bit ops I've ever seen obscures rather than clarifies what's going on. In nearly all cases the motivation is microoptimization. That's what I got the impression was what you were trying to do, from the looks of your code.

As for the other points: apologies, I did more than just replace the bit ops by equivalents and didn't test sufficiently. I even wrote a note that the code probably contains off-by-one errors, but then tested it a bit and it seemed to be ok. The two issues are orthogonal though, it's the other changes in the structure I made that introduced the problems, not the bit op replacement.

As for performance in general, I think it would be worthwhile to do this in XS and stick it into List::Util.