Audit usage of nsLayoutUtils::IsAncestorFrameCrossDoc
Categories
(Core :: Layout, enhancement, P3)
Tracking
()
Fission Milestone | M7 |
People
(Reporter: emilio, Assigned: jwatt)
References
(Blocks 1 open bug)
Details
Replacing for IsProperAncestorFrame if needed.
Updated•6 years ago
|
Updated•6 years ago
|
Assignee | ||
Comment 1•6 years ago
|
||
I didn't see any consumers of this function that we need to worry about. Things like event targeting and display list pruning for pages are not something we need to worry about across iframe process boundaries.
Assignee | ||
Comment 2•6 years ago
|
||
Oops, I misread the function name in the title.
Assignee | ||
Comment 3•6 years ago
•
|
||
I cleaned up my own notes that I had on the consumers of this method:
nsDisplayPerspective::CreateWebRenderCommands:
mattwoodrow says this should not be using the cross-document variant
of this function. I think we could fix this as part of bug 1523500
(I mentioned it over there in bug 1523500 comment 5).
nsIPresShell::ClearMouseCapture:
Covered by bug 1533499.
nsRootPresContext::ComputePluginGeometryUpdates:
Covered by bug 1525720?
GetClosest in PositionedEventTarget.cpp:
Henri/Botond are working on making sure that events go to the correct process
using the APZ code, so this reaching across document boundaries should be
okay at that point.
DOMIntersectionObserver::Update:
This is fine. Script can't observe elements that it can't reach (i.e. anything
cross-document would need to belong to the same origin etc.). The exception to
that is when script passes 'null' to observe intersection with the root
viewport.
nsLayoutUtils::IsProperAncestorFrameCrossDoc
No issues - see comment 1.
Comment 4•4 years ago
|
||
jwatt, can you re-evaluate this for Fission and triage into appropriate Fission milestone?
M7: launch Beta experiment
M7a: Add Linux and BFCache to Beta experiment, increase beta experiment %age, automation test coverage for Fission
M8: Launch Release experiment
MVP: Fission rides the trains
Assignee | ||
Comment 5•4 years ago
|
||
nsDisplayPerspective::CreateWebRenderCommands seems to be the remaining case and (as mentioned in comment 0) we can use bug 1523500 for that. So it looks like we're now done here.
Description
•