Port to ClojureScript.
Speedups for (simple-benchmark [r (range 100000)] (count (partition-by odd? r)) 10)
V8: 0.91
SpiderMonkey: 0.87
JavaScriptCore: 0.81
Nashorn: 0.96
ChakraCore: 0.87
GraalVM: 0.74
Patch LGTM. It is a little surprising that changing a seq to lazy-seq slows things down so much for the benchmark above. Otherwise the change passes on all of the Canary tests.