Do not let elements inside display:none subtrees become focus targets.
The HTML focusable-area model only allows elements to be focusable when
they are rendered, delegate rendering to their children, or are relevant
canvas fallback content.
Preserve blur for the current focused area after script hides it, since
the unfocusing steps operate on the old focus target. Check display:none
through the flat-tree style parent chain, so slotted controls inside
hidden slot subtrees cannot become focused.
Cover hidden ancestors with materialized computed style, display:none
controls, display:contents, hidden focused controls, and slotted cases
inside hidden and visible shadow-tree subtrees.
I missed when these changes actually happened, but the big differences
here are that a few steps got extracted into their own algorithms; and
the dialog now only does things when it's attached to a fully-active
document, instead of immediately.
I've imported a WPT test that exercises this by calling show() on an
unattached dialog.
We previously had this implemented in an ad-hoc way, where we used a
boolean on the CloseWatcher instead of a proper function with steps.
This worked at the time, but causes problems with the current version
of the spec, so let's just implement it properly.
This commit consciously does not update the spec text, because it's
diverted quite a lot. That will happen in a subsequent commit.
Update a couple of focus-related spec steps and their implementations.
The most relevant change is that we no longer allow focusing on elements
that return false for `->is_focusable()`, which necessitates fixing a
broken test that tried to `.focus()` on `<div>`s that were not
focusable. That test's output now more accurately reflects the expected
outcome as seen in other browsers.
This fixes Layout/input/dialog-open-modal.html which began flaking
super hard after the preceding commits that reduced style invalidation
for focus-related pseudo class selectors.
Resulting in a massive rename across almost everywhere! Alongside the
namespace change, we now have the following names:
* JS::NonnullGCPtr -> GC::Ref
* JS::GCPtr -> GC::Ptr
* JS::HeapFunction -> GC::Function
* JS::CellImpl -> GC::Cell
* JS::Handle -> GC::Root
Now that we have RTTI in userspace, we can do away with all this manual
hackery and use dynamic_cast.
We keep the is<T> and downcast<T> helpers since they still provide good
readability improvements. Note that unlike dynamic_cast<T>, downcast<T>
does not fail in a recoverable way, but will assert if the object being
casted is not a T.
