849922 - Panning is janky on Facebook's mobile site, breaks and jumps back to panning start position

Status: VERIFIED FIXED

(Firefox for Android Graveyard :: Toolbar, defect)

Description

[Wesley Johnston (:wesj)]

I noticed that when I pan on touch.facebook.com, I often see the scroll position jump as new content is loaded. I'm not sure if Facebook is inserting content that changes our scroll position or not, but it would be nice to know at least (and fix if possible).

Comment 1

[Mathieu Pellerin]

I'm also seeing this. When panning down on Facebook's mobile site, the panning often breaks / stop abruptly and the view port jumps back to the panning's start position. 

I'm increasing the importance to major as this, left unresolved, would have a negative impact on a huge number of Firefox Android users visiting a popular website.

This regressed in the beginning of March 2013. A March 4 nightly build doesn't exhibit the symptoms. Adding a "regressionwindow-wanted" keyword.

Comment 2

[Aaron Train [:aaronmt]]

I don't see this at all on my Nexus 4 (Android 4.2.2), Nightly (03/20); which device are you using?

Comment 3

[Mathieu Pellerin]

Aaron, I'm seeing this both on my Samsung Galaxy Note 2 and my partner's Samsung Galaxy S3.

I have the feeling it has to do with the (exciting) commit of the feature that hides toolbar when panning down. It's certainly around this time that the regression started occurring.

Comment 4

[Mathieu Pellerin]

I should add that both devices run on the latest Samsung firmwares (Android 4.1.2)

Comment 5

[Mark Finkle (:mfinkle) (use needinfo?)]

I see this on a Nexus S running Gingerbread. The page will jump back in a large increment.

I also see it on a Galaxy Nexus running ICS. The jump back is much smaller though.

To reproduce this, sign in and wait for page to load. Then flick the page to pan down a large amount. Once the pan stops, wait a few seconds. The jump happens after the page stops panning.

Comment 6

[Wesley Johnston (:wesj)]

I also see it on my S3. The jumps are pretty large and they make panning pretty much impossible. I'll try and take a look sometime, but if kats or cwiiis have time earlier, they know this code better.

Comment 7

[Kartikaya Gupta (email:[e-mail])]

Might want to see if it still happens with the pref turned off.

Comment 8

[Kartikaya Gupta (email:[e-mail])]

(The browser.chrome.dynamictoolbar pref, that is)

Comment 9

[Chris Lord [:cwiiis]]

I can't reproduce this locally, so this may have been fixed already - Possibly by bug 852565 which I haven't pushed yet (waiting on try).

Comment 10

[Aaron Train [:aaronmt]]

(In reply to Mathieu Pellerin from comment #3)
> Aaron, I'm seeing this both on my Samsung Galaxy Note 2 and my partner's
> Samsung Galaxy S3.

I too have a Note II (Android 4.1.2) and don't see this at all; panning is rather snappy with no 'jank'.

Comment 11

[Chris Lord [:cwiiis]]

Going to close this, but reopen if you get solid STR on more than one device (even if they're the same model).

Comment 12

[Wesley Johnston (:wesj)]

I can see this pretty easily by just trying to pan quickly on Facebook. Flipping the pref off fixes it on the current nightly.

It seems like we do have pretty solid STR on more than one device (although I don't know why that matters). I'm going to reopen until this is fixed with the dynamic toolbar on. I'll try to put together a build with the patch from bug 852565 and see if it helps.

Comment 13

[Chris Lord [:cwiiis]]

(In reply to Wesley Johnston (:wesj) from comment #12)
> I can see this pretty easily by just trying to pan quickly on Facebook.
> Flipping the pref off fixes it on the current nightly.
> 
> It seems like we do have pretty solid STR on more than one device (although
> I don't know why that matters). I'm going to reopen until this is fixed with
> the dynamic toolbar on. I'll try to put together a build with the patch from
> bug 852565 and see if it helps.

Do you think you could upload a video perhaps?

Comment 14

[Demetrius S.]

This is very noticeable happening on polygon.com it makes the site unnavigable.

Comment 15

[Chris Lord [:cwiiis]]

(In reply to Demetrius Schwab from comment #14)
> This is very noticeable happening on polygon.com it makes the site
> unnavigable.

This is interesting, because I browser Polygon all the time and haven't experienced this (and checking now just in case, it's still fine) - could you provide some details about what build you experienced this on and what device?

Comment 16

[Demetrius S.]

(In reply to Chris Lord [:cwiiis] from comment #15)
> (In reply to Demetrius Schwab from comment #14)
> > This is very noticeable happening on polygon.com it makes the site
> > unnavigable.
> 
> This is interesting, because I browser Polygon all the time and haven't
> experienced this (and checking now just in case, it's still fine) - could
> you provide some details about what build you experienced this on and what
> device?

I noticed it on the latest build 2013-03-21 with my nexus 4.

Comment 17

[Chris Lord [:cwiiis]]

(In reply to Demetrius Schwab from comment #16)
> (In reply to Chris Lord [:cwiiis] from comment #15)
> > (In reply to Demetrius Schwab from comment #14)
> > > This is very noticeable happening on polygon.com it makes the site
> > > unnavigable.
> > 
> > This is interesting, because I browser Polygon all the time and haven't
> > experienced this (and checking now just in case, it's still fine) - could
> > you provide some details about what build you experienced this on and what
> > device?
> 
> I noticed it on the latest build 2013-03-21 with my nexus 4.

If you could try again on the 03-22 build when it's available, that'd be helpful - I think bug 852565 will be included with that.

Comment 18

[Chris Lord [:cwiiis]]

ok, I see this on Nightly but didn't see it in my local build based on 125375:2c41bf87b4e5 with patches on top.

Let's get a regression window on this. I don't think this was caused by bug 716403, at least certainly not any of the patches that are attached to that bug (possibly from some follow-up?)

If we could stop blaming bug 716403 before actually confirming, that'd be great too :)

Comment 19

[Chris Lord [:cwiiis]]

hmm, just noticed my update was out of date, I tried on 03-21. This could still be bug 852565.

Comment 20

[Mathieu Pellerin]

Bug 852565 seems to have taken care of this bug, I can't reproduce the panning break / jump back since it was committed to mozilla-central.

That said, I'm pretty sure it led to a new regression, easily reproduced:
1. Launch Firefox
2. Open a web site (for the sake of continuity, let it be m.facebook.com)
3. When trying to touch / click hyperlinks, the viewport mapping of elements, such as hyperlinks, is shifted down vertically, roughly the height of the toolbar.

Comment 21

[Mark Finkle (:mfinkle) (use needinfo?)]

(In reply to Mathieu Pellerin from comment #20)
> Bug 852565 seems to have taken care of this bug, I can't reproduce the
> panning break / jump back since it was committed to mozilla-central.
> 
> That said, I'm pretty sure it led to a new regression, easily reproduced:
> 1. Launch Firefox
> 2. Open a web site (for the sake of continuity, let it be m.facebook.com)
> 3. When trying to touch / click hyperlinks, the viewport mapping of
> elements, such as hyperlinks, is shifted down vertically, roughly the height
> of the toolbar.

Please file a new bug with these steps to reproduce. Thanks

Comment 22

[Mathieu Pellerin]

Mark, filed bug 852565.

Comment 23

[Kartikaya Gupta (email:[e-mail])]

(In reply to Mathieu Pellerin from comment #22)
> Mark, filed bug 852565.

That should be bug 854289.