Data for DOMRectList seems to be bogus

[bzbarsky]

DOMRectList.item and DOMRectList.length are claimed to be Firefox-only APIs. But a simple testcase shows they exist at least in Chrome and Safari:

<script>
  alert(document.documentElement.getClientRects().item + "\n" +
        document.documentElement.getClientRects().length);
</script>

[foolip]

Blink has [ NoInterfaceObject ] interface DOMRectList, which matches https://drafts.fxtf.org/geometry/#DOMRectList

I'm pretty sure this changed in the not too distant past.

[bzbarsky]

Ah. Yeah, and imo that spec change was bogus and it should be reverted. See https://github.com/w3c/fxtf-drafts/issues/233

But the bigger point is that confluence is getting the information wrong, because it claims the APIs don't exist in browsers where they very much do exist. That is a confluence bug.

[mdittmer]

Do we already have a more general bug about using instance factories to discover APIs? The current procedure for Confluence should consistently exclude any NoInterfaceObject API that is not accessible via instances hanging off the global object. The real bug is that we haven't implemented a slew of factories to produce instances of such APIs for inspection.

On Thu, Jan 11, 2018 at 8:35 AM, Boris Zbarsky [email protected] wrote:

Ah. Yeah, and imo that spec change was bogus and it should be reverted. See w3c/fxtf-drafts#233 https://github.com/w3c/fxtf-drafts/issues/233

But the bigger point is that confluence is getting the information wrong, because it claims the APIs don't exist in browsers where they very much do exist. That is a confluence bug.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/GoogleChromeLabs/confluence/issues/236#issuecomment-356935575, or mute the thread https://github.com/notifications/unsubscribe-auth/ABsWSJ4m0Cck4h5zS7i1ZKiBN7zJYpMxks5tJg41gaJpZM4QU3Ec .