999352 - [Tarako][Dialer]Press down and drag keypad character will get character always highlight

Status: VERIFIED FIXED

(Firefox OS Graveyard :: Gaia::Dialer, defect)

Description

[Mike Lien[:mlien]]

Attachment: logcat.txt

[Device]
  Tarako
---------------------------------------------
[Reproduction build] - 0417pac + PVT v1.3T
  Gaia      3f1d8332d127f1d6bc0fdbb08146ce1deb0efbc0
  Gecko     https://hg.mozilla.org/releases/mozilla-b2g28_v1_3t/rev/5da76152c2bd
  BuildID   20140421164001
  Version   28.1

---------------------------------------------
[Reproduce Steps]
  1. Launch Dialer app
  2. Press down and drag keypad character
---------------------------------------------
[Expected Result]
  After press up should release the highlight
---------------------------------------------
[Actual Result]
  Character always highlight
---------------------------------------------
[Reproduce Rate]
  100%

Comment 1

[Joe Cheng [:jcheng] (please needinfo)]

ni? Etienne

Comment 2

[Etienne Segonzac (:etienne)]

Can't reproduce the bug neither on master nor the current 1.3t.

Comment 3

[bhavana bajaj [:bajaj]]

Its a minor UI glitch and functionality is not broken, you can double-tab and remove the highlight or tap another digit and still get going.

Comment 4

[Mike Lien[:mlien]]

verify with today's daily build, still can reproduce this issue
Gaia      706f474230432c8cb01e124aee2c055ec902aa1d
Gecko     https://hg.mozilla.org/releases/mozilla-b2g28_v1_3t/rev/26ef3dded9ff
BuildID   20140424014003
Version   28.1

build path is
https://pvtbuilds.mozilla.org/pvt/mozilla.org/b2gotoro/nightly/mozilla-b2g28_v1_3t-tarako/

Comment 5

[Joe Cheng [:jcheng] (please needinfo)]

QAWANTED on master
if this doesn't reproduce on master, let's close this bug. thanks

Comment 6

[jcastro]

Unable to reproduce on Buri with master (2.0 from 4/30)

2.0 Environmental Variables:
Device: buri 2.0 MOZ
BuildID: 20140430040206
Gaia: badc73ee7f108fa631150bded0cc57e92aad810e
Gecko: e19812f56952
Version: 32.0a1
Firmware Version: V1.2-device.cfg

Comment 7

[ying.xu]

easy to reproduce, with latest build

on dialer window
press number 1, then move to number 2, release

refer
http://bugzilla.spreadtrum.com/bugzilla/show_bug.cgi?id=311074

Comment 8

[Evelyn Hung]

(In reply to ying.xu from comment #7)
> easy to reproduce, with latest build
> 
> on dialer window
> press number 1, then move to number 2, release
> 
> refer
> http://bugzilla.spreadtrum.com/bugzilla/show_bug.cgi?id=311074

You should describe a clear reproduce step like statement above when you filing this bug, so we won't waste time on "can't reproduce" conversation. Also, please don't put your bugzilla link here, because it doesn't help - most people can't read Chinese. Thanks.

Comment 9

[Kai-Zhen Li [:seinlin][:kli]]

Attachment: 2012-01-01-18-08-57.png

I can reproduce this bug and other two similar bugs on partner's bugzilla. Attached file is a sample taken from Gallery.

STR:
- In Music, Gallery or Dailer
- Press down and drag far away from hightlight button
- Release press, hightlight won't be turn off

Comment 10

[ying.xu]

TabChild::UpdateTapState(const WidgetTouchEvent& aEvent, nsEventStatus aStatus)
{
  
  bool currentlyTrackingTouch = (mActivePointerId >= 0);   //here
  if (aEvent.message == NS_TOUCH_START) {
    if (currentlyTrackingTouch || aEvent.touches.Length() > 1) {
      // We're tracking a possible tap for another point, or we saw a
      // touchstart for a later pointer after we canceled tracking of
      // the first point.  Ignore this one.
      return;
    }
    if (aStatus == nsEventStatus_eConsumeNoDefault ||
        nsIPresShell::gPreventMouseEvents ||
        aEvent.mFlags.mMultipleActionsPrevented) {
      return;
    }

    Touch* touch = aEvent.touches[0];
    mGestureDownPoint = LayoutDevicePoint(touch->mRefPoint.x, touch->mRefPoint.y);
    mActivePointerId = touch->mIdentifier;
    if (sClickHoldContextMenusEnabled) {
      MOZ_ASSERT(!mTapHoldTimer);
      mTapHoldTimer = NewRunnableMethod(this,
                                        &TabChild::FireContextMenuEvent);
      MessageLoop::current()->PostDelayedTask(FROM_HERE, mTapHoldTimer,
                                              sContextMenuDelayMs);
    }
    return;
  }

  // If we're not tracking a touch or this event doesn't include the
  // one we care about, bail.
  if (!currentlyTrackingTouch) {    //here, return directly when handling touchend event
    return;
  }
...
 switch (aEvent.message) {
  case NS_TOUCH_MOVE:
    if (abs(currentPoint.x - mGestureDownPoint.x) > sDragThreshold.width ||
        abs(currentPoint.y - mGestureDownPoint.y) > sDragThreshold.height) {
      CancelTapTracking();   //here, reset mActivePointerId to -1
    }
    return;

Comment 11

[Etienne Segonzac (:etienne)]

I can reproduce the issue now but it looks like a platform issue.

Comment 12

[Steven Yang [:styang]]

Viral, could you help to check this issue? thanks.

Comment 13

[yang.zhao]

It caused by bug 972081.

Comment 14

[viral [:viralwang]]

I'm in business trip now.

