[Webkit-unassigned] [Bug 236952] Permission API: handle non-fully active documents

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=236952

Marcos Caceres <marcos at marcosc.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                URL|                            |https://github.com/w3c/perm
                   |                            |issions/pull/365

--- Comment #4 from Marcos Caceres <marcos at marcosc.com> ---
(In reply to youenn fablet from comment #3)
> Comment on attachment 452717 [details]
> Patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=452717&action=review
> 
> > Source/WebCore/Modules/permissions/Permissions.cpp:88
> > +    if (is<Document>(context) && !downcast<Document>(*context).isFullyActive()) {
> 
> https://w3c.github.io/permissions/#query-method is not clear about this
> check, is there a corresponding PR that should be merged?

Sorry I should have pasted the link:
https://github.com/w3c/permissions/pull/365

> Are other browsers already implementing this check?

See links to other implementations. 

> > Source/WebCore/Modules/permissions/Permissions.cpp:94
> >          promise.reject(Exception { InvalidStateError, "The context is invalid"_s });
> 
> Maybe we should do this check before the fully active one.
> 
> > LayoutTests/imported/w3c/ChangeLog:10
> > +        * web-platform-tests/permissions/resources/empty.html: Added.
> 
> Are these new WPT tests that need to be upstreamed or are they already
> merged?

PR at https://github.com/web-platform-tests/wpt/pull/32921

> > LayoutTests/imported/w3c/web-platform-tests/permissions/non-fully-active.https-expected.txt:6
> > +TIMEOUT Permission change events shouldn't fire on non-fully active document Test timed out
> 
> We usually do not like tests to timeout. Can the test be updated to check
> for onchange and fail early if not present?

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20220221/42060fed/attachment.htm>