Specify window.webkitURL as an alias to URL

[zcorpan]

Fixes #135.

---

Preview | Diff

[annevk]

I'd kinda like to file a Firefox bug first to give them some time to object, but I personally agree that this is the way to go.

This might also be something that Node.js doesn't want to copy. I think we can leave that up to them to decide though.

[zcorpan]

Filed https://bugzilla.mozilla.org/show_bug.cgi?id=1366734

[zcorpan]

Bug for Edge
https://developer.microsoft.com/en-us/microsoft-edge/platform/issues/12091311/

[domenic]

This might also be something that Node.js doesn't want to copy. I think we can leave that up to them to decide though.

I guess technically Web IDL mandates [LegacyWindowAlias] to apply to the "primary global interface". Maybe we could tweak it to apply to Window only somehow.

But this is minor as Node doesn't implement the global-ness of URL anyway (yet?).

[annevk]

@domenic aaah right, that'll be fixed with whatwg/webidl#365 so I think we're totally in the clear then. (Except that at some point Node might want Exposed entries for its global...)

[zcorpan]

Relevant issue for chromium https://bugs.chromium.org/p/chromium/issues/detail?id=348985

[achristensen07]

Is there a way that chrome stats could measure the uses of webkitURL without URL as a fallback?

[zcorpan]

I don't know how that could reasonably be implemented, since JS execution doesn't reach the fallback, there's no way of knowing what the fallback is.

But if there's still interest to remove it in WebKit, Chromium, and EdgeHTML, it seems plausible that it would be successful since Gecko doesn't have it and has no bugs reported.

Edit: quick compat analysis at #135 (comment)