canmove, Confirmed users, Bureaucrats and Sysops emeriti
960
edits
Platform/2009-08-18: Difference between revisions
→Startup Performance: wikilink
(→Startup Performance: wikilink) |
|||
| (33 intermediate revisions by 16 users not shown) | |||
| Line 2: | Line 2: | ||
=== Notices / Schedule === | === Notices / Schedule === | ||
'''[[Releases/Firefox 3.0. | '''[[Releases/Firefox 3.0.14|Firefox 3.0.14]] / [[Releases/Firefox_3.5.3|Firefox 3.5.3]]''' | ||
* code frozen | |||
* targeting release for early/mid September | |||
'''[[Releases/Firefox_3.5. | '''[[Releases/Firefox 3.0.15|Firefox 3.0.15]] / [[Releases/Firefox_3.5.4|Firefox 3.5.4]]''' | ||
* no schedule set yet | |||
* blocker owners will get emails this week | |||
'''Firefox 3.6a1''' | '''Firefox 3.6a1''' | ||
* Active Daily User numbers stabilized | |||
** 13,000 on Firefox 3.6 Alpha 1 | |||
** 5,000 on Namaroka nightly builds (2k on mozilla-central nightlies) | |||
'''Firefox 3.6b1''' | |||
* string-frozen (need dates and deadlines) | |||
* co-ordinated with Fennec RC? | |||
* component leads need to identify P1 issues for this beta | |||
'''mozilla-1.9.2''' | |||
* branch took place on August 13th, 2009 | |||
* mozilla-central re-opened for checkins at noon PDT | |||
* last common changeset is [http://hg.mozilla.org/mozilla-central/rev/c95ef80a5f20 c95ef80a5f20] | |||
* first divergent changeset is [http://hg.mozilla.org/mozilla-central/rev/93a7af291db1 93a7af291db1] (the ones before that landed on both branches) | |||
* if you started using a trunk nightly before August 13th, 2009, you are now a mozilla-1.9.2 nightly user; to get back on to mozilla-central nightlies, get a fresh [ftp://ftp.mozilla.org/pub/firefox/nightly/latest-trunk/ latest-trunk build] | |||
=== Blocker Report === | === Blocker Report === | ||
| Line 19: | Line 38: | ||
=== Browser / Front End === | === Browser / Front End === | ||
* see our [[Firefox/Projects|active projects]] and get involved / propose others | |||
** multitouch JS API prototype work (see [[Firefox/Projects/Multitouch_Polish/DOM_Events|proposed DOM events]]) | |||
** identifying places where animation in the browser would help the UX (blog post coming) | |||
** new UI mechanism for application notifications (see [[Firefox/Projects/Doorhanger_notifications|the spec]]) | |||
** struck a team to evaluate if it will be possible to [[Firefox/Projects/Personas_Uplift_Exploration|integrate Personas]] for Firefox 3.6 | |||
** meeting with the JetPack team to discuss how to build Firefox in ways that help them (providing APIs, etc) | |||
** also looking at protecting users with out of date plugins, see our [[Firefox/Projects/Plugin_Update_Referrals|current ideas]] | |||
** also looking at Ts and Tsnap priorities | |||
=== GFX Update === | === GFX Update === | ||
* Decode-on-draw is getting close to review-worthy. {{bug|435296}} | |||
* Mark Steele starting work on WebGL implementation. | |||
** Updating Vlad's work with new spec changes - native arrays | |||
=== Layout Update === | === Layout Update === | ||
| Line 41: | Line 71: | ||
=== Platform-specific Support Update === | === Platform-specific Support Update === | ||
* Good progress on 64-bit Mac OS X builds. Once patch for {{bug|435041}} lands, only remaining work is complex text input (masayuki) and Cocoa printing (mstange). | |||
* Gecko1.9.2/Firefox3.6 and Java for Mac OS X: There is currently no Java plugin for Firefox 3.6 on Mac OS X due to the removal of OJI, LiveConnect, and XPCOM plugin support. JEP requires those components. A new Java plugin is due to be released within a month which will be compatible with Firefox 3.6. | |||
=== JS === | === JS === | ||
* 1.9.2 blockers need to be re-triaged | |||
* JS trace tests refactored | |||
* large parts of nanojit merge complete | |||
* GC responsiveness issues continue to improve (code cache no longer flushed on GC) | |||
=== Startup Performance === | === Startup Performance === | ||
* | |||
* Joel | |||
** read up on previous optimization efforts and will start running Firefox in Shark very shortly. | |||
* Taras | |||
