From MozillaWiki
Jump to: navigation, search

« previous week | index | next week »

Notices / Schedule

Firefox 4 beta 8

  • Waiting on sync changes
  • Driving blockers to zero, hopefully will converge in the next couple days
  • QA needs at least a week. If we build Thursday we might not be comfortable releasing by the following Thursday

Firefox 3.6.13 and 3.5.16

  • Created build #3 of 3.6.13, released it to beta on Monday
  • Still plan to release on this Thursday, December 9th

Blocker Report

Firefox 4 Beta

Blocker report for Firefox 4.0.beta8 (as of 2010-12-07 11:07:47.826151-08:00)


      blocking: 193
         fixed: 181
          open:  12

Open Blocker Assignees

   Philipp von Weitershausen [:philikon]:   5 +++++
                         Patrick McManus:   1 +
                  Blake Kaplan (:mrbkap):   1 +
              Michael Morgan [:morgamic]:   1 +
                        Myk Melez [:myk]:   1 +
                      Axel Hecht [:Pike]:   1 +
    Nobody; OK to take it and work on it:   1 +
                      ben turner [:bent]:   1 +

Outstanding Review Requests

                  1 +
              1 +
            1 +

        Total outstanding flagged review:   3

Firefox Development

(from our goals):


  • [ON TRACK] API-Complete Jetpack
    • Joint with Jetpack team
    • Currently working towards feature-complete beta SDK
  • [ON TRACK] Stand up basic Firefox with Electrolysis

Developer Tools

  • [NEW] Ensure that Firefox 4 ships with a compatible version of Firebug
  • [NEW] Publish a roadmap for integrated developer tools in Firefox, including console, inspection, and js debugging
  • [NEW] Develop long term plan for Skywriter maintenance/support


(there is a team-by-team goals breakdown, as well)

  • [MISSED] Finish Firefox 4 blockers.
  • [DONE] Support Firefox Team to stand up a browser on Electrolysis.
  • [MISSED] Ship Firefox 4 final with layers composition accelerated using:
    • OpenGL on Mac OS X;
      • This is turned on by default for all supported hardware, and there were few blocking bugs by the end of the year.
    • Direct3D on Windows.
      • This is turned on by default for all supported hardware, and there were few blocking bugs by the end of the year.
  • [MISSED] Ship Firefox 4 final with content accelerated using Direct2D on Windows Vista/7.
      • This is turned on by default for all supported hardware, and there were few blocking bugs by the end of the year.
  • [MISSED] Ship Firefox 4 final with better Opentype font feature support using Harfbuzz on all tier-1 platforms, and text rendering performance equal to or better than Firefox 3.6.
    • We landed Linux Harfbuzz support before the end of the year, and there are no known text rendering regressions.
  • [MISSED] Ship Firefox 4 final with full WebGL 1.0 support turned on by default on all platforms that support OpenGL.
    • We missed this goal because our WebGL implementation didn't pass all parts of the WebGL testsuite, which is required for "full WebGL 1.0 support." However, it is turned on by default for all supported hardware.
  • [DONE] Integrate Opentype santizer library
  • [MISSED] Play back video on Fennec without blocking on main content thread
  • [DONE] Integrate codec ARM optimizations
  • [MISSED] Ship indexedDB - indexedDB is done and landed, has been for some time. Minor bugfixing ongoing, but nothing beyond the norm.
  • [MISSED] Ship HTML5 parser - Done, just didn't have anything to ship it in.
  • [MISSED] Ship HTML5 forms (current set) - Done, just didn't have anything to ship it in.
  • [MISSED] Ship GC compartments work - Done (though some bugfixing and GC work remains), just didn't have anything to ship it in.
  • [MISSED] Proof of concept implementation of a proxy based DOM binding - Plan in place, but no work done here yet.
  • [MISSED] 30% performance improvment on the Cal_* tests in Zimbra
  • [DONE] Performance 15% improved over Oct 1 Sunspider perfomance

Improved over 30%.

  • [DONE] Performance 15% improved over Oct 1 v8 performance

Improved over 30%.

  • [MISSED] Continuous ARMv7 Android and Maemo JS Shell regression and performance testing in place

Have much better ARM testing than we used to, but not quit the JS shell regression tests we were hoping for.

  • [MISSED] Drive blocker list to zero. [This was a super-stretch goal]
  • [DONE] Develop implementation plan for OOP content accessibility. [Note it may change once we ramp up again post-FF4]
  • [DONE] Fix and ship multi-process plugins on Mac
  • [MISSED] Fix and ship asynchronous plugin painting on all plaforms
  • [DONE] Fix and ship content processes on Fennec


  • 119 betaN, final, and Fennec blockers (-4 w/w)
  • OpenGL composited layers on Fennec update (jrmuizel)
    • Can be turned on (layers.accelerate-all pref) on trunk.
    • Known bug: Changing orientation of device, or resuming fennec, breaks all drawing.


  • Blocker count slowly dropping, now 103
    • Many bug fixes in hand


  • Major remainders from compartment landing fixed (beta8 blockers). Many thanks to mrbkap and gal.

Tree Management

  • Fallout from our fun last week hunting down new orange caused by DNS:
    • Friendly reminder that the RelEng Sheriff can re-trigger tests or even builds of specific changesets for you.
    • Cutting off external network access to test machines. See dev.tree-management thread here
  • IT needs to rearrange test minis in the new colo. Proposed dates are December 17th, December 23rd. This will require a full working day with little or no test coverage. bug 616658
  • Win7 debug unit tests enabled
    • last known perma-orange for xpcshell - bug 610654
    • who can make xpcshell run faster on Win7? [1] It runs pretty slow. The other test suites did not get affected (except reftest a little bit)
      • xpcshell takes quite long on win7 and increases the end-to-end times
      • shall we keep on running xpcshell only on win2k3? shall we proceed with the switch? who can take ownership of making xpcshell faster on Windows 7?
      • we would like to stop running unit tests on win2k3 a couple of days after the last perma-orange on win7 gets fixed but this depends on answering the previous question.


  • Pulling WebSockets for Firefox 4 due to a security problem with proxy upgrade negotiation. (inherent problem with standard, spec in flux)
  • String freeze: [blocking with strings] has 10 bugs, 6 prelanded strings, 4 new strings. I'm losing track, bug 616015 and bug 616014 were filed without/wrong whiteboard and blocking+. (Axel) [blocker noms with strings]
    • file bugs with [strings] in the whiteboard, request blocking, and CC :Pike. Johnath is to triage those.
  • Blocker re-triage next week.
  • Branching for Fennec4.0b3, Firefox4.0b8, and what branches to use so that post beta3/beta8 fixes can land, to be revisited later this week. (discussion in meeting with legneato, stuart, joduinn, catlee).