Allow same identifier in static and regular ops

[lucacasonato]

Closes #1097

• At least two implementers are interested (and none opposed):
  • Deno
  • Cloudflare Workers
  • Chromium
  • Mozilla
• Tests are written and can be reviewed and commented upon at:
  N/A
• Implementation bugs are filed:
  • Chrome: https://bugs.chromium.org/p/chromium/issues/detail?id=1305399
  • Firefox: https://bugzilla.mozilla.org/show_bug.cgi?id=1765992
  • Safari: https://bugs.webkit.org/show_bug.cgi?id=239672
  • Deno: already supported
  • Node.js: Support regular and static operations with the same identifier for WebIDL nodejs/node#42831

---

Preview | Diff

[Okiamgood]

[spam]

[lucacasonato]

@TimothyGu / @EdgarChen would you mind taking a look? This is a prerequisite for whatwg/fetch#1392.

[EdgarChen]

This looks good to me.
I checked how we handle the overloaded operations, there are separated overload set for regular and static operation, so I think it is fine to have same identifier in regular and static operations.

[EdgarChen]

@TimothyGu, WDYT?

[EdgarChen]

BTW, I filed https://bugzilla.mozilla.org/show_bug.cgi?id=1765992 for Gecko.

[lucacasonato]

Thanks @EdgarChen. Added the link to the PR description. I'll open a WebKit issue in a moment.

Can I count Gecko and Chromium as interested implementers for the working mode requirements?

[jasnell]

+1 to this. Definitely useful

[lucacasonato]

cc @tabatkins, as this may affect bikeshed IDL handling too.

[tabatkins]

Yeah, thanks for the ping. I will indeed need to make a change in how Bikeshed handles these. (The change makes sense tho, so Bikeshed shouldn't be a blocker here.)

[lucacasonato]

Support for this was merged in Chromium (https://chromium-review.googlesource.com/c/chromium/src/+/3623640), so we have two "interested implementers" now. @EdgarChen think we can land this?

[foolip]

I think this will impact MDN documentation, because static and instance members aren’t distinguished in the URL convention. Does anyone have a plan for that?

[domenic]

From what I understand MDN has had some consensus to move away from their current pretend-all-methods-are-static style for some time, but I haven't seen any actual updates: https://github.com/mdn/content/discussions/5121 and some other discussions.

[foolip]

It's been discussed, but there weren't any plans for anyone to work on it. I've commented on https://github.com/mdn/content/discussions/5121 and filed openwebdocs/project#104 to put it on the radar, but I would expect Response.json to not be documented (or in BCD) until that's resolved.