Using backspace, enter or slash causes a comment field in Facebook to jump instantly to the bottom of the screen.
Categories
(Core :: Panning and Zooming, defect, P2)
Tracking
()
People
(Reporter: akayanni, Unassigned)
References
Details
Attachments
(3 files)
User Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:102.0) Gecko/20100101 Firefox/102.0
Steps to reproduce:
Type in a "comment" field in a Facebook page.
Actual results:
When you enter return or backspace (possibility other keys) the comment field jumps to the bottom of the screen.
Expected results:
Page should not scroll.
Checked in safe mode.
Checked with "suggested text" on and off
Checked in Chrome and the error wasn't present.
It's been going on for some months. Likely been reported in a more technical manner. It's extremely annoying, surely has been reported?
Comment 1•4 years ago
|
||
The Bugbug bot thinks this bug should belong to the 'Core::Panning and Zooming' component, and is moving the bug to that component. Please correct in case you think the bot is wrong.
Comment 2•4 years ago
|
||
In another bug one user reported disabling a Facebook addon related to old or new layout of Facebook worked for them, but you said you tried to safe mode that should have disabled addons. We also haven't gotten a regression range from someone seeing this using https://mozilla.github.io/mozregression/ which pin point which change in Firefox (if any) caused this.
Comment 3•4 years ago
•
|
||
(In reply to Timothy Nikkel (:tnikkel) from comment #2)
We also haven't gotten a regression range from someone seeing this using https://mozilla.github.io/mozregression/ which pin point which change in Firefox (if any) caused this.
To emphasize: if this bug is affecting you (in a clean profile / with no addons), and you've got 20-30 minutes to spare, please consider using the mozregression tool to find the change that introduced the regression. We've had multiple reports of this, but have been unable to reproduce it ourselves, and without a regression window from someone who can reproduce it, we are unlikely to be able to reach a diagnosis and fix.
If you have questions about how to set up or use the mozregression tool, we're happy to provide details.
I've gone back as far as 2021-01-15
Far as I can remember this bug is about 4-7 months and not that far back. I can't imagine I've been putting up with it for over a year without writing a report but that is possible.
I'd say it was connected to JavaScript in Facebook somehow. Something they changed causes this bug to manifest.
Comment 5•4 years ago
|
||
I have been able to reproduce with the following keystrokes in a comment field:
- forward slash ==> Jumps
- space ==> Jumps
- backspace ===>no response
- space===>jump. Scroll page up, backspace ===>jump
I'm on Win10, Running 101.0b5
Comment 6•4 years ago
|
||
there are screenshot of similar issue in bug 1755158 comment #11 and bug 1755158 comment #12 .
Does your case with safe mode exactly match to them?
If not, can you post a screenshot before and after the issue happens?
Yes, that is the issue.
Goes with the spell check issue I reported months ago. You may or may not see the red line.
What happens on Facebook is critical to user experience. These 2 issues combined are making FF are real poor choice, at least for anyone using Facebook, which is 1/2 the known universe (unfortunately).
Comment 8•4 years ago
|
||
I found a reddit thread which looks the exact same issue but on Edge. https://www.reddit.com/r/edge/comments/syb76s/backspace_key_in_facebook_on_edge_makes_my_window/
Comment 9•4 years ago
|
||
I can reproduce in Edge with slash and backspace as Hiro pointed out. This isn't just a FF issue....
Comment 10•4 years ago
|
||
Possible cases that I can think of:
(1) The browser window size
Given some users say the issue disappears by setting to 90% zoom or smaller, the issue might be dependent on the window size.
So, in case the issue disappears by zooming out, it would be nice to check the window size, and also test in larger window size if possible.
The scroll behavior gets affected by what content is visible in the current scroll position.
Also, if the website applies responsive design, the issue might be specific to certain window width.
(2) User Agent or some browser-detection
If this happens mostly on non-Chromium browser, the website might be behaving differently for each browser,
for example, returning different HTML/CSS/JS depending on User Agent of the request,
or checking the User Agent in the JS and doing different thing.
So, testing with different User Agent string might give us some hint.
Comment 11•4 years ago
|
||
FWIW, for my setup, it seems to stabilize at 80% zoom for both FF and Edge...
Comment 12•4 years ago
|
||
Can you test if the issue happens again with 80% zoom + reducing the window width?
Comment 13•4 years ago
|
||
The stability/instability of the action appears to be independent of window width on my setup for both FF and Edge.
Stable action at 80% Zoom
Reducing window width at 80% - still stable
Unstable action at 90, 100% Zoom
Reducing window width at 90 & 100% - still unstable
Comment 14•4 years ago
|
||
Thank you for testing!
sorry, can you also check with reduced height?
I just realized that the height also affects the scroll behavior.
