Platform/2009-07-21
From MozillaWiki
< Platform
« previous week | index | next week »
Contents
Notices / Schedule
- shipping today
- work on your blockers
- code freeze on August 4
- please request approval before landing
- Over 35 MILLION downloads so far!
- shipped an early release last week due to 0-day
- using the new flags for triage
- goal of this release should be a quick-turnaround that:
- topcrash fixes, security/stability patches
- narrow scope, small changes
- contrary to some reports on the Internet, this is the usual process for Firefox and software releases; the 3.5 release was strong, stable and solid, and feedback has been extremely positive. Near the end of the release we become extremely conservative about patches to accept; the 3.5.1 release is a quick update to fold in some patches that came up late in the 3.5 release cycle.
- we will be going through approval1.9.1.1 flags and triaging them; no need to change them to approval1.9.1.2 flags
Firefox 3.6a1 Notes from today's meeting on branch and alpha schedule:
- Our current plan is to branch for 1.9.2 on July 31st. There are several items that must land and become stable before we branch (All discussed below).
- There are several bugs related to ireflow that bz must fix.
- Phase I of Compositor is to land after dbaron finishes review.
- All bugs that block the branch must be marked blocking-1.9.2+ and have P1 priority status. Please mark immediately.
- We will try to do an alpha before we branch on July 31st. Since time is short, we will make the call whether or not an alpha is even possible before we branch, on Friday, July 24th. Even then, things are tight, and it's quite possible that we won't have time for an alpha.
Blocker Report
Gecko 1.9.1.2 / Firefox 3.5.2
Gecko 1.9.2 / Namoroka
- 298 open blocking1.9.2?, 35 open blocking1.9.2+
- 79 open blocking-firefox3.6?, 3 open blocking-firefox3.6+
Browser / Front End
- Asynchronous location bar: one visual glitch outstanding
- Focus rewrite: two bugs left, one needs Karlt
- Form awesomecomplete: done this week
- Responsiveness: ongoing
- Ctrl-tab: moved behind pref
- Polish update: Firefox is 59% shiny (0% change)
- Thanks go to: dougt (P2), dao (P2)
- 46 remaining easy polish bugs (whiteboard [polish-easy])
- 38 remaining hard polish bugs(whiteboard [polish-hard])
GFX Update
- bug 753 - landed!
- Landing bug 753 means that it's now possible to make informed decisions about the individual pieces that depend on it.
- Decode-on-draw is working pretty well, but has a lot more bugs to go. Bobby Holley is going to be sprinting on it this week, and we'll likely be able to make a call on it this time next week.
- Aero peek backend, bug 501490, is just awaiting review, but also depends on some other bugs. Will be ready.
- Backend code also contains a taskbar framework that other Windows 7 enhancements depend on.
- Windows 7 enhancements that depend on Aero peek's taskbar/superbar framework:
- Aero peek frontend bug 474056 is coming along, and will likely be ready in time.
- Download status in the superbar, bug 474060, probably won't make it for 1.9.2, but sid0 is working on it.
- Jumplists, bug 473045. Should be in 1.9.2, will probably make it?
- All the above should be ready by mid-August - soon enough?
- Windows/Windows 7 enhancements that don't depend on taskbar/superbar:
- Window docking on Windows 7, bug 489258, is just waiting on review for both 1.9.1 and 1.9.2.
- Parental Controls:
- bug 355555 - Content blocking by parental controls. In progress.
- bug 423587 - Extension manager install blocking parental controls - up for review.
- Cleartype
- bug 504698 - Windows XP doesn't have Cleartype (subpixel text AA) enabled by default, which has upset some downloadable font-using authors. Fix is going to be to add a hidden pref, default to on, to force Cleartype on downloadable fonts. This will be submitted for 1.9.1 and 1.9.2.
- Question: Do we want to default Cleartype on for all content? That's what IE does. (Probably want to check in backend support for this, then let products make their own call. 1.9.2 only.)
Other news:
- topcrasher bug 470487 fix seems to have worked! Backport to 1.9.1?
Layout Update
- Patches on 1.9.2 radar
- Compositor phase 1
- CSS gradients
- CSS background-size
- bz's framelist cleanup
- Core Text?
- wanted for better OpenType support (esp. Arabic, relevant for @font-face use), and for broader testing as Core Text is required for 64-bit
- requires moving builds to 10.5 SDK (bug 504179), then adding --enable-coretext configure option to mozconfig
- changing SDK introduces risk of incompatibility with 10.4; need to maintain testing on 10.4 until we officially drop it as a supported OS
- known sqlite issue on 10.4 in bug 423691 (has workaround)
- Patches probably waiting for post-1.9.2
- Font system unification cleanup
- Remove float placeholder splitting
- Disable interruptible reflow for 1.9.2 default
Content Update
- Making progress on XBL2!
- Continuing followup work for the HTML5 parser (check out parser/html/java/README.txt)
- Working out what it means to have more than one Java plugin.
- Focusing on the remaining interruptible reflow regressions.
Mac OS X Update
- Markus Stange is very close to having a completed patch for fullscreen in 1.9.2, bug 420491.
- 64-bit Mac OS X port blocked on gfx bug 493280.
- current intention is to land 493280 on trunk right after 1.9.2 branches; is it wanted earlier?
- Cocoa NPAPI event model close to finished, part of it landed, most of the rest under review. Lacking complex text support at the moment.
- Steven Michaud fixed two topcrashers (bug 503196 and bug 503391). landed a patch to support plugin printing (bug 191046) and is reviewing Josh's Cocoa NPAPI event model patch (bug 435041).
JS
- GC plan underway JavaScript:SpiderMonkey:GC_Futures
- converging on JIT with Adobe
- measure, measure, patch
Security
- Content Security Policy wrapping up implementation. Hoping to land before 7/31 1.9.2 branch date, but ok with missing alpha (if there's an alpha before branch)
Electrolysis
- Fixed MSVC8 and GCC 4.1 build bustages
- Buildbots should be in production tomorrow, if things go well
- Fighting with mixed size of wchar... chromium requires 4-byte and we require 2-byte for NS_LITERAL_STRING to work. Several complicated preprocessing schemes are being discussed, but none are especially pretty
- Starting a small subproject to demonstrate multiprocess+fennec: because Fennec doesn't use native widgets the same way, we need to ship bitmaps across the wire as well as remote mouse and keyboard events
Tree Management
- intermittent woes:
- talos redness caused by failed posts to new graph server. bug 476208
- win32 build failures caused by NTFS file system woes "circular directory" bug 496712
- Places project branch handed over last week - no complaints??!?
- Electrolysis branch coming online tomorrow morning.
- Please start gently, while we watch overall load
- has usual builds/unittests, but not talos Details in bug 500755
- http://tinderbox.mozilla.org/showbuilds.cgi?tree=Electrolysis
- filling out Tracemonkey
- enabling mobile linux-arm builds tomorrow morning bug 505219
- unittests, talos coming next week?
- Infrastructure load:
