buttersink uses same as snapshot as basis

[amescornish]

buttersink uses same as snapshot as basis

[AmesCornish]

Simon,

Thanks for the feedback. I specifically added some intelligence to buttersink to find the best set of diffs to use, which is not always simple chronological order, though it considers chronological order as one of the factors. It also looks at the total and "exclusive" size of each snapshot, and will sometimes measure actual diff sizes. It tries to minimize the total size of the diffs, but also to optimize other factors, like a long "chain" of diffs is less reliable that a single diff. If you have a specific example where the optimization seems way off to you, then it would be helpful to get a debug log (using the --logfile option), and it might help me to improve the algorithm.

Thanks!