Add doc is fully active check

[marcoscaceres]

Closes - no issue filed...

The following tasks have been completed:

• Modified Web platform tests -wakelock-active-document.https.window.js and Screen Wake Lock: Add test for non-fully active doc web-platform-tests/wpt#45599

Implementation commitment:

• Chromium - already implemented
• Gecko - already implemented
• WebKit - already implemented

---

Preview | Diff

[EdgarChen]

I wonder if the fully active should be checked again in the task queued in step 7.3. of https://www.w3.org/TR/screen-wake-lock/#the-request-method, as the active state might change, similar to visibility.

[marcoscaceres]

Yeah, good point.

[marcoscaceres]

Integrated @EdgarChen's feedback... now also check if fully active at the turn of the event loop.

[marcoscaceres]

Added an additional test for the update web-platform-tests/wpt#45599

[marcoscaceres]

@EdgarChen, does Gecko already implement the fully active checks or do you want me to file bugs for those?

Never mind... found the answer.

[marcoscaceres]

@reillyeon, this should be good to go, I think.

@rakuco, are you still editing this spec? Should I be pinging you for review? (didn't get any response from previous pings, so was unsure)

[rakuco]

As discussed in one of the inline comments, this check was already present in the spec, so I think the PR should be contain more information and indicate that this is moving the existing check and also checking if the document's fully active in the parallel steps.

Speaking of moving the existing check, you mentioned earlier that "the fully active check should be done first". Could you elaborate a bit? I remember asking ages ago if there was a recommended order for the permissions policy + permissions API checks in general, but thought it didn't matter if the fully active ones came before or after those.

[rakuco]

@rakuco, are you still editing this spec? Should I be pinging you for review? (didn't get any response from previous pings, so was unsure)

Sorry for the silence, it wasn't intentional! I haven't had much time to work on this spec lately; now that the DAS/WebApps situation seems to have progressed I think I can finally become just an interested contributor :-)

[marcoscaceres]

Speaking of moving the existing check, #367 (comment) that "the fully active check should be done first". Could you elaborate a bit? I remember asking ages ago if there was a recommended order for the permissions policy + permissions API checks in general, but thought it didn't matter if the fully active ones came before or after those.

Sure. Without a document (or the document being null or an iframe being detached) the permission check doesn't make sense - so, basically the fully active check just confirms the document is something in the web page and not just a dangling reference that didn't get garbage collected or accidentally stored in a variable.

[rakuco]

lgtm, and thanks for the clarification!

[rakuco]

As discussed in one of the inline comments, this check was already present in the spec, so I think the PR should be contain more information and indicate that this is moving the existing check and also checking if the document's fully active in the parallel steps.

It'd be good to still do this though -- the title does not fully cover what the change is doing.

[marcoscaceres]

Made sure the commit message better reflects the change.