1026720 - Consider making hasInterfacePrototypeObject always true for non-callback interfaces

Status: NEW

(Core :: DOM: Core & HTML, defect)

Description

[:Ms2ger (he/him; ⌚ UTC+1/+2)]

We added an exception for interfaces without concrete descendants, because the URL.createObjectURL() spec required it (as we didn't support URL objects at the time). I don't know if we have any interfaces that hit this case at this point. We should probably change either our implementation or the spec.

Comment 1

[:Ms2ger (he/him; ⌚ UTC+1/+2)]

This affects:

AbstractWorker
ChromeUtils
ChromeWindow
CSS
KeyEvent
PromiseDebugging
SVGUnitTypes
SVGZoomAndPan
ThreadSafeChromeUtils
WebrtcGlobalInformation

According to gsnedders, <http://software.hixie.ch/utilities/js/live-dom-viewer/saved/3697> shows Edge matching us for CSS.

Comment 2

[Boris Zbarsky [:bzbarsky]]

AbstractWorker is [NoInterfaceObject]; it's a mixin, not an interface.  So not relevant here.

ChromeUtils, PromiseDebugging, ThreadSafeChromeUtils, WebrtcGlobalInformation are [ChromeOnly], so exact behavior doesn't really matter.

ChromeWindow is [Func="IsChromeOrXBL"] so exact behavior doesn't really matter.

KeyEvent is a mixin, not an interface.  It only has an interface object because of the constants.  I'm not even sure it's standard; e.g. I don't see it in Chrome.

SVGUnitTypes, SVGZoomAndPan are mixins but have constants.  I have no idea what the status of the SVG spec is here.

So in practice the only standardized interfaces affected are CSS and the two SVG ones.  It's kinda silly to give them interface objects, but we could do that, I guess.