Platform/2010-10-19
From MozillaWiki
< Platform
« previous week | index | next week »
Contents
Notices / Schedule
Firefox 4 Beta 7
- still waiting for GC compartments to land on the relbranch
- compatibility issues with JetPack and Firebug
- stability issues (see roundtable)
- minor theme tweaks are being landed in both places
Firefox 3.6.11
Firefox 3.5.14
Blocker Report
A handy list of Firefox 4 triage queries is available:
- Beta 7: 19 blockers
- Beta: 486 blockers
- Final: 389 blockers
- TOTAL: 875 blockers
- nominations: 183 nominations
- ready to land: 12 blockers ready
Firefox Development
- Need to find owners for the high priority unowned theme bugs
- Need help making bug 588764 performant (roc/vlad?)
(from our goals):
Firefox
- [AT RISK] Finish Firefox 4
- [ON TRACK] API-Complete Jetpack
- Joint with Jetpack team
- Currently working towards feature-complete beta SDK
- [ON TRACK] Stand up basic Firefox with Electrolysis
- Joint with Platform
- UI may not be functional, should launch and open a web page.
- Project tracking page
Developer Tools
- [NEW] Finish Firefox 4 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
Platform
Breakdown of all 2.0 blocker and current assignments. Shows a lot of items in unassigned lists (nobody, general.js)
There are about 31 crash bugs where the blocking status has changed since Oct 13th. We removed the dups.
There are about another 22 crash bugs in the nomination queue we would like to get triaged so we have a handle on all the blockers.
GFX
- Direct3D 10 layers backend will be turned on by default RSN.
- Needs one or two prerequisites to land first.
- Will give us some pretty great Ts/Txul wins because we no longer create a D3D9 device.
- Will also help us with newer hardware, like NVIDIA Optimus, which doesn't handle the D3D9/D3D10 interop very well at all.
- Should also get us some perf wins on some heavy rendering sites.
- This means that DirectX 9 will only be used on computers incapable of Direct2D, such as XP machines.
Layout
- dbaron's CSS 2.1 testsuite results: [1]
- Significantly better than Webkit and Opera, a little behind IE9
- Blockers (including SVG and Widget)
- 2 beta8 blockers
- 35 betaN blockers
- 96 final blockers
- 14 requests need triage
- Risky issues:
- Laptop trackpad breakage due to hacky drivers sniffing our HWNDs: bug 605357, bug 594977, bug 593372
Video/Audio
- Blockers
- 0 beta8 blockers
- 2 betaN blockers
- 10 final blockers
JS
- ES5
- Landings
- strict this landed in TM, also in m-c, to be in b7 soon (bug strictThis
- Current work
- Making eval in strict mode not affect the variable set of the calling code (bug 514568)
- Refactoring of eval code (bug 604504)
- Landings
Tree Management
- Sunday's downtime successful. New infrastructure in place, so 650castro has better redundancy, and better bandwidth capacity.
- new firewall deployed during Sunday downtime had a bug related to h323 data
- this caused intermittent network failures whenever anyone started h323 traffic.
- Firewall patched with fix from vendor at ~09:30 Monday, so we *think* all is good.
- meanwhile, if you still see builds/tests failing because of network issues, please file bug in mozilla.org/ReleaseEngineering and we'll triage.
- yesterday's power outage in 650castro
- cause unknown, IT still investigating
Roundtable
- [blizzard] I need some XUL experts
- [smooney] Problems with Stability of the Trunk - less stable then we were before.
- 2010-10-17, 4500 crashes on the trunk with 31K active users
- 2010-10-13, 1400 crashes on the trunk with 30K active users
- With a similar user set, our crashes have increased by more than 3X.
- More details at https://wiki.mozilla.org/CrashKill/Firefox4#Status.
- [ctalbert] Mcote recreated the blocker report stuff that we had in the run up to 3.6
- Take a look and let us know if there's anything you like to see throw comments our way in bug 599582