damons

Done:

• patches on three blockers (bug 574859, bug 566135, bug 531551)
• unblocked on crash bug 561252 due to low crash activity.
• posted patch to rtl blocker bug 588735
• misc. reviews

Done:

• bug 575950 SSL certificate is not confirmed for secure webpages ; but probably still has problems.
• some minefield bugs reported
• bug 587146 e10s http: fix auth test in testresumablechannel.js
• bug 588298 bad QI of nsNSSCertificate because of wrong typecast

Done:

• bug 571054 - WebGL: emulate vertex attrib 0 on desktop GL
• bug 592416 - several WebGL fixes, about Framebuffer methods and VertexAttribPointer
• bug 591084 - several WebGLTexture improvements
• bug 589814 - fix copyTexImage2D validation
• worked on a few issues blocking enabling WebGL tests by default, see bug 589814 , bug 591438 , etc
• worked upstream to fix a couple of WebGL tests, especially http://www.khronos.org/bugzilla/show_bug.cgi?id=349

Done:

• Finished up bug 584175 to the point where the ETW probes gave MSR what they wanted, though I don't know the outcome from their side.
• Split 588537 (ETW implementation) out of 584175 (probe unification). 584175 is r+'ed, NPOTB. Currently trying to land.
• PTO
• Acquired new desktop system. Set up as distcc server, VM host. (Windows 7 only so far.) My laptop is useful again! (Had to dedicate it to Windows 7 build stuff; now back to Linux.)
• Acquired Mac laptop for repairing dtrace support.

Done:

• Fixed Ts regression by writing on shutdown. Relanded.
• Did some investigation on allHUDServiceTests.js ,I think that the issue is some timing problem (not really caused by startupcache).
• Got talos Ts numbers for scache + mozJSCmpLdr + xulPrototypeCache. Doesn't show any perf win, this is bad/confusing.

Done:

• review bug 559942: Heuristic to pick default HTTP cache size
• reviewed bug 589292 - e10s necko: add contentDisposition prop to nsIChannel
• fixed bug 591552 - SetupReplacementChannel has bogus cast to nsHttpChannel
• started design for bug 591707 - e10s: handle redirects from HTTP to a different protocol

Done:

• bug 589368: "Few wrinkles" took a bit longer than I expected. Posted a patch for review.
• bug 581606: Did some more polish on less-sqlite-fragmentation patch. Ready for landing.
• bug 533038: Reviewed mwu's jar changes
• Bug 590242: Do not open omnijar 3x. got r+

Done:

• fixed regression in painting plugins on windows that was caused by semantics-preserving refactoring not quite preserving semantics
• settled on platform APIs for fennec, have a buggy implementation kinda working

Done:

• Processed a lot of bugmail and other email.
• Tried to figure out what's going on with Hotmail attachments; made progress but the exact cause still unknown; help from Microsoft would be great
• Carefully evaluated the situation with createContextualFragment and Outlook Web App; concluded that an old Gecko quirk needs to be approximated; wrote a patch.
• Investigated yet another case of document.write() blowing away the document. This time due to crazy use of jQuery with a twitter widget. (Developing a fix.)
• Started importing HTML5 parsing algorithm test cases that the html5lib team and the WebKit team have developed.

Done:

• More benchmarking, testing of case-insensitive UTF8 comparator (bug 145975, patch up for review). Working on integrating it into the location bar (bug 570975).
• Work on resource packages + incremental extraction.
• Fixed pushState scroll issue, bug 590573. More pushState spec discussion. (And you thought it was over!)

Done:

• At least ten reviews, mostly for the developer tools group. Some day I'll list the reviews I do in these updates, but today is not that day
• Hacked a bit on compare-talos (file issues if you find bugs or want a feature!) to make it more useful for the next item
• Fixed bug 583611 (blocker)
• Fixed a pymake bustage so people on windows could build with it again
• Created a patch which got checked in (albeit with a poor commit message) to fix an issue where the implementation didn't do what the docs said in docshell code for bug 587573 (blocker)
• Fixed bug 590654 (blocker)
• Fixed bug 519769 (blocked bug 583882)
• Fixed bug 583882 (blocker)

Done:

