On my project, this makes ESLint as a whole roughly 8% faster, and it seems to result in consistently faster benchmarks.
Note a behavior change - member expressions used to find the first matching rule, while they now favor the more general rule if relevant. (I believe this change is an improvement: "Promise is not supported in IE 10" is a more helpful message than "Promise.reject() is not supported in IE 10".)
@amilajack, could you rerun the workflow for this branch? I believe the earlier test failures were due to issues that had already been fixed in the main branch.
On my project, this makes ESLint as a whole roughly 8% faster, and it seems to result in consistently faster benchmarks.
Note a behavior change - member expressions used to find the first matching rule, while they now favor the more general rule if relevant. (I believe this change is an improvement: "Promise is not supported in IE 10" is a more helpful message than "Promise.reject() is not supported in IE 10".)