Firefox/Channels/Meetings/2016-08-23
Channel Meeting Details
- Tuesdays and Thursdays at 10am Pacific Time
- irc.mozilla.org #planning for backchannel
- Past meeting notes: https://wiki.mozilla.org/Firefox/Channels/Meetings
Video/Teleconference Details - NEW
- 650-903-0800 or 650-215-1282 x92 Conf# 99951 (US/INTL)
- 1-800-707-2533 (pin 369) Conf# 99951 (US)
- Vidyo Room: ReleaseCoordination
- Vidyo Guest URL
Contents
Attendees
philipp, calixte, Jlund, Kevin, Blassey, RyanVM, Sylvestre, Lizzard, Ritu, Jst, Milan, shell, andym, dveditz
Previous Actions
Schedule Update
- 48.0.1
- Update are enabled for
- Windows user running 48.0
- GNU/Linux users (all versions)
- Mac OS X users (all versions)
- Disabled
- Windows running < 48
- Update are enabled for
- 48.0.2
- We have a workaround for the websense startup crash - https://bugzilla.mozilla.org/show_bug.cgi?id=1294650
- landed in nightly
- but we don't know how to test it for sure
- some volunteers proposed to help
- Asked to Jim to request the uplifts
- Hopefully will build today and ship... when we are sure we don't break firefox
- Interesting discussion on whitelisting dlls.
- 49.0b6 desktop released this morning
- mobile - built this morning, should release after testing.
- we just received the sign off
- (Callek) OpenH264 v1.6 shipped to *DESKTOP* ESR/release Wednesday.
- (jlund) desktop 1.6 shipped on Aug 17th.
- Delayed for Android on release (and esr incase someone built it locally) -- pending more complete testing from Maire's team (bug 1296002)
- android is still in same state as line above afaik. bhearsum is on the hook while callek is out
- beta, esr, release, android updates haven't gone out. yet. (not often used... .h264 devices to firefox
- Delayed for Android on release (and esr incase someone built it locally) -- pending more complete testing from Maire's team (bug 1296002)
Add-ons
- Update from Kev on 48.0.2, he found a product contact at ForcePoint who has promised an Engineering contact at Websense by the end of the day
- let's also talk with dees and see if we can enlist him here.
Stability
A Modest Dashboard Collection beta crash-stats: https://crash-stats.mozilla.org/topcrashers/?product=Firefox&version=49.0b more beta data: https://health.graphics/crashes/beta https://sql.telemetry.mozilla.org/dashboard/beta-crash-rates and more: https://crash-analysis.mozilla.com/rkaiser/crash-report-tools/longtermgraph/?fxbeta e10s/non crash rate: https://telemetry.mozilla.org/crashes/office-dashboard.html addons e10s crash rate (currently beta): https://sql.telemetry.mozilla.org/dashboard/stability-comparison-for-e10s-add-ons-experiment
Aurora
Top crashers - Bug 1051567 - [e10s] Crash in [@ mozilla::ipc::MessageChannel::OnChannelErrorFromLink]
Bug 1273770 - AsyncShutdown timeout in session store
Beta
[philipp] 49.0b5:
- crash spike specific to particular intel cpus that started in beta 4 is continuing (6-7% of all crashes): https://bugzilla.mozilla.org/show_bug.cgi?id=1296630
- suggest we consider this a blocker
- crash in nsXPCWrappedJS::nsXPCWrappedJS https://bugzilla.mozilla.org/show_bug.cgi?id=1295033 is 1.1% of all browser crashes
- crash in MustSkipMarking<T> https://bugzilla.mozilla.org/show_bug.cgi?id=1275172 is 0.6%
- regressing in volume in 49 but no activity in the bug report
- crash Abort | ... | mozilla::dom::PBrowser::Transition https://bugzilla.mozilla.org/show_bug.cgi?id=1269036 is 0.5%
- network top crash has been mitigated somewhat but not fixed in beta 5. https://bugzilla.mozilla.org/show_bug.cgi?id=1292181
Release
browser crash rate: 0.92 (green), content: 0.01
- working on a crappy fix addressing the startup crash with websense endpoint (using a dummy dll) - https://bugzilla.mozilla.org/show_bug.cgi?id=1294650
- we depend on affected users replying in the bug report to verify if this works
- are we in communication with the vendor?
Nightly
Mobile
Performance (Thu)
READ ONLY raw notes (with links to graphs, etc.) - https://public.etherpad-mozilla.org/p/vtNqXxJuTY
Aurora / Dev Edition
* pending bugs: * win7 canvasmark: https://bugzilla.mozilla.org/show_bug.cgi?id=1287707 (fix in progress) * linux64 tabpaint: https://bugzilla.mozilla.org/show_bug.cgi?id=1287707 (r?) * winxp cart: https://bugzilla.mozilla.org/show_bug.cgi?id=1289321 (ni? - thanks for marking this a blocker)
* e10s vs non-e10s: * kraken - osx10 getting slightly worse than non-e10s (seen a few weeks ago: https://public.etherpad-mozilla.org/p/5hVO4pOYIA#lineNumber=56 ) * tart - win8 getting slightly worse than non-e10s (TODO: figure out the difference)
Beta
- no pending bugs to resolve- all changes are acknowledged
- e10s vs non-e10s
* kraken - linux64/osx10 getting slightly worse than non-e10s (seen a few weeks ago: https://public.etherpad-mozilla.org/p/5hVO4pOYIA#lineNumber=50 ) * tart - win8/winxp getting slightly worse than non-e10s (TODO: figure out the difference)
Release
Release health
Beta: 2 blocking, 12 new, 62 carryover Aurora: 3 blocking, 59 new, 117 carryover
- Triage day changing to Wed 11am PT, async triage still happening as well
QE
Nightly
Aurora / Dev Edition
Beta
Action items
Mobile
- Fennec 49 Beta 6 signed off without major issues - https://wiki.mozilla.org/QA/Fennec/49/Beta/6
RelEng
(callek) * There was a planned WNP for zh-TW for beta5, driven by mkaply's group. -- Testing with it with beta4 reminded us we can't show a WNP without a Gecko Milestone bump. -- We'll still likely need a WNP for that locale for 48.x-->49.0
Liz will talk with mkaply. ** Bug 1292637 ** Update: This can still happen for 49 release but testing will be last minute. (We may not need a relbranch here, just a test update channel )
Marketing/Communications
User Advocacy
[philipp] sumo user reporting hangs switching between tabs with e10s on 48: https://support.mozilla.org/en-US/questions/1135499
Roundtable
* did i miss mention of beta 5 - non-e10s app crash spike starting Aug 19th https://telemetry.mozilla.org/crashes/office-dashboard.html
* seeing spike in add-ons dashboard https://sql.telemetry.mozilla.org/dashboard/stability-comparison-for-e10s-add-ons-experiment
* could be a data spike at start of Beta (but seems like enough days to have sufficient data to level), but will follow up to see if correlates with soccoro
- result of followup, no, this was a data aggregation timing issue not a real spike
Special Topics
- open h264 release follow up
- android N stability?
Aurora/Beta Feature Review
Post-Mortem (Tues 2wks from GA Release)
- post mortem still postponed till we know more about e10s. Erin will also run a postmortem.