Closed bovee closed 4 years ago
@boydgreenfield this is not likely the cause of regression as 0.1.1 has similar speed: https://app.zenhub.com/workspaces/primary-5c917e1983c46e630dc17e32/issues/onecodex/mgo/91#issuecomment-595156249
I ran the benchmarks again today and they are similar so I think we can merge that in the meantime.
This "fixes" the bridging over the mutable/immutable mmap issue we've had for a while (which is blocking e.g. read-only file access at the system level) by wrapping them in a enum. It also includes some combinatorial logic from upstream that's useful for e.g. RRR and includes a little more testing.
We're already depending on this upstream so we should probably merge into master, but it would be nice to do a cursory pref check to make sure everything runs at the same speed.