- Finished proof of concept for Spark UI proxy extension.
- Started cleaning up proof of concept code to create a utility that can be used by myself or others for features. This includes adding some missing parts like error handling, testing on different versions of Python, etc.
- Continue cleaning up Spark proxy extension and submit a pull request once it's ready.
- Pick up one of the other open issues and begin investigating it.
- posted patch: Bug 1247641 - New Tab page creates thousands of requests to my server
- uplifted: Bug 1236364 - address bar pull down delete deletes wrong entry
- worked some more on: Bug 1235397 - Investigate merging the action row with the input field
- reviewed: Bug 524757 - Add architecture and operating system to about:support
- reviewed: Bug 1243779 - Remove uriIsPrefix option from nsINavHistoryQuery
- needinfoed: Bug 1155824 - Restart Firefox badge on menu changes position when window loses focus
- holiday next Monday
- Met with Desktop Retention team in SF
- Got started on onboarding art direction
- Retention team sprint planning
- Complete initial art direction
- Catch up on bug questions, prioritization and filing
- Filed lots of oranges
- Fixed?: Bug 1241692 - Intermittent test_restyles.html | CSS animations running on the compositor should not update style on the main thread - got 3, expected +0
- Wrote patches: Bug 1242872 - nsAnimationManager::CheckAnimationRule should not create temporary animations
- Wrote patches: Bug 1196114 - Animation.IsRunningOnCompositor should indicate its state for each properties
- Wrote a patch: Bug 1246893 - nsSMILKeySpline::GetSplineValue should return 0 or 1 on its boundaries
- Wrote patches: Bug 1216842 - implement effect-level easing
- Filed: Bug 1247800 - Need to investigate why transform or opactity animations in invisible elements run both on main-tread and compositor on Android
- Wrote an automation test: Bug 1218620 - Opacity animation with transform animation which is disabled to running on compositor because of Gecko's bug should be running on compositor
- Tutored an intern
firstname.lastname@example.org 2 days ago
- Refactor/implement the new HTTP-redirect Fetch step
- More tests for still-untested aspects of Fetch
email@example.com 3 days ago
- Bug 873163 - Investigating nsGlobalWindow leak with vimperator addon.
- Bug 1242084 - Investigating nsTerminator pseudo leak.
- Ported over my BFS shortest-path find_roots algorithm to the CC version.
- landed: Bug 1245673 - Remove useless NS_INT32_TO_PTR from nsTraceRefCnt.
- filed: Bug 1247025 - [e10s] Various Windows-specific content process graphics leaks.
firstname.lastname@example.org 3 days ago
email@example.com 3 days ago
firstname.lastname@example.org 3 days ago
Automation with V:
- Generally lots of bugging people on irc to figure out how to do various things.
- Discover that the default V on Ubuntu 12.04 is too old. Build a version of the trunk for it instead. Place at valgrind.org and do unholy hacks in the Docker file to make it be downloaded/installed for the automation runs.
- Hassling to get logging working again (irc w/ jonasfj w/ DNS)
- Look at timeout problems -- mochitest timeouts, IPC timeouts. Difficult to figure out what is relevant and what isn't. Set up a very slow machine (Intel Atom) so as to be able to study that more locally.
- Learnt about faster Amazon EC2 node types, but decided to stay with the slowest kind for now.
- Set up chunked runs. Run tests split into up to 40 pieces.
- Work around issues with the treeherder error filter, so the "Error summary" pane actually shows Valgrind errors.
- First plausible run! https://treeherder.mozilla.org/#/jobs?repo=try&revision=3de21c4e5b75 . This represents around 100 to 200 CPU hours on the slowest node type.
- Bugs: bug 1229348 (landed), bug 1245566 (created), bug 1246719 (created, landed), bug 1246962 (created)
- Still some way to go before landing it all. The Automation people seem to want this to be all-green before landing, but I am not sure how realistic that is.
- bug 1136762 (races in nsPipe3.cpp). Made fix, in review.
- Started to look for a next-race to fix, not much progress.
- xpcshell tests on tsan started hanging. Urr.
- Set up 2fa
- A bit more Valgrind bug triage
More of the same, but with somewhat increased emphasis on race removal, which has fallen behind recently.
email@example.com 4 days ago
Reviewed: - bug 1153053, bug 1243233, bug 1243349, bug 1243861, bug 1243918, bug 1244128, bug 1204202, bug 1243750, bug 1240284, bug 1242724, bug 926980, bug 1239828, bug 1242632, bug 1245269, bug 1245745, bug 1245476, bug 1246040, bug 700997, bug 1246333, bug 1246334, bug 1246550, bug 1246505, bug 1242641, bug 1246592
- bug 1244941: Don't fill install manifest for artifact builds
- bug 1244997: Remove TOOLKIT_EM_VERSION from toolkit/xre/Makefile.in
- bug 1244999: Move icon-related DEFINES to moz.build in browser/app
- bug 1245013: Move CMFLAGS/CMMFLAGS from config.mk to configure
- bug 1245015: CompileDB doesn't contain commands for ObjC sources
- bug 1245022: Kill stlport's Makefile.in
- bug 1245027: Move LOCAL_INCLUDES to moz.build in media/libvpx
- bug 1245035: Move LOCAL_INCLUDES to moz.build in media/omx-plugin/lib/ics/libvideoeditorplayer
- bug 1245055: Remove workarounds for assembly using r7 on thumb
- bug 1245422: Cleanup some crufty C*FLAGS on Windows
- bug 1245763: Don't emit Sources objects when there is no Linkable in the same directory
- bug 1245764: Move Windows RTL flags (-MD/-MT) flags out of config.mk
- bug 1238769: Add a -xpcshell option to Firefox
- bug 1246779: Stop disabling optimizations for test plugins on Windows
- bug 1246874: Unify GRE_BUILDID and MOZ_APP_BUILDID at the build system level
Waiting for review:
- bug 1246871: Generate platform.ini with the preprocessor instead of an ad-hoc script
- bug 1246881: Generate a header defining MOZ_BUILDID and kill config/buildid
- bug 1246894: Remove msys-perl-wrapper
- More of the same
firstname.lastname@example.org 4 days ago
- Bug 1178298 (APZ ride the trains) - feedback
- Bug 1213434 (facebook loads faster from search in chrome) - (content-perf) feedback
- Bug 1217364 (fennec scroll not smooth) - (content perf) feedback.