• Digging out from under a week's worth of bugmail backlog
• Reviewed calc() for heights and remaining widths
• Other reviews
• Landed some more layout and xpconnect performance wins

Done:

Landed:

• bug 590246 - Multitouch swipe gestures
• Bug 566151 - Add "Maemo" to UA string.
• bug 582676 - Slow CSS selectors (part 4)
• bug 590848 - Fix newTab error in prefs-about-button
• bug 559939 - Add basic context menu for
• bug 590246 - Swap meanings of up/down swipe gestures
• bug 589800 - Refactor code to fix pinch zoom on local tabs (fixup)
• bug 586306 - Hide sharing dialog before showing more dialogs
• bug 586313 - Fire an event when panning finishes
• bug 584785 - Restrict sharing based on URI scheme

Awaiting review:

• bug 583280 - Align optgroup labels with top-level options

Other work:

• New add-on: Quit Fennec - https://addons.mozilla.org/en-US/mobile/addon/221517/
• Updated https://wiki.mozilla.org/Mobile/Projects/Sharing with proposed Maemo integration plans and web auto-discovery feature.
• Updated https://wiki.mozilla.org/Mobile/Fennec/Extensions/UserInterface with some initial docs on the new pageaction and contextmenu APIs.

Done:

• Landed bug 586115 (cache compiled scripts for e10s)
• bug 588808, release TabChildGlobals earlier
• Fixed bug 588643 (a crash regression from bug 462076). nsCOMArray::InsertObjectAt works in a very strange and undocumented way.
• Waiting review for bug 585173, process message manager
• an sg:* bug. Waiting for approval.
• W3C
• Reviews

Done:

• IPDL docs updated - the main guide now discusses modern destructors, the multiple forms of constructors, protocol deletion, and ActorDestroy (https://developer.mozilla.org/en/IPDL/Tutorial)
• IPDL best practices guide created - I have grown tired of pointing out the same mistakes with every new protocol created (https://developer.mozilla.org/en/IPDL/Best_Practices)
• e10s geolocation segfault (bug 589721, needs landing)
• fixed incorrect timespan test (bug 561528)
• added gczeal() to content process xpcshell (bug 586362, needs landing)
• fixed pausing and resuming downloads in f10s (bug 588351, needs landing)
• cottaging

Done:

• Progress on message display
• DXR generification
• Prototyping Dehydra and macro support

Done:

• Landing a LOT of code! :)
• formmethod, formenctype, formtarget and formaction attributes (bug 582412, bug 566160, bug 566064)
• Constraint validation API for <object> (bug 589490)
• Fix action IDL attribute (bug 566128)
• autocomplete attribute (bug 557628)
• form attribute (bug 588683)
• UI work for form validation (bug 561636)
• Crying like a baby in front of labels attribute...
• Misc: bug 589026, bug 587875

Done:

added a few last minute changes to my static-checking fixes (bug 573786, bug 572819, bug 576214, bug 575720, bug 578235). dwitte will land once the tree opens.

asked for review on callgraph changes: bug 588002. Using the finish_decl callback from bug 584164, all the info needed for dead code can be gleamed from treehydra alone.

Experimented a little more with post callgraph generation dead code scripts:

Couple of things:

-Can speed things up considerably by computing transitive closure of main and other start points, then computing path relationships for only the remaining dead nodes.

-Whitelisting exported functions (those with default visibility) is too conservative.

-Regarding the display of dead code, the most useful grouping of dead nodes is not the most concise. ie it's better to separately register two dead subgraphs even if one is contained in the other.

Also:

looked a bit into ssa passes in treehydra as possible future project. I can get at the def of an ssa name (possibly a phi node) right now, but the really useful stuff will require hacking convert_tree.js

see: http://gcc.gnu.org/onlinedocs/gccint/SSA-Operands.html

Done:

Recovered fully from bad crash on bicycle (I hope :) )

bug #546606 (async redirect API part 2): Finalized and landed (thanks Honza and jst!)

bug #585214 (Favicon fails to load on some sites): Analysis

bug #586264 (Improve/add logging to async redirect code): filed and added trivial patch this to get better logging for bug #585214

bug #495115 (Implement STS): reviewed async-redirect part

Done:

• IndexedDB event security work
• Moving cycle collector off the main thread

Done:

• bug #513008 - async cache read
• review bug #280661 - SOCKS
• bug #486199 - Support HTTP over SCTP

Done:

Landed:

• TM: remove object allocation debug hook: bug 578159

Done:

Last report was many weeks ago....

