Add Response.json static method

[lucacasonato]

This commit adds a Response.json static method that can be used to
create well formed JSON responses will very little effort.

The JSON response is not pretty printed.

Closes #1389

---

• At least two implementers are interested (and none opposed):
  • Chrome
  • Firefox
  • Deno
  • Cloudflare Workers
• Tests are written and can be reviewed and commented upon at:
  • Fetch: add tests for Response.json web-platform-tests/wpt#32825
• Implementation bugs are filed:
  • Chrome: https://crbug.com/1305358
  • Firefox: https://bugzilla.mozilla.org/show_bug.cgi?id=1758943
  • Safari: https://bugs.webkit.org/show_bug.cgi?id=240375
  • Deno: Implement Response.json()  denoland/deno#13902

---

Preview | Diff

[domenic]

Note that we'd need to fix Web IDL first to allow regular and static operations with the same name:

The identifier of a static operation also must not be the same as the identifier of a regular operation defined on the same interface.

from https://webidl.spec.whatwg.org/#idl-operations

[lucacasonato]

@domenic Do you think that is desirable to change? It may be confusing to have both a prototype and static method with the same name. Maybe this method should be called fromJSON.

[domenic]

I think it is desirable to change and do not see anything confusing about having such methods, personally.

[annevk]

This looks reasonable, but I think it can be improved by making initialize a response also take a body and Content-Type header value (potentially as a tuple?). That would allow for sharing even more steps.

I also wonder what we should do here if browser interest isn't forthcoming. Perhaps there's a way in which we can still have mutual agreement, but it's not exposed in browsers for now. Perhaps this is worth bringing up in the HTML triage call (if it doesn't otherwise get resolved).

[jasnell]

Just fyi... I have an implementation of this in Cloudflare Workers just waiting on this to land or at least signal that it's moving forward.

[domenic]

I don't think we have any signal that this is moving forward in browsers, unfortunately, due to it just being low-priority sugar for browser teams. As such I don't think it'll land in Fetch, at least until a couple of browsers have engineers looking for starter projects.

I'd really encourage you to use this as a case study for https://github.com/whatwg/js-hosts . Figure out what sort of documentation you want to create there, which would enable you to land such things with confidence, and go for it! I'm happy to participate and secure guarantees like "Chromium will never ship a Response.json() with different semantics than those specified here".

[lucacasonato]

@domenic Would this get a 👍 from Chromium if we did the implementation work?

[domenic]

@ricea would be the decider there but I can predict with very high probability he would support it :).

[annevk]

Mostly nits and request for a follow-up. Shared algorithm looks rather good! Thanks.

[foolip]

This change isn't possible to roll into WPT's idlharness.js test, see web-platform-tests/wpt#34155. Does anyone want to give it a go to update idlharness.js to avoid this problem?