Project roadmap

[ljharb]

Now that I've got the tests running in Github Actions, here's my rough plan for the package, not necessarily in order:

• the zuul tests only work on local branches, because secrets aren't available on PRs from forks, unless a Github Action workflow that uses pull_request_target is in place.
  • this can be resolved if @defunctzombie adds the current credentials to the repo's secrets, and then I can migrate the travis run to Actions
  • alternatively, we can remove these tests
• add "native" tests, for each node version's url implementation, and figure out which node version(s) we currently accurately target
• let's say url changed in node 0.10, v4, v6, and v10: this module should have v0.10.x, v4.x, v6.x, and v10.x versions published, that match those node implementations. So, the task is to publish those (this includes Update to newer version #24, update to node 7.2.1 #30, Support searchParams #36, [bug] URL is not a constructor #37, url.format ignores the "slashes" prop #48, etc)
  • an alternative would be, include all of those implementations in a v1.0.0 under different entry points, and ask users to configure their bundler to point to the one they want to target
• update the readme to make it clear that no features will be added to this package that node itself does not support

After this is all done, we should be able to easily remain "in sync" with node, for the duration.

[flaviotordini]

Hello, maybe I'm asking too much but is there a timeframe for whatwg-url support? In a project I'm using webpack with some nodejs modules to target a platform that has no URL implementation and this is going to be a showstopper.

[ljharb]

No, there’s no timeframe at this time.

There are separate whatwg URL polyfills you can certainly use in the meantime.

[rifler]

Sorry for silly question, what is the purpose of this project?
In node.js 16 if I do require('url') it uses core module

[ljharb]

@rifler older nodes exist, and browsers exist.