Notices / Schedule
- Firefox 3.0.12
- missed code freeze
- please work on your blockers
- Firefox 3.0.13 schedule and blocker list coming soon
- rc1 released to beta channel early last week
- rc2 released to beta channel and posted on web
- will likely need an rc3 for very localized fixes (no additional qa cycle)
- making a call on shipping at tomorrow's meeting
- 2 blocker nominations - trying to get more information on them
- 0 code blockers left without a fix
- 11 total outstanding blockers not marked fixed1.9.1/verified1.9.1 (0 of which are code-related)
- 6 patches approved over the last week
Browser / Front End
- moving on to TSnap bugs which aim to make Firefox observably faster
- some of us are working on WinCE compatibility issues
- also returning to polish bugs and exploratory sprints
- Reminder to review the Firefox.Next plan and add your comments/feedback on the discussion page or in the dev-apps-firefox thread
- Polish update: Firefox is 60% shiny (+2% change as some were set invalid)
The polish priority is based on a combination of UI location / frequency / severity.
- UI location: is the polish issue in the main interface, a secondary interface, or even deeper into our UI?
- Frequency: how often the user is likely to encounter the issue?
- Severity: when the user encounters the issue are they even able to notice it?
Here are general descriptions of the levels:
- P1 - Polish issue that appears in the main window, or is something that the user may encounter several times a day.
- P2 - Polish issue that is in a secondary interface, occasionally encountered, and is easily identifiable.
- P3 - Polish issue that is in a secondary interface, occasionally encountered, or is not easily identifiable.
- P4 - Polish issue that is rarely encountered, and is easily identifiable.
- P5 - Polish issue that is rarely encountered, and is not easily identifiable.
- CSS3 Fonts working draft published!
- Goals work - please comment and/or add or edit goals in the dev.planning thread or the wiki page.
- Landed the new Win32 rasterizer for great justice. Have already seen some perf improvements (Tsvg) come through on dev.tree-management.
- Removing gfxIImageFrame and nsIImage - please comment in bug 753 if you depend on them.
- Bobby Holley (bholley) joined us again! He'll be working from Paris.
- No 1.9.1 activity
- Compositor phase 1 passing mochitests; needs IME/accessibility testing
- CSS transitions demo (dbaron)
- SMIL CSS patch awaiting review (dholbert)
- Font refactoring and Harfbuzz work (jfkthame)
- Video poster element (cpearce)
- Working on goals
- Work on HTML 5 parser
- No reports of webcompat problems.
- Actual parser changes being reviewed.
- Talos perf problem partially understood, but not yet fixed.
- Hope to land preffed off soon.
- Slim wrappers
Mac OS X Update
- some blockers landed today (a leak, a security fix)
- 1 hard to repro crash showing up on the reporting server
- tracing recursion underway
- tracing a few more cases to help with benchmarks
- ARM / NJ fixes landing
- discussing OOM policy in the newsgroups
Goal discussion ongoing: in particular, the security team is going to discuss which aspects of Mozilla need fuzzing.
Jesse and dveditz are going to start looking at the backlog of crash bugs with the intent of reducing the number of open crash bugs by 20%. They will be trying to using a GTD workflow to help track. More details to follow later, see Jesse's blog for a basic outline.
Multiprocess tabs are very basically working! See cjones' video.
Stage I is almost complete. We're finishing up small tasks so that phase II can begin:
- sub-protocols and protocol management (cjones)
- removing the writable string buffers/ldscript/short-whcar hack (??)
- integrate IPIDL into the build system (bsmedberg)
- landing the code into the project repository, with basic review and TODOs clearly marked (bsmedberg)
- working on Windows again (bent)
- IPIDL and general IPC documentation (cjones)
Phase II is where we can begin parallel hacking of different subsystems. The immediate list:
- networking (jduell)
- global history... needs async history (sdwilsh)
- docshell hookup, session history, and link targeting (bz)
- preferences (bsmedberg)
Additional items to spin up as we get more comfortable.
Multi-process plugin work needs love. cjones got something hooked up, but the plugins are not drawing, and it needs a lot of fleshing-out (kinetik?)
- converted almost all the talos slaves from "dedicated" to "pool-of-talos-slaves"
- reduced/eliminated wait times?
- when do 4hr downtime for try-talos. details in bug 448047
- when do downtime for graph server and talos db changes?
- cleaning out old builds to free up space on ftp.m.o: long-ish discussion about the value of old and very-old builds. beltzner proposes that we do not archive localized nightly builds because they are much less useful for regression testing. Various proposals floated about using S3 or less reliable and therefore cheaper storage to avoid having to delete old builds... please respond to the newsgroup post in dev.builds
- CSS 2.1 test suite (dbaron) -- the CSS WG is building the testsuite for CSS 2.1, which serves multiple purposes and for the purpose of deciding whether CSS 2.1 is interoperably implemented, will only be counting tests submitted by September. If you have tests you'd like to submit, please contact dbaron or fantasai. The WG has tentatively agreed to accept tests in reftest format in addition to the standard format.