« previous week | index | next week »

Notices / Schedule

• New working Firefox 3.6.4 schedule
• New working Firefox 3.5.10 schedule
• Code complete for 3.6.4 and 3.5.10 is TODAY, April 27 2010 @ 11:59 pm PST
• Schedule risk to 3.5.10 if 3.6.4 requires additional resources

Blocker Report

Firefox 3.6.4

• 1 open blocker for 3.6.4. It is getting pretty good investigation, though root cause hasn't been tracked down yet
• 2 security bugs (561031 & 534666) still need to be landed. I'll be increasing the poking today

Firefox 3.5.10

• 1 open blocker
  • Bug 532246 - Fixed in 1.9.2, just needs to land on 1.9.1 (and the same patch should apply)
  • Bug 424558 - May not apply to 1.9.1, asked in the bug Doesn't apply to 1.9.1
• 3 bugs need landing, I'll bug people to get them in by tonight

Browser / Front End

You know how I always say that planet firefox has the details? Still does!

Thing that might impact your team this week: startup timeline work in bug 560647 is on fire. If the people instrumenting code need help understanding your chunk of it, please help them.

This is what it looked like several hours ago - there's more now.

GFX Update

• Cairo update finally landed. Now Jeff can start updating Cairo again.
• It's Benoit Girard's last week as an intern at the Mozilla Corporation! Thank him (BenWa on irc) if he's made your life better.

Hardware Acceleration Update

• Matt Woodrow has a patch to enable GL layers for accelerated video on Mac, should land today or tomorrow

Layout Update

• Chris Double landed new YUV conversion code ... faster and more accurate
• liboggz/liboggplay/libfishsound libraries are now unused and will be removed any day now
• David Baron landed new -moz-any selector

Content Update

• Lots of investigation and fixing relating to DOM performance
• HTML5 parser very close to being enabled by default.

Platform-specific Support Update

JS

https://wiki.mozilla.org/JavaScript:SpiderMonkey:GC_Futures

Electrolysis

Startup Performance

• Startup timeline: Vlad landed the core startup timeline code (bug 480735). Now he, Ehsan, Shawn and a host of others are cranking on adding instrumentation all over the codebase. Follow along in bug 560647.

• Taras' research found that the giant binary approach in bug 525013 is more complicated and reduces our options for improving binary read performance. The new effort is being tracked in bug 561842.

Security

Breakpad

• Ted has a cute baby!
• Jim has reassigned the upstream patches to the new reviewer, but we're going to take local patches for now, because the lack of x86_64 Mac support blocks our ability to complete the unit test on Talos mac project.

Tree Management

• Fedora oranges - Help wanted - New set of fedora oranges
  • bug 560882 - [tracking bug] Reftest Fedora oranges
    • bug 558910 - Several reftest tests are failing for Fedora unit tests from antialiasing
    • bug 558977 - Rev3 Fedora permaorange: reftests/svg/objectBoundingBox-and-fePointLight-01.svg
    • bug 558981 - Rev3 Fedora permaorange: reftests/text-decoration/underline-block-propagation-2-standards.html
    • bug 559125 - Rev3 Fedora permaorange: reftests/bugs/180085-1.html and reftests/bugs/359903-1.html
    • bug 559426 - Rev3 Fedora permaorange: borders in layout/reftests/box-shadow/boxshadow-inner-basic.html and reftests/bugs/456219-1c.html
  • Big big thanks to ehsan, zwol, mak, davidb, philor, dwitte and others who have actively helped armenzg with the other oranges

• Going ahead with rollout of gcc 4.5 on Linux. ETA 2 weeks. Will happen on mozilla-central first.

• Connectivity issues between MPT and Castro still biting us bug 555794
  • bug 561419 landed to only upload breakpad symbols to FTP. Saves around 1GB per checkin.
  • We're in better shape, but still at risk of saturating the link. Faster link is coming soon.

Automation & Tools Update

• Fennec Electrolysis mochitests
  • Finding a means to enable developers to easily run/debug these
• Crash Automation System Rollout beginning

Roundtable