Few quick questions here:

(In reply to Etienne Segonzac (:etienne) from comment #11)
> I can reproduce the issue now but it looks like a platform issue.
Any information can help to us to understand it's platform issue?

(In reply to yang.zhao from comment #13)
> It caused by bug 972081.
is that means the bug 999352 is disappear if you revert the bug 972081.

Comment 15

[Kai-Zhen Li [:seinlin][:kli]]

After revert this commit, https://hg.mozilla.org/releases/mozilla-b2g28_v1_3t/rev/dcaf06de49f5

The problem on dailer will be fixed. 

But it still happens the on "return" button gallery edit page and music playing page.

Comment 16

[Joe Cheng [:jcheng] (please needinfo)]

ni? Steven
the triage team does not feel this is a blocker.

Comment 17

[James Zhang (Spreadtrum)]

(In reply to Joe Cheng [:jcheng] from comment #16)
> ni? Steven
> the triage team does not feel this is a blocker.

It's our blocker issue, we have found the root cause, why don't fix it?
revert this commit, https://hg.mozilla.org/releases/mozilla-b2g28_v1_3t/rev/dcaf06de49f5

Comment 18

[Steven Yang [:styang]]

(In reply to Joe Cheng [:jcheng] from comment #16)
> ni? Steven
> the triage team does not feel this is a blocker.

This is a partner-blocker.

Comment 19

[Joe Cheng [:jcheng] (please needinfo)]

Responding to James. 
we should understand what's the original purpose of this fix? https://hg.mozilla.org/releases/mozilla-b2g28_v1_3t/rev/dcaf06de49f5 and what happens when it gets reverted? Thanks  

also ni? Fabrice

Comment 20

[thomas tsai]

apzc is disabled in tarako.

Comment 21

[Vivien Nicolas (:vingtetun) (:21) - (NOT reading bugmails, needinfo? please)]

(In reply to Joe Cheng [:jcheng] from comment #19)
> Responding to James. 
> we should understand what's the original purpose of this fix?
> https://hg.mozilla.org/releases/mozilla-b2g28_v1_3t/rev/dcaf06de49f5 and
> what happens when it gets reverted? Thanks  
> 
> also ni? Fabrice

The original purpose of this fix was to fix the highlighting interaction with APZC enabled. On the Tarako branch, APZ is disabled from what I understood. So in order to fix the issue the best think is likely to revert the patch locally (https://bug972081.bugzilla.mozilla.org/attachment.cgi?id=8394685)

Comment 22

[[:fabrice] Fabrice Desré]

I backed out on 1.3t only:
 https://hg.mozilla.org/releases/mozilla-b2g28_v1_3t/rev/a3a14fb5ad51

Comment 23

[Kai-Zhen Li [:seinlin][:kli]]

Attachment: close button is highlighted

Similar issue still happens on a different place.

STR
- Open gallery
- Open an image
- Into edit image
- make a touch on the screen
Then the 'close' button on the left top will always be highlighted. (as attached file)

Comment 24

[Naoki Hirata :nhirata (please use needinfo instead of cc)]

Khai-Zhen is that the same bug as 998086?  We noticed the highlight in other areas as well.

Comment 25

[Naoki Hirata :nhirata (please use needinfo instead of cc)]

Gaia      917174ee3812a43758bf43f7ba5f9416dcdb2ca8
Gecko     https://hg.mozilla.org/releases/mozilla-b2g28_v1_3t/rev/a3a14fb5ad51
BuildID   20140520014006
Version   28.1
ro.build.version.incremental=eng.cltbld.20140520.050850
ro.build.date=Tue May 20 05:08:57 EDT 2014
Tarako

Comment 26

[Kai-Zhen Li [:seinlin][:kli]]

Naoki, Yes it is the same as in bug 998086. So we don't need to file a new bug for this. Thanks!

Comment 27

[ying.xu]

we meet a crash after the modification in 
Comment 22  Fabrice Desré [:fabrice]  2014-05-19 13:27:45 PDT


Crash reason:  SIGSEGV
Crash address: 0x3c

Thread 0 (crashed)
 0  libxul.so!mozilla::dom::TabChild::RecvRealTouchEvent(mozilla::WidgetTouchEvent const&, mozilla::layers::ScrollableLayerGuid const&) [nsCOMPtr.h : 554 + 0x2]
     r4 = 0x42840880    r5 = 0xbec6a808    r6 = 0x41ecb92c    r7 = 0xbec6a740
     r8 = 0x00000001    r9 = 0xbec6a8b8   r10 = 0xbec6a7b8    fp = 0xbec6a7a8
     sp = 0xbec6a738    lr = 0x409f1b91    pc = 0x40e9834e
    Found by: given as instruction pointer in context
 1  libxul.so!mozilla::dom::PBrowserChild::OnMessageReceived(IPC::Message const&) [PBrowserChild.cpp : 2131 + 0xd]
     r4 = 0x42840880    r5 = 0xbec6aed4    r6 = 0xbec6a808    r7 = 0xbec6a8b8
     r8 = 0xbec6a970    r9 = 0x00000000   r10 = 0x00000003    fp = 0x00000000
     sp = 0xbec6a7e8    pc = 0x40b2b315
    Found by: call frame info
 2  libxul.so!mozilla::dom::PContentChild::OnMessageReceived(IPC::Message const&) [PContentChild.cpp : 3170 + 0x7]
     r4 = 0x4318d818    r5 = 0xbec6aed4    r6 = 0xbec6aef8    r7 = 0x41ecb92c
     r8 = 0x00000006    r9 = 0x4031e90c   r10 = 0x00000003    fp = 0x00000000
     sp = 0xbec6aa20    pc = 0x40b44339
    Found by: call frame info