Firefox/Channels/Meetings/2017-09-07
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
lizzard, jcristau, jlorenzo, RyanVM, bogdan_maris, philipp, cornel, Julia, brindusa,jorge,Timea, Roxana, Hani, Ovidiu, Ninu, Tom, Adrian, nicole, selena,Sorina, petruta
Previous Actions
Schedule Update
- 56 Beta 10 gtb today
- schedule change for 57 beta: https://docs.google.com/document/d/1jeypuqBqEyIh-4qxXT0UnE2aVjetN7uVD8W7L7TbWKg/edit
Add-ons
Shield studies
- Quick review of shield studies or experiments.
- Is anything waiting for signoff or testing?
- Shield studies in flight: https://gregglind.github.io/x-shield-inflight-dashboard
Activity stream study continuing
Stability
Nightly
awsy rate: 4.32 (browser 2.10, content: 2.22) telemetry (m+c-s) from tuesday: 7.46 (week before: 4.95, this time last cycle: 12.49)
[philipp] top crashers in current nightly builds - most of them have fixes already...
- Crash in RtlEnterCriticalSection | mozilla::net::CacheStorageService::CacheQueueSize: https://bugzilla.mozilla.org/show_bug.cgi?id=1395938 (45% of browser crashes)
- Crash in gfxPlatform::InitAcceleration | gfxWindowsPlatform::InitAcceleration: https://bugzilla.mozilla.org/show_bug.cgi?id=1397214 (28% of content crashes)
- Crash in mozilla::image::ImageSurfaceCache::SuggestedSize: https://bugzilla.mozilla.org/show_bug.cgi?id=1397235 (16% of content crashes)
- Crash in nsPIDOMWindowener::UpdateWebSocketCount: https://bugzilla.mozilla.org/show_bug.cgi?id=1396368 (9% of content crashes)
Beta
awsy rate: 1.19 (browser 0.92, content 0.28) telemetry (m+c-s) from tuesday: 5.04 (week before: 5.12, this time last cycle: 4.33)
win64 migration with 56.0b9 unearthed a number of startup crash issues with 64bit (startup crashes on awsy have gone up 50% from 0.29 to 0.43):
- Win8 crashes with Kaspersky's klsihk64.dll: https://bugzilla.mozilla.org/show_bug.cgi?id=1268470 (20.7% of browser crashes)
- in communications with kaspersky about this; will try if blocklisting the crashing module has an effect
- Is there a replacement webextension for crashmenow ?
- For now we use extensions.allow-non-mpc-extensions;true & extensions.legacy.enabled;true and we add the xpi directly. It will appear in Menu Bar- Tools.
- https://ftp.mozilla.org/pub/utilities/crashfirefox-intentionally/ works too
- Crash in CrashReporter::OOPInit: https://bugzilla.mozilla.org/show_bug.cgi?id=1347867 (17.2% of browser crashes)
- Crash in abort | _acrt_RtlGenRandom with Lenovo OneKeyTheater: https://bugzilla.mozilla.org/show_bug.cgi?id=1369361 (0.7% of browser crashes)
- Lenovo utility preinstalled on some older ideapad notebooks, will try to confirm with affected users if blocklisting the modules works
- Crash in ... | mozilla::Preferences::AddBoolVarCache: https://bugzilla.mozilla.org/show_bug.cgi?id=1276488 (1.7% of browser crashes)
- diagnostics patch waiting for uplift approval
- Crash in Abort | new XPCJSContext failed to initialize. | mozalloc_abort (QuickHeal Antivirus): https://bugzilla.mozilla.org/show_bug.cgi?id=1397750 (1% of browser crashes)
Release
awsy rate: 0.81 (browser 0.59, content 0.23) telemetry (m+c-s) from tuesday: 3.08 (week before: 3.32, this time last cycle: 2.61)
no news
Mobile
awsy rate - release 0.66, beta 0.81, nightly 3.31
nightly rate a fair bit higher than reported on tuesday, but top crashers all received patches:
- ExpirationTrackerImpl<T>::RemoveObjectLocked: https://bugzilla.mozilla.org/show_bug.cgi?id=1397223
- java.lang.AssertionError: Expected condition to be true at org.mozilla.gecko.media.GeckoHlsPlayer.assertTrue(GeckoHlsPlayer.java): https://bugzilla.mozilla.org/show_bug.cgi?id=1396268
- mozilla::WrapNotNull<T>: https://bugzilla.mozilla.org/show_bug.cgi?id=1397235
beta & release rate is flat
- Crash in ResolvePromise with ublock origin webextension still slightly on the rise: https://bugzilla.mozilla.org/show_bug.cgi?id=1347984
Performance (Thu)
Aurora / Dev Edition
Beta
Release
QE
[Cornel]
- e10s multi testing [1393638] completed [Fx 55] - GREEN
- report: https://goo.gl/hJx1cN
Nightly
[Adrian]
- QA Photon work on all areas: status @ https://docs.google.com/spreadsheets/d/1ERWXWhTDgPi15py-rCLuYGO5kgp6MtrocxvSATjMIEE/edit#gid=0
- QA Photon: waiting for the TP sign-offs and preparing for the pre-Beta test runs and pre-Beta sign off
Aurora / Dev Edition
Beta
[Cornel]
- 56.0b9 signed off without any issues
- 64-bit migration testing completed on "beta-cdntest", "aurora-cdntest", "aurora" and "beta" - GREEN
- Pre-release sign-offs are currently underway (due Sep 8)
[Bogdan] [Photon Preferences Search] We are concerned about this feature since there still are several new bugs that affect both the usability and design of the feature. We are inclined not to let this feature reach 56 release only after fixing the newly found issues. Here is a list with all the new bugs that need fixing: https://goo.gl/GkQoRm [Julia]
- DevTools GoFaster - Per engineering, the feature is missing 57 due to blockers on the add-on startup code and the stabilization issues (see https://goo.gl/umcTfC )
Action items
Mobile
[Ninu]
- Tuesday - Signed off the 56.0b9 build with no major issues
- Today - Signed off the Nightly Pre-check part 2 session with no major issues
Performance
RelEng
Marketing/Communications
User Advocacy
Roundtable
- [julien] some users get 3 watersheds on 56.0bX, do we want/need to avoid that?
beta 3 - signing/compression beta 8 - to get the mig64 flag 64-bit - beta 9 to be migrated to 64-bits
- can we repackage beta 8 signing?
- required for 56.0 RCs https://bugzilla.mozilla.org/show_bug.cgi?id=1393789
- no QA resource available to test that
- did we turn automated update tests back on? - liz will follow up with florin
- planned for next week
System Add-ons
[Adrian] - not sure where this item should go
- Shield Opt-Out - QA mostly done, waiting for two fixes to land so we can smoke test it and sign it off