Platform/2009-05-19

From MozillaWiki

« previous week | index | next week »

Notices / Schedule

Firefox 3.0.11

• Code frozen!
• Builds have started, nay, Mac and Linux builds are available for QA

Firefox 3.0.12

• Schedule on the wiki

Firefox 3.5 RC1

• goal is to code freeze tomorrow
• build can produce RC builds in 24-36 hours thanks to timezone fun
• QA would like 7 full business days for RC tests

Blocker Report

• 16 blocker nominations (-12 w/w)
• 53 blockers left without a fix (-13 w/w, 40 are code related)

Since 1.9.1 branch …

Past 2 weeks …

[ Platform Blocker Queries | Front End Blocker Queries ]

Summary

• great pace, though still adding a lot of unresolved blockers
• nominations need constant vigilance; should be doing it every day
• thanks for landing fixes on branch – let’s get the rest!

The Breakdown

• 115 blockers marked fixed1.9.1 over last week (+lots w/w, yay landings!)
• 14 bugs nominated for blocking over last week (-9 w/w)
  • blocking-firefox3.5+: 6 new unresolved (-1 w/w), 3 new resolved blocking-firefox3.5+ added (+1 w/w)
  • blocking1.9.1+: 18 new unresolved (+6 w/w), 41 new resolved blocking1.9.1+ added (+14 w/w)
• 57 patches approved for 1.9.1 over last week (+18 w/w)
• 52 blockers are FIXED but not yet landed on 1.9.1 (-48 w/w)
• 41 bugs with approval1.9.1+ that are not yet landed on 1.9.1 (+11 w/w)

Browser / Front End

• Blockers: 15 remaining & 0 nominations
  • should be tight, but we can make tomorrow
  • only one bug that’s unscoped, which is bug 490035 – might depend on bug 492410 about slow script warnings?
  • our other areas of interest are process elevation on Vista installer, TSnappiness bugs, easy polish fixes

• Polish update: Firefox is 56% shiny (+1% change)
  • Thanks go to: matthew, mossop, dao (2)
  • 43 remaining easy polish bugs (whiteboard [polish-easy])
  • 39 remaining hard polish bugs(whiteboard [polish-hard])

GFX 1.9.1 Update

• 0 blockers open, 0 blockers remaining to be landed on 1.9.1
  • GFX mostly code-complete for 1.9.1
  • One windows 7 bug may be landed on 1.9.1 yet, but the release doesn’t block on it.
• 1 nomination
  • security bug – might block, awaiting decision.

Layout Update

• 0 layout blockers (2 noms)
  • bug 488649 should probably block, need to get a simple patch
  • bug 493189 should probably not block
• 0 SVG blockers
• 4 Video blockers
  • bug 486899 unsure how to proceed
  • bug 487519 number of crasher bugs in libvorbis and other libs, doublec working with upstream, probably should not block
  • bug 493109 patch backed out, should block, roc on it
  • bug 493678 probably should block, making progress
  • Lots fixed over last week: playback perf, A/V sync, many edge cases fixed, playing back strange content, some random failures
  • One more thing: fix for seeking artifacts
  • Very good shape
  • Quite a lot of merging to branch to do

Content Update

• Content Blockers
• Not disabling XPCOM plugins until Java SE 6 Update 14 is released. That is the first release that has no XPCOM dependencies whatsoever, though the plugin has been NPAPI (not XPCOM) since Java SE 6 Update 10.

Mac OS X Update

• Steven Michaud released a JEP update for the top Mac OS X crash in Firefox 3.5 Beta 4.
• Josh looking into consuming Objective-C exceptions vs. aborting, at least in some cases.
• 64-bit Mac OS X builds coming along, builds up to gfx without patches and builds through gfx with patches in review. Next module that needs to be ported is plugins, will take longer than other components since a new API implementation is required (Cocoa event model).
• Trunk and all branches of Firefox run without any major issues on the latest Mac OS X 10.6 seed. Can successfully compile 32-bit Firefox via cross-compile on 10.6 if you disable crashreporter (default architecture on 10.6 is x86_64 via gcc-4.2).
• Josh at WWDC in SF June 8-12.

JS 1.9.1

• 9-10 JS Blockers

General 1.9.1

These are bugs that fall outside of components covered by the Gfx, Content, Layout and JS groups:

• blockers
• untriaged nominations

Mobile 1.9.1 Update

Security

• Security Reviews

Multi-Process

• bent working on chromium library in his user repo.
  • base and chrome code exported from chromium svn, integrated into build system on windows (libxul required…)
  • glue library started to provide libxul and appshell integration
  • child process shell built, uses xpcom embedding apis
  • test-ish parent app built using chromium’s base library, still being tweaked, but currently launches child process and waits around for something to do

Tree Management

• Splitting out unittests to run concurrently
  • running along with “traditional unittests” while we debug different results and geolocation test bustage
    • http://tinderbox.mozilla.org/showbuilds.cgi?tree=Firefox-Unittest
    • http://tinderbox.mozilla.org/showbuilds.cgi?tree=Firefox3.5-Unittest
    • http://tinderbox.mozilla.org/showbuilds.cgi?tree=TraceMonkey-Unittest
  • details in catlee’s blog and in bug 383136

Roundtable

• Blog funnel starting up: Evangelism/Firefox3.5/35Days – will ask for help/writing
• New Firefox icon:
  • Design work is all public: http://blog.mozilla.org/faaborg/
  • It is ok to provide feedback (we won’t think you are bikeshedding)
• How many people are running Windows 7?
  • you can get the RC for free
  • it runs in VMWare, Parallels and VirtualBox as if it were Linux
• QA will concentrate testing in the following areas during RC1 testing:
  • Video performance and Controls, Geolocation, Session Restore, Tabbed Browsing, Drag and drop functionality, Clear history/ private browsing/ forget this site, topsites layout and performance, windows 7 adhoc testing. Please suggest any other broad regression areas here.