1121805 - [Dialer][dolphin][FFOS7715 v2.1] [performance] The time spent on hiding callscreen after hangup is much longer

Status: RESOLVED WONTFIX

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

Description

[helloarvin]

*** Build Information
Gaia-Rev        17c7ad2e4919a994f0844239b483116090412dee
Gecko-Rev       4f7c43e577a1a43be0af2c02383c1c56c8c4b756
Device-Name     dolphin(ea)

*** Description
Hang up the call, the time spent on hiding callscreen is much longer.

*** Steps to Reproduce
Pre-conditions: Call-logs(50);Contacts(250);Messages(100);Pictures(300);Songs(150);Videos(150)

1. Click the hangup button to disconnect the active call;
2. Stop the timer once the call-screen hide and the background app display completed.

*** Expected Results
The performance should be the same as 7715ea-android.

*** Actual Results
The time took for hiding callscreen after pressed to hangup is worse than 7715ea-android and 7715ea-FFOSv1.4.

The average time
7715ea-FFOSv2.1 ：3.816s
Flame-FFOS  ： 3.366s
7715ea-Android: 1.372s
7715ea-FFOSv1.4 : 1.956s

Comment 1

[Steven Yang [:styang]]

According the results, seems like an serious regression. 2.1S?

Comment 2

[Steven Yang [:styang]]

William, need your help to update the statistics on 2.1s pvt build.

Comment 3

[William Hsu [:whsu]]

(In reply to Steven Yang [:styang] from comment #2)
> William, need your help to update the statistics on 2.1s pvt build.

Roger that!
I will take some actions after PVT build is ready

Comment 4

[William Hsu [:whsu]]

Dolphin v2.1 (512MB) test result.
Visually complete time: 3710 ms

@ PVT Build information:
 - Gaia-Rev        2d0df3907319edf55a643b7d4a103534579ebef0
 - Gecko-Rev       https://hg.mozilla.org/releases/mozilla-b2g34_v2_1s/rev/2a3804a0911a
 - Build-ID        20150128161200
 - Version         34.0
 - Device-Name     scx15_sp7715ea
 - FW-Release      4.4.2
 - FW-Incremental  eng.cltbld.20150128.193949
 - FW-Date         Wed Jan 28 19:40:01 EST 2015

Comment 5

[Steven Yang [:styang]]

It should be a performance regression, please arrange resource to profile for the bottleneck.

Comment 6

[William Hsu [:whsu]]

NI?whsu to do extra performance comparison. (Flame 512 MB: v1.4 vs v2.1)

Comment 7

[Sherman Chen [:chens] (inactive)]

I think this is intended to have 2 seconds delay before we close callscreen, and quote from bug 1068109 comment 0, this delay is needed so that the user can see the duration of the call that just ended. 
And it has reasonable reading (3.816s/1.956s) from QA, which is exactly the delay time we need.

Comment 8

[Vance Chen [:vchen][skype:vance.lucida][[e-mail]]]

Hi Chens -

Is this delay configurable? I am thinking if SPRD is not happy about the 2 seconds delay, maybe they can reduce it to one to meet their QA requirement

Thanks

Comment 9

[Sherman Chen [:chens] (inactive)]

Yes it is configurable, try change `callEndPromptTime`[1] to a proper value may work. But can't say there will be no side effect once we change that.

[1]: https://github.com/mozilla-b2g/gaia/blob/862521678b7e77ef606ec483caf50d549f78c34f/apps/callscreen/js/call_screen.js#L11

Comment 10

[Vance Chen [:vchen][skype:vance.lucida][[e-mail]]]

Hi Arvin -

As Chens mentioned in Comment#7, we deliberately have 2 seconds delay while closing the call screen. The reason for that is to give user enough time to check the call duration. So this is per design. We would suggest you to communicate with your QA and close this issue. However, if you guys really want to reduce the time, you can try to modify the time value Chens mentioned in Comment#9, but again, can't guarantee there will be no side effect, and Mozilla won't make this modification on our 2.1s branch

Thanks for your understanding

Comment 11

[helloarvin]

Thanks for chens and vance.

I'd discussed with our QA before on the '2 seconds delay' while closing the call screen but they insisted that the time consuming of hiding callscreen should at least not be more than that on v1.4. So I'll try to change the 'callEndPromptTime' to check if it can satisfy the minimum requirement(1.9s on v1.4).

Keep ni me flag alive.

Comment 12

[Vance Chen [:vchen][skype:vance.lucida][[e-mail]]]

Per Comment#11, remove the nomination first. Please ni me again if you need further discussion about this bug

Comment 13

[helloarvin]

sprd fixed locally.

Thank you all.

Comment 14

[Johan Lorenzo [:jlorenzo]]

Moving the resolution to WONTFIX as it won't be fixed on Mozilla's trees.