Platform/2009-08-11

« previous week | index | next week »

Notices / Schedule

Firefox 3.0.14

  • code freeze today
  • 3 blockers left that need 1.9.0 fixes
  • (one of those we'll take past code freeze for sure)

Firefox 3.5.3

  • code freeze today
  • 1 blocker left that needs a 1.9.1 fix
  • might push one day to pick up some non-blockers
  • mostly a 'sync up' release with 3.0.x

Firefox 3.6a1

  • shipped on Friday
  • not a ton of traction yet (more users on old trunk nightlies)

Blocker Report

Gecko 1.9.2 / Namoroka

  • Release Blockers (flag: blocking1.9.2 or blocking-firefox3.6)

Browser / Front End

  • see our active projects and get involved / propose others
    • multitouch JS API prototype work
    • streaming video controls
    • new UI mechanism for application notifications ("doorhangers")
  • triage to be finished this week
  • feature priorities are:
    • asynchronous location bar (landed)
    • awesomeform complete (landed)
    • potential uplift of personas (under evaluation)
    • responsiveness increases

GFX Update

  • bug 507970 - New font file formats for the web: EOT Lite and ZOT - Support in development.
  • Jeff & Joe were in New Orleans for SIGGRAPH 2009. Blog posts about interesting stuff upcoming.
  • Decode-on-draw work ongoing in bug 435296.

Layout Update

  • WebOTF (jdaggett, jfkthame)
  • roc away next week

Content Update

  • infallible malloc ready to land bug 441324
  • HTML5 work moving along (hsivonen has the details).
  • multi process network implementation starting to materialize.
  • XBL2 implementation moving along (jonas has details).
  • peterv found some problems with XPConnect perf bug 499199 and has been working on ways around that.
  • peterv got ibench up n' running and will be looking into why don't perform better than we do.
  • mrbkap working with labs to iron out a security model for JetPacks.
  • HTML5 History API progress
  • File drag'n'drop implemented. File API progress

Platform-specific Support Update

  • Final Windows 7 SDK came out on Friday, testing in progress
  • Windows 7 integration work continues

JS

  • ES5 Status (tracking bug bug 445494):
    • Meta-object API:
    • Strict mode:
    • Regexp:
      • bug 98409 (waldo): literal regexps and lastIndex.
      • bug 465199 (sayrer): Regexp lastIndex property type.
      • bug 501739 (unassigned): string match, replace vs. global lastIndex.
    • Misc:
      • bug 430930 (bruce): Date.parse and ISO 8601.
      • bug 496923 (unassigned): Integrate ECMAScript 5 conformance test suite
      • bug 509733 (unassigned): ES5 chimera arguments object
  • DWARF Call Frame Info
    • needed for breakpad stacks with -fomit-frame-pointer optimization
    • Linux dumper re-architect is landed in Google's SVN repository
    • DWARF-on-Linux patch queue is complete, seems to work, needs more testing
    • Next step is to add CFI support
    • DWARF reader and processor shared with Mac, so we should get CFI on Mac for free or cheap

Security

Electrolysis

  • plugin work (NPObject) ongoing...
  • need help with windowless plugins
  • need lots of help with Mac
  • tab work is going slowly... bz will be able to return to that during/after the workweek

Tree Management

  • Branching for mozilla-1.9.2 bug 504143
    • Planning to bring up all normal Firefox builds (linux, linux 64-bit, mac, win32) + l10n, maemo/winmo + l10n. WinCE to come with or shortly after bug 485109. Anything else?
    • Will be cloning mozilla-central and l10n repositories today or tomorrow, please don't touch them yet.
    • When are we doing this? Before or after Mobile "go to build"?
      • impact of possible Fennec respins
      • No downtime is needed.
    • Will show up on http://tinderbox.mozilla.org/showbuilds.cgi?tree=Firefox3.6
  • Fun and games with colo outage Sunday; fast recovery
  • Now that tp4 rolled out in production, planning to:
    • turn off fast-talos machines bug 509744
    • stop running tp3
  • Last batch of new slaves added late last week; will monitor waittimes to see about adding more

Roundtable

  • Update on Code Coverage (timr)
    • Latest results: c/c++ and JS
    • We had a good discussions with the Content and JS teams last week. Some issues we are looking into is specific files with low coverage, how coverage for in-lined code is handled, and annotating the coverage results pages with info like: a file was recently re-designed and therefore previous bug data may no longer be valid.
    • Also see the new FAQ and new Code Cov one stop shopping wiki
    • Other next steps: Identify with the Dev teams which files should be targeted for additional testing. Work with other teams like the Firefox team to understand how this info could be used. Feel free to contact Murali or myself if you want to dig into the data with us.
  • Content intern project progress