Not able to scroll page "ps-blog-game-of-the-year-2021-the-winners" (Lady Dimitrescu keeps reappearing in scrollport)
Categories
(Core :: Layout: Flexbox, defect)
Tracking
()
Webcompat Priority | P2 |
People
(Reporter: patrickj, Unassigned)
References
(Regression, )
Details
(Keywords: regression)
Attachments
(1 file)
Steps to reproduce:
-
Navigate to https://blog.playstation.com/2021/12/20/ps-blog-game-of-the-year-2021-the-winners/
-
Scroll down page using scrollbar
Expected result: Able to scroll page
Actual result: Page gets stuck in a loop and won't scroll past a certain section
Regression - last good build 2020-11-03, first bad build 2020-11-04 - unable to bisect further
Comment 1•4 years ago
|
||
Working on Nightly on macOS.
Comment 2•4 years ago
|
||
On Windows Nightly I can see the issue here. Scrolling down shows several times the same content and after scrolling down a while, scrolling up will show a blank canvas. Let's start the investigation from DOM: Events.
Comment 3•4 years ago
|
||
This is not a recent regression. To me it happens also with Fx90.
Comment 4•4 years ago
|
||
Possible regression range based on comment 0:
https://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=8e58925909cccd52a321809d5094d96e7979ed0e&tochange=2977d4891ceeed019dc906dff3feb6c9f530213c
At first glance, this stands out:
Bug 1671526, use DispatchInputEvent for drag events on windows, r=kats
https://hg.mozilla.org/mozilla-central/rev/b6d258866337047ef3aa51f629e026c9d5dd5cf4
Comment 5•4 years ago
|
||
Unlikely. That code will only get called during a drag/drop operation.
This looks more like a rendering issue, and I see similar glitchiness on Mac as well.
Comment 6•4 years ago
|
||
(In reply to Jens Stutte [:jstutte] from comment #2)
On Windows Nightly I can see the issue here. Scrolling down shows several times the same content and after scrolling down a while, scrolling up will show a blank canvas. Let's start the investigation from DOM: Events.
I could also reproduce the same glitchiness on Mac and I got following regression range,
4:13.91 INFO: Last good revision: fb2b08c6bc8c320127586fcedb2626e496bf3dcf (2020-07-07)
4:13.91 INFO: First bad revision: ee89deeb4fb6d12456c551c5ab4aaae151ceedae (2020-07-08)
4:13.91 INFO: Pushlog:
https://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=fb2b08c6bc8c320127586fcedb2626e496bf3dcf&tochange=ee89deeb4fb6d12456c551c5ab4aaae151ceedae
![]() |
||
Comment 7•4 years ago
|
||
Scrolling down shows several times the same content
I can reproduce the issue on Windows10.
Regression window:
https://hg.mozilla.org/integration/autoland/pushloghtml?fromchange=3a38af2b7e83456941bed532016ae31208589dd0&tochange=1aa53285261e17aa6a8b36e05c7208d616c24aa4
Comment 8•4 years ago
|
||
Bug 1492538 appears in both regression ranges. Daniel, could you take a look and in case confirm the regression?
Comment 9•3 years ago
|
||
(In reply to Jens Stutte [:jstutte] from comment #8)
Bug 1492538 appears in both regression ranges. Daniel, could you take a look and in case confirm the regression?
I confirmed it is a regression from that change, yeah.
Here's a diagnostic patch that lets you disable the optimization that we added in that bug. With this patch applied (and with the included pref layout.css.flex.perf.cache-final-reflow.disabled
manually set to true
), the bug does not reproduce.
My STR for testing this are as follows (simpler/more-reliable than using the scrollbar IMO):
(1) Load https://blog.playstation.com/2021/12/20/ps-blog-game-of-the-year-2021-the-winners/
(2) Press pagedown 3-4 times, until you see the picture of a Lady Dimitrescu | Resident Evil Village
(3) Press pagedown a couple more times.
ACTUAL RESULTS: Lady Dimitrescu's photo keeps on reappearing at or near the top of my scrollport. And if I scroll up (with e.g. uparrow), it looks like a bunch of blank space has been inserted between her and the previous text (Now, on to the winners!
).
Updated•3 years ago
|
Updated•3 years ago
|
Comment 10•3 years ago
|
||
Calling this S2 since it makes this article pretty unusable/unreadable, and it seems to be a bug in an optimization that was not meant to impact user-perceivable behavior.
Updated•3 years ago
|
Updated•3 years ago
|
Comment 11•3 years ago
|
||
Set release status flags based on info from the regressing bug 1492538
Updated•3 years ago
|
Comment 13•3 years ago
|
||
Daniel, do we have a potential fix for this bug? I am assuming that we probably won't have time to uplift a fix to 98, is that correct?
Comment 14•3 years ago
|
||
No fix or thorough diagnosis yet -- and correct, it's unlikely we'll have a fix for 98. (FWIW we've been shipping with this bug since Firefox 80; this seems to be a bug in an optimization that we shipped as part of that release. Hoping to fix this during the near future, though.)
Updated•3 years ago
|
Updated•3 years ago
|
![]() |
||
Updated•3 years ago
|
Comment 16•3 years ago
|
||
Setting WebCompat Priority P2 for now: we don't have evidence of many sites breaking because of this, but where it breaks, it makes content inaccessible.
Updated•3 years ago
|
Updated•3 years ago
|
Updated•3 years ago
|
Updated•3 years ago
|
Comment 17•3 years ago
|
||
This seems to work for me now.
Comment 18•3 years ago
|
||
Yeah, works for me as well, using STR from comment 9. It even works in Nightly 2022-01-10 (the datestamp from comment 9). So possibly there was a site change here. Too bad; that might mean we're unable to investigate further / know-if-it's-still-a-problem.
Closing as WORKSFORME, but please reopen if anyone can still reproduce.
Updated•3 years ago
|
Description
•