• Lots of reviews! Thanks for all the patches :)
• Getting e10s stable in Fennec
• Working on list of features for 2.0 (http://wiki.mozilla.org/Mobile/Planning/2.0)
• Bug fixing
• Docs and blog posts for add-ons using e10s
• Summit presentation on using e10s
• Pushed a few e10s platform patches through reviews and into tree
• Started implementing "Contacts in form autocomplete" using ctypes
• Started enhancing the Fennec nsISessionStore to store more state, support undo and support Sync
• Made a patch to remove canvas tiles from Fennec and use browsers only!

Done:

• Patched and landed fix for bug 572149.
• Reviewed the evalInSandbox parts of bug 534398.
• Helped come up with a list of todo items for Nashville XPConnect hacking work week.
• Got mochitest-chrome passing with my patch for bug 523994 (still need to deal with mochitest and browser-chrome).

Done:

• Worked on the tabs API, landed for the Jetpack SDK 0.5 freeze
• Worked on a demo of the tabs API
• Lots of review iteration on the Console patch
• Reviewed expiration correctness and perf fixes for Mak
• Bugfixes and such for Jetpack SDK 0.5 blockers

Done:

• paternity break
• bug 536273 (Http POST)
• bug 506269 at last :-) (506269)
• bug 543739 (xpcshell tests for libpref) two quick fixes:
• bug 566869
• bug 569270

Done:

• Lots of IndexedDB spec work. Including meeting with the CouchDB guys
• Security bugs
• id-handling rewrite almost done, but have run into a last-minute (merge?) bug

Done:

• initial work to get tests running in fennec+e10s: https://bugzilla.mozilla.org/show_bug.cgi?id=567417
• mochitests on android
• google summer of code tests are making progress: https://bugzilla.mozilla.org/show_bug.cgi?id=489158

Done:

• Have buildbot infrastructure up and running for slow-network simulation to test HTML5 speculative parsing performance.
• Had meeting with jorendorff, bz, and gal about implementing virtual properties in the JS engine.
• Made progress towards implementing the final nsIJetpack/PJetpack multi-process API (thanks to recursive IPDL structs/unions).

Done:

• 3.6.4 hang and crash analysis
• 3.6.4 blockers: https://bugzilla.mozilla.org/buglist.cgi?quicksearch=FIX%20blocking1.9.2:.4+%20@benjamin@smedbergs.us

Done:

• landed patches for bug 560199 and bug 560273
• finished patches for bug 560462 (Use fast unwrapping for more quickstubs)
• worked on patches for bug 564266 (DOMCI GetItemAt/GetNamedItem should return nsWrapperCache)

Done:

Working on bugs/enhancement requests on the topfails website.

Tried [ unsuccessfully ] to create a backPort tracking application for changeset landings. The idea is to ask for a release branch and check if all the relevant landings on the branch are ported appropriately to all the current & active named-branches. Unfortunately, the logic of deciphering what are current active named branches on a given release is not trivial.

Done:

• Fennec is string frozen for the 1.1 release (do patches, reviews)
• Site menu has landed! bug 456077
• Save page has landed! bug 465305
• Work on clean site menu (bug 552526)
• Work on Theme for web content forms (bug 516641)
• Work on a few other patches

Done:

Followed up with jwalker/bespin... was asked to wait until Q2, when the code will be in a better state -- current idea is to have pluggable views (in MVC sense). Some hacking on our caret orange. Some patch prep for landing.

Done:

Signed up. Joined the Accessibility project. Caught up on stuff that had to be pushed back due to the all-hands.

Done:

• xpcshell builds and runs

Done:

Work on bug 516749, finalized the parent/chrome side.

Done:

password echoing follow ups

Done:

• Just finished helping to track down and kill what is hopefully the last ARM bug merged from Tracemonkey (bug 518227).