I just took a really quick glance at the code. These changes might significantly slow down performance, but if you wrote a quick benchmark it should tell us. Would you be willing to add a quick benchmark to this pull request?

Ya, sure thing!

Because the emit method returns as soon as propagation is stopped, execution could actually end up being considerably faster when an early listener stops propagation; on the other hand, it does add an extra comparison to each successful if all listeners are processed as before.

To me, it sounds like the current benchmark scripts should be able to cover these changes, unless I’m missing something. I’d be happy to add another benchmark if you can point me in the right direction.

Here is a comparison of the current benchmarks run from both branches:

stopPropagation

EventEmitterHeatUp x 1,416,737 ops/sec ±0.71% (100 runs sampled)
EventEmitter x 1,532,806 ops/sec ±0.46% (98 runs sampled)
EventEmitter2 x 9,411,991 ops/sec ±0.56% (98 runs sampled)
EventEmitter2 (wild) x 6,423,381 ops/sec ±9.22% (93 runs sampled)

EventEmitterHeatUp x 1,294,572 ops/sec ±0.45% (100 runs sampled)
EventEmitter x 1,475,587 ops/sec ±0.63% (97 runs sampled)
EventEmitter2 x 8,797,374 ops/sec ±0.47% (98 runs sampled)
EventEmitter2 (wild) x 6,255,054 ops/sec ±9.13% (95 runs sampled)

EventEmitterHeatUp x 1,430,530 ops/sec ±0.56% (99 runs sampled)
EventEmitter x 1,397,150 ops/sec ±0.52% (97 runs sampled)
EventEmitter2 x 8,770,814 ops/sec ±0.45% (99 runs sampled)
EventEmitter2 (wild) x 4,457,747 ops/sec ±5.88% (96 runs sampled)

EventEmitterHeatUp x 1,420,539 ops/sec ±0.71% (98 runs sampled)
EventEmitter x 1,502,386 ops/sec ±0.70% (98 runs sampled)
EventEmitter2 x 9,216,787 ops/sec ±0.40% (96 runs sampled)
EventEmitter2 (wild) x 6,311,002 ops/sec ±9.28% (98 runs sampled)

master

EventEmitterHeatUp x 1,416,757 ops/sec ±0.76% (96 runs sampled)
EventEmitter x 1,538,006 ops/sec ±0.42% (99 runs sampled)
EventEmitter2 x 9,082,098 ops/sec ±0.52% (99 runs sampled)
EventEmitter2 (wild) x 6,159,633 ops/sec ±9.24% (96 runs sampled)

EventEmitterHeatUp x 1,382,359 ops/sec ±0.31% (99 runs sampled)
EventEmitter x 1,355,995 ops/sec ±0.58% (101 runs sampled)
EventEmitter2 x 9,289,798 ops/sec ±0.42% (99 runs sampled)
EventEmitter2 (wild) x 6,560,096 ops/sec ±9.21% (92 runs sampled)

EventEmitterHeatUp x 1,407,961 ops/sec ±1.34% (99 runs sampled)
EventEmitter x 1,525,306 ops/sec ±0.49% (102 runs sampled)
EventEmitter2 x 9,093,574 ops/sec ±0.48% (94 runs sampled)
EventEmitter2 (wild) x 6,156,239 ops/sec ±9.19% (92 runs sampled)

EventEmitterHeatUp x 1,273,150 ops/sec ±6.06% (87 runs sampled)
EventEmitter x 1,479,227 ops/sec ±0.85% (100 runs sampled)
EventEmitter2 x 8,806,507 ops/sec ±0.70% (94 runs sampled)
EventEmitter2 (wild) x 5,237,295 ops/sec ±8.16% (91 runs sampled)

@0x00A Do the bechmarks look good enough to you?

This would be a nice addition.

@elmatou If you're up for fixing the conflicts and testing that it works, I'm down for getting it in. :)