- Convergence milestone plan is incoming
- Quick note on priorities of bb+ bugs
- B2G Test Drivers program is going swimmingly well – thanks for your feedback, and keep it coming
- This call only: if you’re an engineer and have STR, please file a bug (template) instead of using the feedback app – saves a step
- We’re expecting to push another stable (dogfooding) build out today, with new fixes/regressions :)
- First FOTA (kernel) update test should be happening in the next couple of weeks
- 3 sets of weekly smoketests data on Otoro Master, Otoro Nightly, and Unagi Nightly. Weekly Test Results.
- Dogfood build received 2 updates already since its infancy. Reports are rolling in nicely through feedback and bugzilla.
- Team continues mapping of testcases to usecases in v1 FFOS tracking sheet. See areas done .
- up to 1500 testcases in MozTrap to date.
- Created a first set of test results from test execution in MozTrap. Exportable via PDF.
- Worked through the blocking+ list of bugs with engineering
- Wrapping up sprint for device storage, contacts, and alarm mochitests. Expect tracking bugs up by EOD next Tues.
- Still some breakage in mochitest. Running the sanity suite shows various tests erroring out of the run with focus restore issues. Don’t currently think this is the tests, so going to put them up for review anyway, but would block putting them on TBPL unless the current mochitest work fixes it.
- Still waiting on A-Team to get mochitests up on TBPL so we can expand framework.
- QA driving discussion on performance and stability for v1. Met with reps from slimfast, memshrink, platform, and gaia teams to brainstorm ideas.
- Issues / Risks:
- Packaged apps is entirely broken on trunk
- Still blocked for B2G on mochitest-chrome
- There are still blockers that are taking >2 days to resolve. Many of these are found in smoketests
- TEF is still on Otoro devices, and have yet to move to Unagi. We need to collaborate on one testing device.
- Video recording is still not done
- Lots of dupe bugs coming in from dogfooding. Working with Relmgmt on best process to cover these.
- Taking an alternate tactics w/ dchan and ptheriault to test permissions manager against desktop Gecko to get at least some early assurance. Once mochitest-chrome lands, will switch priorities back to tests running on B2G.
- Payments and Identity UX and user stories still isn’t complete
- App Install and Errors UX and user stories still isn’t complete
- Smoketest Blockers:
- Builds are still crashy and rebooting. just doing a smoketest, we’re averaging 5-6 crashes and 2-3 reboots last week
- bug 805967 – [OTA update] much of my user data blown away in today’s update
- bug 805966 – 10/26 daily build is hosed (boots only to black screen) – this blocked smoketesting on Friday.
- bug 805377 is now generating a crash whenever you select the video selection in the camera app.
- bug 805967 [OTA update] much of my user data blown away in today’s update
- bug 800263 – [FM Radio] Launch the FM Radio with headphone already plugged-in will cause screen freeze
- Functional Blockers:
- bug 796172 – [Homescreen] Takes ~6 seconds to appear after pressing the home button
- bug 800203 – [camera] Camera shows major graphical glitch right after taking a picture
- bug 796818 – [video] video is not showing up after recorded from camera
- bug 798636 – [Gallery] Crash upon finishing loading all of my pictures, including a picture just taken from the camera
- bug 796814 – [email] Email cannot launch if it does not think that the account does not have an inbox
- bug 796082 – [Music] Music player crashes when scrolling mix screen
- Blocking basecamp+ count is still high, but trending down…
- Non-blocker bugs need driver approval starting 11/1
- Blocker bugs need driver approval starting 12/1
- Prioritizing work: smoketest keyword
- Visibly transitioning away from feature work
- Etienne and Fabrice landed 3rd party app updates (and other 3rd party built-in app support)
- Margaret landed crash reporting UI
- FF17.0b3 went out last Wednesday, with 16.0.2 security fixes
- FF16.0.2 went out last Friday
- Final CTP block tests are going up today (can also toggle plugins.click_to_play), please also continue testing SocialAPI
- FF17.0b4 is going to build before EOD today
- We’ve only got 2 betas left! If you’ve received a nag email today, please make sure to comment with status on FF17+ bugs assigned to you.
- As a reminder, you can find our release calendar here and we now also have a release Twitter account
- Dao landed a few Snappy fixes:
- Ehsan, Josh and the folks from AppCoast are making good progress on front-end per-window private browsing support (tracking bug: bug fxPBnGen)
- Social is in the last weeks of beta, we’re working as hard as possible to get all the remaining blockers addressed
- Mike Conley is continuing to refine the new download manager panel (bug 746756, bug 766263, bug 801055)
- \o/ Progressive tile drawing is now enabled by default bug 795259, bug 803013
- Also turned off java screenshots with that
- Fix for bug 757680 Too few touch events dispatched for drawing application has landed (yay for parapara fans)
- Pref for disabling Automagic Update Downloads now works (“app.update.autodownload = never) bug 805431
- Project 256
- Expire unused tabs bug 792143
- Backspacing to delete words in text fields now behaves as expected thanks to JChen’s fix for bug 792928. Landed and uplifted to Aurora for Fx18. SUMO, Product, and good citizens who file bugs on-the-go with Fennec rejoice.
