1679320 - hit testing is wrong (or delayed) with an element on top of a cross origin iframe

Status: RESOLVED DUPLICATE of bug 1676466

(Core :: Layout, defect, P2)

Description

[Cameron McCormack (:heycam)]

STR:

1. Load https://www.macrumors.com/.
2. Scroll down to an entry that embeds a tweet where the embedding iframe is behind the "Show Full Article" button.
3. Mouse over the "Show Full Article" button.
4. Click the button.

Expected results:

1. When mousing over the button, it is immediately highlighted and the cursor changes to a hand.
2. When clicking the button, the full macrumors.com article is opened.

Actual results:

1. When mousing over the button, it is highlighted only after a ~0.5s delay and the cursor does not change.
2. When clicking the button, the tweet is opened.

Comment 1

[Hiroyuki Ikezoe (:hiro)]

Attachment: Screenshot_2020-11-26 MacRumors Apple Mac iPhone Rumors and News(1).png

The title of the entry in question is "Reliable Leaker Suggests Redesigned MacBooks in 2021 Will Include Both Apple Silicon and Intel Models". Oddly my daily use nightly does just show the tweet text and is not able to reproduce the issue. But on my local build, it shows embedded tweet image, then is able to reproduce the issue. I am attaching a screenshot of the entry. (I took it without fission, since with enabling fission, the tweet image were not rendered in the screenshot)

Comment 2

[Hiroyuki Ikezoe (:hiro)]

Looks like the text tweet case on my daily use nightly, the tweet content is not iframe-ed at all. I don't know where the difference comes from.

Comment 3

[Hiroyuki Ikezoe (:hiro)]

Attachment: A possible test case

Probably this is a minimum test case, there is an OOP iframe inside a containing block and there is a a position:absolute element inside the containing block and the element has an <a>.

Comment 4

[Hiroyuki Ikezoe (:hiro)]

CCing Timothy. (I have totally no idea how we treat this kind of overlapping elements in APZ)

Comment 5

[Neha Kochar [:neha]]

Timothy, this is really bad, can you fix this please?

Comment 6

[Timothy Nikkel (:tnikkel)]

Redirecting need info to Botond as I believe he is currently working on fission hit testing issues.

Comment 7

[Neha Kochar [:neha]]

I believe Miko is already looking into similar hit testing issues for Fission (thanks Miko).

Comment 8

[Neha Kochar [:neha]]

Botond said he'll look into this.

Comment 9

[Hiroyuki Ikezoe (:hiro)]

I am ~100% sure this is a dup of bug 1676466. (I hadn't noticed the bug before)

Comment 10

[Botond Ballo [:botond]]

Thanks Hiro! I'll mark bug 1676466 as a dependency, and we'll re-test once that is fixed.

Comment 11

[Neha Kochar [:neha]]

Miko, can you test this with your patch for bug 1676466 to see if it fixes this?

Comment 12

[Miko Mynttinen]

(In reply to Neha Kochar [:neha] from comment #11)

Miko, can you test this with your patch for bug 1676466 to see if it fixes this?

It does. :)