Wpt: Spin off testharness?

[tripu]

ie, maintain testharness.js as a separate project and publish releases on npm.

(I think that's a minimum requirement if the authors want it to be adopted by others.)

cf https://github.com/w3c/nodejs/issues/4

[jgraham]

I'm not sure I understand the provenance of this issue. But I think that testharness.js should focus on being the best tool for writing web-platform-tests and not try to be a generic test harness for all scenarios; my experience is that the engineering tradeoffs are very different between a test runner designed for generic libraries and one that's designed specifically for testing the browser.

In terms of the repository, it used to be in a separate repo, but that turned out to have substantial overhead when making changes, so I'd be extremely reluctant to split it out again (this is linked to the goal of being the best thing for wpt tests specifically).

[tripu]

@jgraham: I actually agree with you. I was filing this issue “on behalf” of others who had expressed interest in bringing testharness.js to a wider audience, and who were defending its virtues as a generic JS testing tool (as opposed to just for WPT and the browser).

I'm waiting to see if somebody steps forward to defend the merits of this idea — otherwise, I'm happy to close this issue.

[plehegar]

I don't think we need a separate repo for testharness.js in order to create an npm package....

[tripu]

“I don't think we need a separate repo for testharness.js in order to create an npm package....”

Okay by me; whatever the maintainer(s) find most appropriate :)

[tripu]

Closing this, since there seems to be consensus about testharness not needing to live in a separate GH repository (but I still think it'll have to be published independently via npm if you guys want any other project to adopt it in the future).