Comment 15•4 years ago
|
||
No change in the previously listed behaviors for either an window height change or a combined window size change (both height and width simultaneously) in either FF or Edge.
Comment 16•4 years ago
|
||
Thank you for the investigation!
It would mean that there are more complex condition, instead of just what's visible in the content area.
I'll look into other possibilities.
| Reporter | ||
Comment 17•4 years ago
|
||
(In reply to Frank Griffith Jr from comment #13)
The stability/instability of the action appears to be independent of window width on my setup for both FF and Edge.
Stable action at 80% Zoom
Reducing window width at 80% - still stableUnstable action at 90, 100% Zoom
Reducing window width at 90 & 100% - still unstable
<<<
I can reproduce all of that.
Interesting, I can't get the spell checker to fail on Facebook at 80% or below.
Comment 18•4 years ago
|
||
The severity field is not set for this bug.
:botond, could you have a look please?
For more information, please visit auto_nag documentation.
| Reporter | ||
Comment 19•4 years ago
|
||
For anyone who can set this the bug is S2
S2 (Serious) Major functionality/product severely impaired or a high impact issue and a satisfactory workaround does not exist
It's high impact because it applies to Facebook the most used site in the world and there is no workaround just a PITA.
If this bug goes beyond Nightly then there is a serious issue that will cause people to drop Firefox. It's so annoying that I'm ready to jump to Chrome until it is resolved.
Comment 20•4 years ago
|
||
Findings so far (e.g. comment 8) suggest that this is an issue on the Facebook side, rather than a Firefox issue.
That said, I will still mark this as S2, as the bug is fairly disruptive and we should do what we can to provide additional information to Facebook (e.g. stack traces of their Javascript code causing the undesired scrolling) to help them resolve it.
Comment 21•4 years ago
|
||
Facebook team located one case where such unexpected scroll jump happens with their text editor framework.
it sounds like, it was focusing on an element in the middle of reconcilation, that could move the scroll position in unexpected way.
https://github.com/facebook/lexical/pull/2262
It would be nice to test once the fix gets deployed.
Comment 22•4 years ago
|
||
We deployed that particular change on Wednesday. So if folks can test again and let us know if it has fixed their issue, that would be awesome.
Comment 23•4 years ago
|
||
No change to previously noted behavior on my end. Still jumps at 100% Zoom with front slash, space and backspace.
I'm on Win10, Running FF 101.0
Comment 24•4 years ago
|
||
Would you be able to post a video somewhere so I can see it happening (if possible, I understand if not)?
| Reporter | ||
Comment 25•4 years ago
|
||
Video of the issue, it's rough but shows the backspace issue, the spell check that stops working and how at 80% all is fine.
Comment 26•4 years ago
|
||
Thanks for the video, it helped us isolate an issue in Lexical. https://github.com/facebook/lexical/pull/2350
We deployed the fix yesterday to Facebook. If you could check again and let us know if you still experience the same problem, that would be much appreciated!
Comment 27•4 years ago
•
|
||
I have been able to confirm this issue appears fixed with the following keystrokes in a comment field:
forward slash ==> NO LONGER Jumps
space ==> NO LONGER Jumps
backspace ===>no response (as before)
space===>jump. Scroll page up, backspace ===> NO LONGER Jumps
I'm on Win10, Running 102.0b4 now. Will allow others to confirm but appears good on my end.
PS....no longer evident in Edge either.....
| Reporter | ||
Comment 28•4 years ago
|
||
Bug in spell check is lessened but not 100% resolved.
Comment 29•4 years ago
|
||
I'm unable to reproduce the spell check issue locally. Could it be a browser extension you have enabled? Does it happen on the Lexical playground for you? https://playground.lexical.dev/
Maybe that issue should be flagged in another bugzilla issue, as it's not related to the scroll problem.
| Reporter | ||
Comment 30•4 years ago
|
||
Type misspelled words
Backspace
Type more
Spell check is erratic.
Comment 31•4 years ago
|
||
Yani, can you file another bug for the spell check issue?
It's not a good idea to track multiple issues in a single bug.
Comment 32•4 years ago
|
||
(In reply to Dominic Gannaway from comment #26)
Thanks for the video, it helped us isolate an issue in Lexical. https://github.com/facebook/lexical/pull/2350
Dominic, thanks very much for tracking this down and fixing this!
Based on comment 27, I'm going to close this as having been fixed on the site side.
I did notice in the github issue you wrote:
When you zoom in, and do getBoundingClientRect, you get back decimal numbers, and strangely, doing it of two elements that are the same size, can return slightly different values.
If you're able to share a testcase that demonstrates this, I think that would be valuable for further investigation on our side. (Please feel free to file a new bug in the Core :: Layout component for it.)
Updated•4 months ago
|
Description
•