-
Notifications
You must be signed in to change notification settings - Fork 2
Description
I'm having a hell of a time tricking the combination of benchmark-regression and benchmark.js into doing complex things.
The lack of async support complicates things quite a bit, but I figured out a trick with onStart() that's not too unpalatable.
By far the biggest problem has been trying to write benchmarks for a library that uses node:cluster, which of course tries to run all of the tests in the child process because of the fork() mechanism, and then due to this project not propagating any of the metadata to the tests (unlike benchmark.js) there is no convenient way to propagate that data to the child.
If the goal of the project wasn't to support node:cluster, I'd recommend they replace the whole thing with workers, which would still have the problem of "where is the module I am supposed to be using?" but not the recursive run issue.
tl;dr: I'm motivated to make a few improvements to run() but this project hasn't seen a PR in years and that worries me.