903498 - [B2G][Unagi][Browser] Browser shakes while scrolling down

Status: RESOLVED WORKSFORME

(Core :: Graphics: Layers, defect)

Description

[Allen Maxwell]

Attachment: LogCat of issue

Repro Steps:
1) Updated Unagi to v1.2.0 Build ID: 20130809040201
2) Connect WiFi to a network
3) Open browser
4) Enter any website (I used cnn.com and abc.com)

Actual:
Browser is shakey while scrolling and keeps resetting the scroll position to the top of the screen.

Expected:
User is able to scroll page.

Environmental Variables
Build ID: 20130809040201
Gecko: http://hg.mozilla.org/mozilla-central/rev/e33c2011643e
Gaia: c354940fbe112bcf2d90e2a5654ad1824f3a2348
Platform Version: 26.0a1

Notes:
Repro frequency: 100%
Related to bug 901339
See attached: 
browserLog.txt

Comment 1

[Jason Smith [:jsmith]]

kats - Is this related to the APZC work you've been doing lately? Any ideas why this happening?

Comment 2

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

It might be related. I haven't seen this behaviour on my Peak device so I'm not sure what's causing it. Is it possible to get a regression window?

Comment 3

[Natalya Kot [:nkot]]

regression range:

Build ID: 20130731030205 - reproduces
Gecko: http://hg.mozilla.org/mozilla-central/rev/c2b375f3a909
Gaia: 9bfceaa90e8b92a379432b67121afa3cd3f14c90
Platform Version: 25.0a1

Build ID: 20130730030200 - does NOT reproduce
Gecko: http://hg.mozilla.org/mozilla-central/rev/3d40d270c031
Gaia: ba5ff211fbf6a930326cc6a0d4a1205a7528630b
Platform Version: 25.0a1

*this bug only reproduces for web sites with ads

Comment 4

[Jason Smith [:jsmith]]

Looks like yet another regression from bug 866232.

Comment 5

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

Can you describe the "shakey" behaviour? The scroll resetting should be fixed by bug 895905.

Comment 6

[Jason Smith [:jsmith]]

(In reply to Kartikaya Gupta (email:[e-mail]) from comment #5)
> Can you describe the "shakey" behaviour? The scroll resetting should be
> fixed by bug 895905.

QA Wanted to test with that patch included. If it still reproduces, include a video of this bug to help answer the question in comment 5.

Comment 7

[Allen Maxwell]

Attachment: Video of Issue

Browser continues to be shaky while scrolling and continues resetting the scroll position to the top of the screen.

Build ID: 20130819041202
Gecko: http://hg.mozilla.org/releases/mozilla-b2g18/rev/285e2afe2a3b
Gaia: 124ed9737f0f44781cb75f763468f1574f753ac0
Platform Version: 18.1

Comment 8

[Allen Maxwell]

Incorrect Build data in Comment 7 used as follows:

Build ID: 20130819040203
Gecko: http://hg.mozilla.org/mozilla-central/rev/c8c9bd74cc40
Gaia: f6de05c135913f2cb790759335875bb1b3c4f9bb
Platform Version: 26.0a1

Comment 9

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

(In reply to Allen Maxwell from comment #7)
> Browser continues to be shaky while scrolling and continues resetting the
> scroll position to the top of the screen.

Hm, I'm unable to reproduce this behaviour unfortunately. From the video it looks like there was only one "scroll reset" and that happened right when the page stopped loading, so it's possibly done by the page itself. The unpainted content at the end does look like a real bug, I will try to investigate what's going on there.

Comment 10

[Allen Maxwell]

The shaky scrolling appears to have been resolved, WFM. 
The page reset to the top not only occurred at end of the page load but upon the ad (progressive) reloading. After browsing several sites and pages the page resets upon ad reload on many of them. 

If the reset to the top of the page is normal then this bug could be resolved as WFM.

Build ID: 20130820040204
Gecko: http://hg.mozilla.org/mozilla-central/rev/bb025b6949e8
Gaia: 7249b5d61b955c23efe9e18f80f0da2e78d827eb
Platform Version: 26.0a1

Comment 11

[Jason Smith [:jsmith]]

The page reset rate in comment 10 does seem a bit concerning, so let's open a separate bug for that. Given that the original bug as filed isn't reproducing anymore, let's close as works for me. I'm assuming the unpainted content is being tracked in bug 907179.

Comment 12

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

Yes, the unpainted content should be dealt with in bug 907179. Bug 904648 was also filed about some scroll resetting but I wasn't able to reproduce that either (except for one, which was caused by content). People who could reproduce it previously can also not reproduce it now, so it could be something that was fixed already.