** landed mmap jar io [https://bugzilla.mozilla.org/show_bug.cgi?id=504864 (bug 504864)], combined startup jars [https://bugzilla.mozilla.org/show_bug.cgi?id=468011 (bug 468011)]. Looks like decent Talos improvements across the board. | |||
** Working on moving res/ stuff to a jar [https://bugzilla.mozilla.org/show_bug.cgi?id=508421 bug 508421] | |||
** Up next, moving app prefs into jars [https://bugzilla.mozilla.org/show_bug.cgi?id=507288 bug 507288], optimize jar io further [https://bugzilla.mozilla.org/show_bug.cgi?id=510611 bug 510611], etc | |||
* Drew | |||
** Linking XPTs on Mac landing this week | |||
* Vlad/Ddahl | |||
** Startup timeline has review, landing soon. | |||
* Dietrich | |||
** Enabling PGO for SQLite, mozStorage and Places code on m-c this week | |||
** Got JS components joined (thanks to some help from fqueze), testing now for perf effect | |||
** Dirty profile testing being enabled this week by Alice, testing two different Places profiles. More coming - send suggestions of dirty testing to me | |||
** Details of other ongoing work available on my [http://autonome.wordpress.com/2009/08/14/weekly-firefox-startup-performance-roundup/ weekly update blog post] and on the [[Firefox/Projects/Startup_Time_Improvements|project wiki]]. | |||
=== Security === | === Security === | ||
=== Electrolysis === | === Electrolysis === | ||
Necko: | |||
* Close to having necko IPC protocol hooked up and initialized | |||
* Some high-level design progress (from conversation with BZ) | |||
** IPC probably "main chrome thread <-> content process", not "Socket transport thread <-> content process" | |||
** Probably use "if(ContentProcess)" within existing necko code, rather than create new set of classes for Content-side necko. | |||
* Issue: which protocols (network; file?) need IPC? | |||
* Issue: goodbye standalone necko? | |||
* Set up design discussion for workweek | |||
Plugins: | |||
* NPStream implementation uncovered some nasty race conditions... discussion ongoing in mdt.dom and chromium-dev | |||
* It's at least clear that prior run-to-completion behavior within the plugin itself will no longer be true | |||
Fennelectrolysis: | |||
* joe still working on tilebrowser graphics: ETA undetermined | |||
* need basic graphics before people can really start working on events | |||
=== Tree Management === | === Tree Management === | ||
* TB2.0.0.23 this week | |||
* Turned off TP3, fast-talos machines yesterday. | |||
** yeaaaah for faster, newer, TP4! | |||
* New dirty-profile talos suite coming later this week. | |||
* l10n nightly updates on mozilla-central as of Sunday night (16th). | |||
** users on nightly before 16th need to manually pave-over-install like usual | |||
** working on getting l10n nightly updates on 1.9.2 | |||
* WinCE build-on-checkin now on mozilla-central, mozilla-192 | |||
** no unittests, talos | |||
* proposed changes to fennec linux-arm unittest/talos jobs | |||
** run desktop unittests during the day, on n810s at night | |||
** will add another 20 n810s to the (tiny) pool, and revisit. | |||
=== Roundtable === | === Roundtable === | ||
| Line 58: | Line 140: | ||
** These are automatically updated every morning before 6am Pacific time | ** These are automatically updated every morning before 6am Pacific time | ||
* Who is allowed to mark <code>approval1.9.2+</code> ? (dbaron) | * Who is allowed to mark <code>approval1.9.2+</code> ? (dbaron) | ||
** Answer: all [[Firefox/Drivers|drivers]], not just endgame drivers | |||
* should we mark the set of bugs that are fixed on mozilla-1.9.2 by virtue of branching as fixed1.9.2+ (beltzner) | |||
** Core/Toolkit/Firefox: [https://bugzilla.mozilla.org/buglist.cgi?query_format=advanced&product=Core&product=Firefox&product=Toolkit&resolution=FIXED&chfieldto=Now&field0-0-0=resolution&type0-0-0=changedafter&value0-0-0=2008-12-01&field0-1-0=resolution&type0-1-0=changedbefore&value0-1-0=2009-08-13&field0-2-0=resolution&type0-2-0=changedto&value0-2-0=FIXED 6034 bugs FIXED since Dec-01-2008 and Aug-13-2009] (warning: takes a while to load) | |||
** this would miss some NSS and NSPR bugs, but they're harder to query | |||
** it would also cause a good deal of bugspam :) | |||