- Other happy IME news includes the freshly landed bug 712018 – International hardware keyboards input text as English qwerty (need implement IME support for hardware keyboard)
- bug 805162 – complete re-write of Android IME to make it thread safe has been reviewed, should be landing this week
- bug 802270 – Reuse a11y hover event for a11y focus event
- Stability Wins
- bug 795226 – java.lang.NullPointerException: at org.mozilla.gecko.GeckoInputConnection$2.run(GeckoInputConnection.java)
- bug 777460 – crash in nsNPAPIPluginInstance::CreateSharedHandle on Honeycomb and above
- bug 799977 – java.lang.NullPointerException: at org.mozilla.gecko.Tabs.loadUrlInTab(Tabs.java)
- bug 803289 – java.lang.NullPointerException: at org.mozilla.gecko.BrowserApp.onPrepareOptionsMenu(BrowserApp.java)
- Test builds on the new Razr i (Android x86) are very promising. Chris Lord reports responsivness and performance are very good.
- WinRT widget layer backport to VS2010 is mostly complete. Should land on Elm this week (bug 794983). This will remove the dependency on VS2012 and make it possible to build Metro code without breaking compatibility with Windows XP.
- Continuing elm to mozilla-central merge. Some big blockers like the platform splitting work (bug 755724) that need to get finished up.
- Next: Migrate Windows builders to the Windows 8.0 SDK. Initial testing work started on this, see bug 774910. We’ll be migrating custom elm builders over to VS2010/8.0 next. This is a big change, QA will need to be involved in testing before we can switch. Releng is currently in the drivers seat. Goal is to get this done by year end so we can turn on metro browser bits for mc nightlies.
- Product and UX are working on finalizing user stories/requirements for the initial launch.
- First draft of MVP Feature list available and entered into Bugzilla, queries available at Metro Wiki Page
Stability Report [Weekly]
- Flash – spike in crashes being tracked in bug 805021. Adobe advises in one of the most recent comments that they have determined the root cause of the issue and will push a server side fix. Should see dropoff in 24 hours.
- bug 801516 crash in js::gc::IsObjectMarked with AdBlock Plus
- Trending down in the past few days
- bug 777460 needs nom for aurora/beta
- bug 760392 seems to be fixed? Fix from bug 761706 might need to be pushed to beta?
- bug 803289 seems to be fixed, only affects nightly.
The following have traction, just need implementation which are being worked on:
- still need ui; some bugs have been nom’ed blockers
- still need ADU…
- still need symbols…
- Jared Wein landed a fix to prevent the awesomebar popup from opening and closing repeatedly while typing in the location bar. (bug 804968)
- Dão Gottwald removed some expensive layout flushes when switching tabs if the user isn’t overflowing their tab bar (bug 752376)
- Dão also changed some initialization code to run after Firefox is drawn resulting in ~10% decrease in perceived start-up time in certain cases (bug 715402)
- Timothy Nikkel fixed Flash so that when run in a background tab it no longer degrades rendering perf (bug 642257)
- Vladan Djeric landed a probe to measure various DOM Local Storage overhead (bug 802920) – view telemetry data
|0.0||0.0 days||0.0 days||90 bugs (92)|| 54 bugs(78)|
|Patch Review Time||SG:Crit Resolution Time||Top Crash Resolution Time||Unresolved Aurora Trackers (non-security, not tracked for Beta)|| Unresolved Beta Trackers (non-security)|
- Please report bugs to https://metrics.mozilla.com/projects/browse/BA login with your LDAP to create an account.
- High confidence queries
- Burn up
- Burn down
- All: http://people.mozilla.org/~klahnakoski/
- Programs are defined in http://people.mozilla.com/~klahnakoski/es/js/MozillaPrograms.js
- How reviews end
- Timezone: https://metrics.mozilla.com/projects/browse/BA-54
- No reviews dropping
- Warning vs Prohibited – use cases?
Key Actions & Issues
- jet to staff social performance profiling, layout-side
- gavin to staff social performance profiling, session-restore side
- dmandelin to find owners for 2 JS top crashers
- jet/jkew to take the text, layout top crashers
- Reorganizing Test Directories (Henrik)
- Backout Policy for Ts Regressions (Ehsan)
- (catlee) Downtime notice for Saturday November 3, 9:30am-4:00pm PT 
- HG will be offline for maintenance bug 805851
- all trees will be closed
- (catlee) releng unagi builds on TBPL as of today. these are per-checkin. we’re working on nightly build support w/ updates and signed mars.
- (jst) make “WebIDLFile”-example is your friend if you’re implementing WebIDL interfaces.
- (johnath) need layout owner on the social+flash bustage on nightly/aurora now that we have a regression range and suspect
Platform Meeting Details