[1.Description]: [Flame KK v2.5][Aries KK v2.5][Dialer] When you are dialing, input some digits then suggestion appears. Tap dailing button at once. One call is dialed out, at the same time , another ended call is displayed. Time: 16:53 See attachments: logcat_1653.txt and Aries_KK v2.5.3gp [2.Testing Steps]: Precondition: The DUT has an SIM card, and there are some contacts in contact app. 1.Launch dialer app. 2.Input a part of the phone number. 3.Tap suggestion. 4.Tap dailing button at once. [3.Expected Result]: 4.Dial out only one call. [4.Actual Result]: 4. Two same calls are dialed out, but one of them displays as ended. [5.Reproduction build]: Device: AriesKK v2.5(Affected) Build ID 20150820005800 Gaia Revision 89e0096a3de0378e3eda77e6a2a0bb5ca03eb8bb Gaia Date 2015-08-19 18:28:14 Gecko Revision https://hg.mozilla.org/integration/mozilla-inbound/rev/29b2df16e961fbe9a379362ecba6f888d1754bc3 Gecko Version 43.0a1 Device Name aries Firmware(Release) 4.4.2 Firmware(Incremental) eng.worker.20150820.002839 Firmware Date Thu Aug 20 00:28:46 UTC 2015 Bootloader s1 Device: FlameKK v2.5(Affected) Build ID 20150820150207 Gaia Revision c6705f739fb605031eb2a0b943ba55c64bee5a03 Gaia Date 2015-08-20 14:36:40 Gecko Revision https://hg.mozilla.org/mozilla-central/rev/095988abdc560bf8ba07a94a425c6922a3e9bfd6 Gecko Version 43.0a1 Device Name flame Firmware(Release) 4.4.2 Firmware(Incremental) eng.cltbld.20150820.183055 Firmware Date Thu Aug 20 18:31:07 EDT 2015 Bootloader L1TC000118D0 Device: FlameKK v2.2(Unaffected) Build ID 20150820032506 Gaia Revision 335cd8e79c20f8d8e93a6efc9b97cc0ec17b5a46 Gaia Date 2015-08-14 19:06:41 Gecko Revision https://hg.mozilla.org/releases/mozilla-b2g37_v2_2/rev/ebaa55c4247b Gecko Version 37.0 Device Name flame Firmware(Release) 4.4.2 Firmware(Incremental) eng.cltbld.20150820.065310 Firmware Date Thu Aug 20 06:53:22 EDT 2015 Bootloader L1TC000118D0 [6.Reproduction Frequency]: Always Recurrence,10/10 [7.TCID]: Free test

[Blocking Requested - why for this release]: Functional regression on the Dialer.

[Triage] blocker as it's a basic function regression.

This is actually a similar issue to bug 1193563 and thus depends on bug 1147735 for fixing.

WIP patch, built on top of bug 1147735. This should fix bug 1193563 too.

OK, this turns out to be a lot more complicated than I thought. The time it takes for the call promise to be resolved ranges from seconds to 100s of ms. In the latter case it's too short and doesn't prevent the user from making a second call. The fact that the callscreen and dialer are split doesn't make this any easier as we can't wait for the callscreen to open in response to a call starting, we know nothing about its state.

[Blocking Requested - why for this release]: I'd like to de-nom this bug for a number of reasons. First of all with the dialer and callscreen being split as they currently are it's impossible to fix. The fix here would be to prevent further calls until the callscreen is displayed but there's no way to tell from the dialer. As an alternative we might abuse the settings API for doing pseudo-IPC between the two apps (but the thought of it makes me cringe) or just use a timeout to prevent further calls. Both are horrible hacks, not robust and prone to introducing worse races than this one. The proper fix would be to bring back the callscreen into the dialer first, but this is not doable in the 2.5 timeframe. Additionally bug 1193563 which is this same issue but with a far simpler STR is not a 2.5 blocker, so in my eyes either both bugs are blockers or neither is.

[Tracking Requested - why for this release]: Corner case, that will require mayor work. Triage don't think is worthy to risk. Unnominating, but setting priority to p1 to be fixed after 2.5 branch, if the final fix is acceptable we could uplift to 2.5

Firefox OS is not being worked on