Notices / Schedule
- in QA, on track for mid-December release
- no schedule for either release yet, should have one by the end of the week
- tree should open in early December
Firefox 3.6 Beta
- beta user base is now over 450,000
- beta 3 refresh shipped last week
- component lockdown effects?
Firefox 3.6 RC
- late on RC freeze
- Release Blockers (flag: blocking1.9.2 or blocking-firefox3.6)
Browser / Front End
(Progress reports every weekend on Planet Firefox)
- see our active projects and get involved / propose others
- Namoroka/mozilla-1.9.2 front end development:
- 8 blockers left, 4 are crashkill
- bug 520535 needs a patch
- the other three need review (Neil Rashbrook is sick, Neil Deakin is out)
- should be done in next 24 hours
- keeping nomination list down with constant review, so far no major concerns other than late breaking Toolkit dependencies coming in from Fennec
- 2 blockers; one is waiting on review, the other is a mobile blocker (kerning) that has not seen traction yet.
- 2 noms; 1 might be build config related and if so should not block, the other has a reviewed patch and probably shouldn't block, but should be approved.
- Jonathan Kew's font enumeration Ts patch is looking good, and has passed review from John Daggett.
- And now, for the Direct2D in Firefox show, starring Bas Schouten!
- Bas's blog post on Direct2D was featured in a cnet news article on Direct2D in Firefox.
- Direct2D provides vector graphics rendering on the GPU. It is implemented as a backend to cairo.
- Large performance improvements in scrolling.
- Large improvements in dynamic websites using a lot of transformations and blending operations.
- Possible implementation strategies:
- Switched on for all D3D10+ cards, blacklist buggy hardware.
- Switch off by default, whitelist hardware/driver versions.
- Possibly additionally whitelist D3D9 cards.
- bug 470440 Going to implement sub-pixel positioned text rendering on Linux, hoping to have working code by early next week
- Blocker report
- Several blockers added and fixed in the last week
- 1 nomination (arrived 30 minutes ago)
- 3 blockers with patches
- 1 crashkill (not a hard blocker IMHO)
- Still evaluating crashes revealed by frame poisoning
- Largest-volume crash fixed, probably fixes for one or two others
- Most have no steps to reproduce, some seem inexplicable
- Need a judgment call on remaining volume
- Video on N900
- Theorarm decodes 100fps on "reasonably sized video", beating current DSP decoder (on speed if not power)
- Plan on using GL for YUV decoding, not this quarter
- Experiment with standalone player
- Pulseaudio takes 10-15% of CPU!
- Instructions for setting up record and replay
- Any particular test we should look at?
- Need Windows loopback interface sniffing
- One blocker left (security bug)
- Ben Newman got JPW talking across processes.
- Jonas Sicking went through and caught up with the spec and cleaned up some minor problems in our new DOM file API additions.
- Continuing with crashkill work
- HTML5 parser update (hsivonen)
Platform-specific Support Update
- Some fixed on TM tree, one test failure after m-c merge holding things up
- Merging to 1.9.2 ongoing
- Many new fuzzbug / QA bugs recently
- JS team will meet to discuss each JS blocker at 2pm PST
- Read the latest weekly update.
- View snapshot results for startup and all other tests.
- bug 517804 was a big win for Mac warm startup, putting 3.6 at 35% faster than 3.5.
- Joel has some data on his super-static Firefox in bug 525013. Inconclusive so far.
- Ben is making progress on the fastload replacement in bug 520309. Some fastload/invalidation discussion in this dev.platform thread.
- Taras has patches up for service caching (bug 516085
- Super-fast-path-ing of Components.* needs some JS team help: (bug 512584).
- Looks like a final patch on bug 519445, for reducing Mac startup time spent in font system initialization.
- Ryan Flint put a patch to minify JS on bug 524858, needs work still, but significantly reduces the size of shipped JS files.
|Windows TSF integration (1.9.2)||unscheduled||Jim Chen, roc|
|DNS Prefetching||unscheduled||Patrick McManus||Jesse, bz, reed(?), ctalbert|
|New system metrics (and media queries)||unscheduled||?||dbaron|
Session Cookie Policies
- Lucas filed bug 530594 to propose that we forcibly expire session cookies after 7 days; we need feedback from people here and from web developers
- also wondering if we should have an explicit option to always expire session cookies whenever the browser quits, even if the user chooses to save tabs or if there is a restart for add-on or application install/update.
- Related bugs:
- Try server repo needs resetting. Will be down for around 1 hour. Tomorrow night or next tuesday.
- b4 / rc1 scheduling
- we think we've fixed a sigificant number of the 3.6b3 topcrashes (see Sam's nice table) (dbaron)