On: WebKit on GitHub

[usefulcat]

> no patch can land if it regresses benchmarked performance

..unless it fixes some important security vulnerability, one hopes..

[btown → usefulcat]

Via https://webkit.org/performance/ :

> If a patch lands that regresses performance according to our benchmarks, then the person responsible must either back the patch out of the tree or drop everything immediately and fix the regression.

I imagine that a security hotfix would lead almost immediately to the second situation (perhaps as soon as the implementor had gotten some sleep!)

[mhh__ → usefulcat]

the trick is to change the benchmark at the same time