Slab mapped, bug fixed

Okay, got Slab tempo-mapped ... in doing so found a bug that didn't come up when I did Cheap Thrills. That's why you have to test extensively, kids!! Turns out if I had several regions (between "key beats") in a row that were all exactly the same length to the sample, my interpolation algo would inadvertantly cause a division by zero. Can that really happen on a human performance? It certainly can if there's any editing done beforehand! So I'm glad this happened; glad I caught it early and fixed it.