bigeasy / cadence / 784

92%

push

travis-ci

Bring `try/catch` into `invoke`.

Left it in its own little function because it seemed to be a wee bit
faster there still. `try/catch` used to prevent JIT compilation which is
why it was in its own function to begin with, it would otherwise cause
`invoke` to be uncompiled and terrible slow.

 % node benchmark/increment/call.js
_cadence call 1 x 5,225,931 ops/sec ±2.80% (90 runs sampled)
 cadence call 1 x 5,160,580 ops/sec ±2.94% (90 runs sampled)
_cadence call 2 x 5,210,175 ops/sec ±2.59% (92 runs sampled)
 cadence call 2 x 5,270,328 ops/sec ±0.68% (94 runs sampled)
_cadence call 3 x 5,288,174 ops/sec ±0.64% (94 runs sampled)
 cadence call 3 x 5,153,411 ops/sec ±2.18% (93 runs sampled)
_cadence call 4 x 5,256,624 ops/sec ±0.74% (91 runs sampled)
 cadence call 4 x 5,214,441 ops/sec ±1.99% (94 runs sampled)
Fastest is _cadence call 3, cadence call 2,_cadence call 4, cadence call 4,_cadence call 2, cadence call 1
 % node benchmark/increment/call.js
_cadence call 1 x 5,176,786 ops/sec ±3.14% (92 runs sampled)
 cadence call 1 x 5,178,754 ops/sec ±2.93% (91 runs sampled)
_cadence call 2 x 5,231,882 ops/sec ±2.60% (94 runs sampled)
 cadence call 2 x 5,337,209 ops/sec ±0.80% (94 runs sampled)
_cadence call 3 x 5,288,183 ops/sec ±0.60% (93 runs sampled)
 cadence call 3 x 5,220,239 ops/sec ±2.19% (90 runs sampled)
_cadence call 4 x 5,258,038 ops/sec ±1.09% (95 runs sampled)
 cadence call 4 x 5,211,841 ops/sec ±2.23% (94 runs sampled)
Fastest is  cadence call 2,_cadence call 2

68 of 68 branches covered (100.0%)

196 of 196 relevant lines covered (100.0%)

203.68 hits per line