2.3.0: An _Even_ Faster Boat
After some really fantastic work by @jshier and @RuiAAPeres, we were able to track down some serious performance issues related to our (internal) use of reduce
. @jshier was able to remove all instances of reduce
in favor of using for in
loops, and ended up with a 2600% improvement in performance. This release is to get that into users hands while we work on moving towards version 3.0, which should introduce even more performance increases as well as reducing the complexity of the library.
- [FIXED]: Aforementioned performance improvement (Jon Shier and Rui Peres)
- [NEW]: We've audited the codebase to use
@noescape
where appropriate (Sid Raval) - [NEW]:
DecodeError
is nowHashable
(Matt Diephouse)