Platform/2009-10-06

From MozillaWiki

« previous week | index | next week »

Notices / Schedule

Firefox 3.0.15 / Firefox 3.5.4

• Firefox 3.0.15: builds started
• Firefox 3.5.4: need one more fix, then we’ll start builds
• Should reopen branches on Thursday or so for 3.0.16 and 3.5.5

Firefox 3.6b1

• code freeze was supposed to be last night, now moved to Tuesday Oct 6 at 11:59pm PDT
• 31 P1 blockers showing as unresolved on 1.9.2
  • only 10 of those do not have fixes on some branch (either mozilla-central or tracemonkey)
  • the rest need to land on 1.9.2 immediately:
• all feature landings needed before b1 cutoff are complete:
  • lightweight themes (done)
  • about:support (done)
  • js c types (done)
  • plugin update referrals (done)

Blocker Report

See more Firefox 3.6 related blocker queries, or learn about the new status and blocker flags

• Release Blockers (flag: blocking1.9.2 or blocking-firefox3.6)
  • 124 OPEN (-19 w/w)
  • 37 FIXED but not yet fixed on mozilla-1.9.2 (-39 w/w)
  • 35 nominations (+18 w/w)
  • Handy charts: Blocker snapshots, Blocker and Noms trends
• Approvals
  • 74 requests (+2 w/w)
  • 37 approved but not yet fixed on mozilla-1.9.2 (+6 w/w)
  • Charts: Nominations snapshots

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:
  • plugin update referrals landed on branch and trunk
  • work to reduce Places overhead landed on branch and trunk
  • about:memory (see bug 515354) is approved to land on branch
  • personas now works with www.getpersonas.com

• Firefox 3.6 b1 blockers:
  • bug 474060 needs to land

GFX Update

• 0 beta blockers; 4 1.9.2 blockers
• Jeff landed a huge QCMS perf win from Steve Snyder in bug 512865. He’s got a blog post upcoming.
• John got a big Mac Ts win from deferred font initialization in bug 517045.
• Bas Schouten is coming on board, and has already got some great work done on OpenGL accelerated full-screen video. bug 510057
• Lots of great news w.r.t. Windows 7 features – see Platform-specific features section below.

Layout Update

• Blockers
  • 4 P1 blockers, all fixed on trunk (still analyzing regression from bug 509693, but can demote it to P2)
  • 24 P2 blockers, at least 3 with patches
  • low nomination rate
• Landed animation of CSS properties with SMIL/SVG Animation (dholbert)
• Landed doubly-linked frame lists (bzbarsky/mats)
• WOFF landed for 3.6 (jfkthame/jdaggett)
• Getting closer with Ogg decoder rewrite (doublec) — seeks, plays chained Oggs
• Depending on how we perform on N900, may need new video/audio goals (or redeploy cpearce/kinetik)

Content Update

• 1.9.2 beta1
  • Working to get the last two remaining P1 blockers in today (bug 504862 and bug 517355)
• Got the JetPack security changes landed (mrbkap)
• Working on top-crashers

Platform-specific Support Update

• Win7 sdk rollout complete. (hat tip bhearsum!)
• Win7 taskbar preview patches (widget/browser) landed on trunk, 1.9.2 last night.
• Win7 progress patches (widget/browser) have final reviews, still need to land.
• Win7 jump list widget patch has final reviews, browser patch needs an sr.

JS

• recursion landed for simple cases

Strange perf hit on a few mac (perf win on windows/linux) talos tests, investigating.

Startup Performance

Highlights from the weekly update post, and work started this week:

• John Dagget’s ~20% cold startup improvement and ~4% warm startup improvement on Mac in bug 517045 landed on branch.
• Taras Glek’s work on combining the different preference files has landed (bug 507288)
• Alfred Kayser’s patch to increase throughput of JAR file reading has landed (bug 510844).
• Ryan landed bug 475289 to lazily initialize the engineMetadataService for an expected 20ms win.
• Ben Hsieh got some stat() removal landed in bug 510991. He’s got even more in bug 511761, just need bsmedberg review.
• I’m no longer able to reproduce the Windows cold startup regression from 3.5 to 3.6 reported in bug 517741. I turned off SuperFetch, which resulted in more stable numbers, but there was no visible regression, even in builds going back a couple of months. Juan tested on Windows XP and found the same. The current numbers show a 27-30% *improvement* in cold startup time on trunk, and up to 8% on 3.6.

Join us on IRC in #startup.

Security

Nothing to report.

Electrolysis

• Windowless plugins animating correctly (implement NPN_Invalidate*)
  • karlt beginning to design asynchronous painting
• Remote canvas.drawWindow in final review. Yay!
  • Now we need to get event handling back up to speed
• NPRuntime in final review
• bent working on Windows deadlocks

Tree Management

• We need a quiet time to land split-mochitests so we can iron out inter-test dependencies. Post-beta?

• Win7 SDK is deployed!

• Migrating 8 Tiger Talos machines to run Linux / Vista on Try Talos

• Turning off idle builds. We currently run a full set of builds and tests if nothing on a branch has changed in several hours. This ties up resources required for other branches.

• 1,774 landings last month, 37% jump over previous recorded highest
  • year to date range: 1,018 to 1,295
  • details here

Roundtable

Firebug

Four bugs outstanding right now for Firebug.

• bug 520572 – funky window property js debugger bug
• bug 500128 – long js files have incorrect line numbers
• bug 520421 – need regression range
• bug 515051 – Stream listener registered in a network request channel eats JS error messages