1293999 - [animation][async XHR][event] content disappears after animated xhr event

Status: RESOLVED INCOMPLETE

(Firefox for Android Graveyard :: General, defect)

Description

[Robert Utasi [:hunboy]]

Testcase: http://utasirobert.com/getter.html

AR: open the testcase, and click to a link. It does new content for #content but disappears. Pan the touch screen, it appears randomly for a short time.
ER: you can browse the menuitems via XHR

This is a really weird bug, I can repro exclusively on my android 4.2 mobile phone with the android firefox 48. (clean install)

I tested it on desktop firefox, but works properly. Also it works properly in chromium (desktop,mobile too).

Although this is a permanent bug on android platform.

I cannot localise the issue, so the testcase is not reduced, but standard and simple. 1 object. When you click on a menuitem, the XHR starts, at state1 fires 1 transition, to hide (fade-out) the actual content. When the transition ends, it loads the new content to the node #content, and fires transition to fade-in the new content. When the content is not ready if the first animation finishes it waits for the state4 of XHR.

Whilst the android is a FreeBSD linux, any confirmation would be appreciated from a linux user, while I have no linux at this moment.

I file this bug for Android 4.2 FF

Comment 1

[Botond Ballo [:botond]]

(In reply to Robert Utasi [:hunboy] from comment #0)
> Whilst the android is a FreeBSD linux, any confirmation would be appreciated
> from a linux user, while I have no linux at this moment.

I can confirm that the page works fine for me on Linux desktop.

It also works fine for me on Android 4.0 and Android 5.0. I don't have a 4.2 device handy to test on.

Comment 2

[Botond Ballo [:botond]]

QA, do you have a 4.2 device you can test on? Can you reproduce?

Comment 3

[Timothy Nikkel (:tnikkel)]

Seems to work fine for me on Android 6 as well.

Comment 4

[Hiroyuki Ikezoe (:hiro)]

I can't reproduce on Android 4.2.2 and 4.4 either.
I don't know what the difference between 4.2 and 4.2.2 is.

Comment 5

[Botond Ballo [:botond]]

Are you able to get a log via "adb logcat" during the problematic interaction with the page?

Comment 6

[Robert Utasi [:hunboy]]

Botond, not really. This is a chinese clone vertis basecore9+ tablet, and i have no experience by using adb, so can't debug. I just purchased it for testing my webpages on android by a clean install browsers. Nothing hacky. I just needed a real end user experience by this tablet.

Comment 7

[Robert Utasi [:hunboy]]

Some additional info, I installed CatLog Viewer to this tablet, and this event caused to work the code.

The LogCat says: Failed to swap drawable from org.mozilla.firefox

In other way this looks to me as a some type of reflow issue. When the tablet went to sleep, and I relogged, the actual state was shown well, but no more activity.

Comment 8

[Botond Ballo [:botond]]

(In reply to Robert Utasi [:hunboy] from comment #7)
> The LogCat says: Failed to swap drawable from org.mozilla.firefox

This message isn't produced by Firefox. A Google search shows that some Android components in SystemUI, like SearchPanelView [1], can produce it. Jim, do you know if we use any such Android components, and if so, whether this message might be related to the problem in this bug?

[1] https://android.googlesource.com/platform/frameworks/base/+/0b68ff4/packages/SystemUI/src/com/android/systemui/SearchPanelView.java#205

Comment 9

[(inactive) Jim Chen [:jchen] [:darchons]]

I'm not sure... redirecting to Sebastian.

Comment 10

[Sebastian Kaspari (:sebastian; :pocmo)]

I don't think this is necessarily related. @Robert: Can you copy more from the log while loading and using the website to pastebin.mozilla.org and post that here?

However I think we will still need to reproduce this (I don't expect the log to provide much here) and we need to find someone from the Gecko team who can debug this. The tablet seems to be pretty cheap so maybe we can just buy one. But it it is not easy to get anymore (at least here).

Apart from a low resolution I can't see anything particular weird:
http://www.imei.info/phonedatabase/13538-overmax-basecore-9/

Comment 11

[Robert Utasi [:hunboy]]

Attachment: catlog screenshot

here is a catlogscreenshot. Actually while I installed the catlog viewer I cannot reproduce anymore the bug itself. so it's weird.

probably we can close this bug as RESOLVED INCOMPLETE

incomplete - because we can't reproduce anymore the bug, but it occured on a clean install tablet. After installing catlog viewer this is not a clean install anymore.

Comment 12

[Botond Ballo [:botond]]

Closing as per comment 11. Feel free to reopen if you can reproduce the issue again.