- 1 Notices / Schedule
- 2 Firefox Development
- 3 Firefox Developer Tools
- 4 Performance
- 5 GFX
- 6 JS
- 7 Layout
- 8 Video
- 9 DOM
- 10 WebAPI
- 11 Network
- 12 Plugins
- 13 Mobile
- 14 Accessibility
- 15 Multi-Process
- 16 Tree Management
- 17 Security
- 18 Stability Report
- 19 Roundtable
Notices / Schedule
- We shipped firefox 6 last week (and yes, we received a cake)
- Migrations went according to plan on beta and aurora, and updates are now out for those users
- Legnitto out this week - poke johnath if needed
- Discussion in rel-drivers about 3.6.21 schedule - should settle down this week (trying to manage the overlap of 3.6.x, FF7 release, moco work week and QA team work week)
- Dao Gottwald and Tim Taubert have been doing excellent work chasing down DOMWindow leaks in bug 658738 (and dependencies)
Firefox Developer Tools
- Orion editor landed for use in Scratchpad (and future tools), disabled on Aurora.
- Enabled on mozilla-central to get feedback on issues raised in the dev-apps-firefox thread.
- Weekly meetings have been moved to Thursdays at 10am.
- Support for antialiasing in WebGL may land soon! bug 615976
- jsdbg2 bug 672829 has landed.
- Congrats to Jason Orendorff and Jim Blandy!
- This is the new debugging API for JS.
- It will make debuggers more robust and enable new capabilities (e.g., eventually, remote debugging).
- Brian Hackett's type inference branch is landing soon (1-2 weeks).
- Looking pretty good, so not too worried
- We do have a backup plan in place: Brian will be maintaining a TI-free branch on the side.
- JS runtimes (class JSRuntime) are going to become single-threaded soon.
- Luke will add assertions that fire if you use a runtime in a multithreaded manner.
- There are still a few MT usages left in the browser (e.g., bug 679036 ), so we'll turn on those asserts only when they are cleared away.
- Mats Palmgren has a patch about to land that changes the way we iterate through frame child lists to a more efficient and maintainable approach. This may cause some bit-rot in pending layout patches.
- Mats landed support for text-overflow:<left> <right> (different markers for the left and right sides of an overflowing block).
- Josh Aas has a patch to move ownership of plugins from frames to content to fix bugs like "display:none plugins don't work". This is getting close to done. We'll need to monitor its progress carefully because it's a scary, regression-prone area. The good news is that another >10 year old architectural botch is going away!
- New Auckland office rocks!
- Paul Adenot (intern) implemented "played" and "seekable" media element attributes.
- Event properties (foo.onclick etc) changing to be defined in IDL bug 659350 (bz)
- Mounir is working on WebSMS. Is able to send SMS messages and get notified when they are sent, as well as get notifications about incoming SMS messages. Will link to draft of API once we have it. Patches in bug 674725
- Kevin Gadd is working on WebUSB on Windows. We can only get full support on Windows 7 for now (and possibly forever) due to platform limitations in driver handling.
- There's a first draft of API for WebTelephony
- Lots of discussions about security model
- Camera work is happening in other teams
- [Patrick McManus] SPDY implementation is progressing nicely.
- [Nick Hurley] Continuing to expand NeckoNet capabilities. New version coming soon (with a ready-to-use Linux VM) which will allow testing from Linux, Windows, and Mac OS X.
- Fix for content ownership of plugins is in review. All tests pass, no known problems. Hopefully it can land within a week.
- There is a new "plugin hang" crash where the Firefox (nightly) process stops responding and the plugin process therefore commits suicide. There are some false positives being investigated.
- Everyone should read "Android/ARM Firefox Performance: where we are and where to go"
- A todo list of things to improve on Android
- Everyone should turn on Telemetry in Mobile
- WebApps JS API needs bug 641552 - Provide hook for adding APIs to window.navigator
- Electrolysis/Short-Term Goals work continues apace
- Update on Tinderbox mail delay
- Maximum delay we are seeing has been cut in half (~8 hours -> ~4h), and spans a shorter period of time.
- IT improved performance on the Tinderbox server in bug 676822 - no longer I/O bound
- Turning off mail for the Try branch cut down load by a lot. bug 679589
- New TBPL is still in the process of being put into production, now taken over by Web Dev bug 669000, no ETA at this time.
- Network issues last week caused a bunch of Windows machines to get into a hung state, causing long wait times for some jobs.
- Now fixed, wait times back to normal.
- Improved monitoring tracked in bug 681111
Reviews this week
- Wed. 24-Aug, 1:00 PM PDT
- 3O - ORLY
- Thur. 25-Aug, 15:00 PDT
- 3P - PB&J
Intern Project for DNSSec
- Fri. 26-Aug 1:00 PM PDT
- 3P - PBJ
SecReview: (double review) Web Activities & F1/Share
- ) Web Activities
- component of the broader Open Web Applications project
- ) F1 retooled version of the link sharing service we looked at in May (https://wiki.mozilla.org/Security/Reviews/F1)
Common Meeting Information
- Introduce Feature (5-10 minutes) [can be answered ahead of time to save meeting time]
- Goal of Feature, what is trying to be achieved (problem solved, use cases, etc)
- What solutions/approaches were considered other than the proposed solution?
- Why was this solution chosen?
- Any security threats already considered in the design and why?
- Threat Brainstorming (30-40 minutes)
- Conclusions / Action Items (10-20 minutes)
- IRC Channel: #security
- Etherpad: http://etherpad.mozilla.com:9000/secreview
- Dial-in Info:
- In office or soft phone: extension 92
- US/INTL: 650-903-0800 or 650-215-1282 then extension 92
- Toronto: 416-848-3114 then extension 92
- Toll-free: 800-707-2533 then password 369
- Conference num 624
- 6.0 data so far - the crash rate tracking very close to that of 5.0
- Next release is 2.3. No ETA but we are working with the Socorro team to clarify our priorities and see what can get in.
- Bug 680862 was a rather high volume regression yesterday and was fixed by backout. Comment 28 in the bug indicates it was visible in tests, but because the color was purple instead of red it was ignored.
- Bug 681385 - Plugin Trunk regression that started with 20110817 build.
Aurora & Beta
- https://bugzilla.mozilla.org/show_bug.cgi?id=680954 Bug 680954] - Mac only, happens across all versions
- http://test.kairo.at/socorro/2011-08-22.firefox.7.0.flashhangdetails.html - Data for 7.0b1 on Flash issues.
- Malware/Extension Related
- A few other things we are tracking...
- Looking at options (if any) for getting plugin URL correlations.
- bug 671960 seems to be resolved in 6; was close to 40 % of crashes in 5
- biggest bug in 6 and 7 seems to be bug 663494
- still a lot of crashes with no signatures in 8, 7, 6, 5
- Fennec 9a1 added; at least 81.2 % of crashers are new (out of 186 crashes); Fennec 8 has a lot of new crashers as well
- 9 and 8 seem to be crashing due to graphic related issues
- Announcing mozilla-history, a git repo containing *all* of our code history (ehsan)
- Switch the Geolocation provider end point to their new service. If you see anything funny, file bugs (dougt)