You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Based on goja, which gives us better optimization options than otto.
Remove the confusing Runtime/Runner distinction.
Split the monolithic API object, it should in itself deal only with core behavior such as groups and checks, everything else should be in subpackages and registered with a module registry.
You should be able to PR new modules with only a single-line modification to the js package, and everything else self-contained.
I'm doing some preliminary work on this right now, putting this out here for discussion.
The text was updated successfully, but these errors were encountered:
Once the reworked JS runner is in place, would one of the side effects be that goja would completely replace otto in k6, or would otto still be a dependency?
I'm asking from the point of view of planning the debian packaging - neither of them are on the repositories and would need to be packaged, and if otto is to be removed in the future (whenever that future is!) it would help me with prioritizing the tasks involved.
Yeah! Using goja is way faster for what we're using it for, so I want to move everything to it. ES6 compilation is already on goja (as otto's regex implementation couldn't run Babel), the remaining bits of otto code are the JS runner and thresholds.
I have a couple of design goals for this:
API
object, it should in itself deal only with core behavior such as groups and checks, everything else should be in subpackages and registered with a module registry.js
package, and everything else self-contained.I'm doing some preliminary work on this right now, putting this out here for discussion.
The text was updated successfully, but these errors were encountered: