Closed bakura10 closed 9 years ago
I'm not exactly sure how you want to have the before/after part measured. I also don't think that the Travis build machines always deliver the same performance.
I'm not particularly thinking about Travis, but your own modification. If I want to test an optimization, it's really useful to have some pre-made benchmarks for that.
Regarding ZF, it would allow mergers to locally test the performance impact of a change before merging. I think generalizing it (at least for critical components) would be really nice addition to ZF.
Ah, I see. Alright, I'll integrate that in a future PR.
Addressed with e7917933e8175110ceb8bd32ccf73f7b466063b3 – Thanks for the suggestion :)
Hi,
It would be awesome if you could move some of your benchmark code directly into Dash. As a rule of thumb, I'd like if we could have this in ZF3 for all critical performance components. That would allow for much simpler optimization work (as we would have a true before/after), as well as being able to quickly have a look at the impact of a new feature on performance, and eventually rejecting it (I'm thinking here about input filter that has many features introduced, without taking care of performance, so now this component is really slow - if we had that we could have quickly rejected some feature, or refactoring them).