- Bug 1217372 (fennc input lag in text boxes) - (content perf) feedback.
- Bug 1233214 (2-5% linux memory regression from uncompressed omni.ja) - analysis, guidance.
- Bug 1238040 (update checkerboard telemetry measure), Bug 1238137 (scroll triggers telemetry), discussions, telemetry analysis - hopefully last probe update will start provide meaningful data.
- Bug 1241362 (BHR jank events don't match about:telemetry histograms) - analysis, landed fixed histogram display, uplifted up to 45.
- bug 1243842, Bug 1242176 (saved logins deleted on update to fx 44) - happened to me, filed one of those, update to 44 halted after ~1/4 updated, discussion on use cases, ended up as 44.0.1 which got pushed to the rest of the users (~3/4) today.
- Bug 1242692 (upto 90% regression on some paint speed tests on linux due to disabling xrender) - discussed with author - the numbers don't mean much, overall it's an improvements, the test is still valid - just the transition should be ignored.
- Bug 1243379 (no way to collect what are the offenders for increasing GC durations) - feedback, long discussions with smaug, might need to end up just observing memory allocations over time to detect leaks instead of getting this info from the GC/CC.
- Bug 1243863 (tresize regressions on train to 46) - some guidance.
- Bug 1244259 (60% xperf regression due to safe browsing changes) - feedback (investigation progressing well without me).
email@example.com 4 days ago
bug 1226564 - Reopened for necko to take a look at. Explained the reason why the experiment's numbers are the ones to look at for e10s comparisons, not the ones on t.m.o. Maybe we should push more on bug 1224264 so we can get that data on t.m.o and show the differences.
preliminary experiment data:
- Ran :poiru's stability analysis with SUBPROCESS_CRASHES_WITH_DUMP instead of ABNORMAL_ABORT as the latter can over-count according to :bsmedberg. The results were largely the same, possibly because of Skype click-to-call filling in the e10s numbers. (bug 1222890)
- Ran an analysis truncating responsiveness measures to the jank threshold (> 100ms). Shows that, on the beta 45 experiment, even the parent+child jank is the same in both cohorts. Ship it!
- Modernized :rvitillo's addon jank analysis to see what addons in the experiment are jankiest with a bias towards the most popular. #1 spot: Adblock Plus.
Ran a couple of SSE2 analyses for :mhoye and :kats for a dev-platform thread. tl;dr - All but about 0.5% of Firefox users on release have SSE2, and almost all of those 0.5% are running the latest version of Firefox on Windows XP.
bug 1231968 - got PROFILER_LABEL to work from JS, upgraded the profiler guts to have better guards against improper RAII usage.
bug 1246205 - What looked like a large regression in INPUT_EVENT_RESPONSE_MS was instead a reflection of no longer creating, emitting, and handling 2.5 Billion "0ms" input events in the parent process, per day.
Learned how to tunnel the spark UI properly to investigate why, sometimes, clusters just don't do things.
firstname.lastname@example.org 4 days ago
Fixed up the tests, refactored the code to reduce copy & pastes, added set-up/tear down code for specific test
Get caught up to speed on the new spark project and readings.
email@example.com 4 days ago
Monday - 2/1/2016
- Cleanup and retesting on Linux
- Expanded test configs
- Worked on getting blog up and running
Tuesday - 2/2/2016
- More cleanup of atsy
- Retested on OSX
- Added basic setup steps, tested script on linux and OSX
Wednesday - 2/3/2016
- Added manual atsy test, tested on all browsers
- Added semi-automated Safari test
- Tracked down various Safari related processes by diffing ps states
Thursday - 2/4/2016
- Ran manual test on Safari
- Ran manual test on IE
- Ran manual test on Edge
- Attempted automated test on Edge, definitely doesn't work currently
Friday - 2/5/2016
- Sketched out e10s memory usage blog post
- Played with WordPress
- Wrap up blogs
- Work on removing libxpcomrt
firstname.lastname@example.org 5 days ago
- bug 1237058 - CrossCompartmentWrapper handling is slow on Google Spreadsheet
- bug 1244222 - Race when setting ToXPCOMCallback to a content function
- bug 1245213 - Servo docs hang for >20s in buildIndex() when Gecko profiler is enabled
- landed: Stylo uplifts Part 2 #9515
- bug 1244222 - Race when setting ToXPCOMCallback to a content function
stylo planning / coordination
landed: Remove the msg dependency from style #9261
- landed: Use cargo features to reduce the dependencies of components/util in GeckoLib builds #9267
- landed: Add Partial Implementation of style DOM traits for Gecko, and some basic glue to let Gecko call into Servo #9284
- bug 1025338 - Remove JSPropertyDescriptor::object()
email@example.com 5 days ago
- Some cleanup after landing Uint32x4 in IonMonkey. We don't have a 1-1 mapping between MIRType and SimdType which caused problems with SIMD boxing/unboxing.
- Bug 1244254. Terrence found a threading problem with TSan. Ion was looking at SIMD template objects from the compilation thread.
- Uint32x4 support for Odin/Baldr. The signed/unsigned integer types in SIMD.js are causing more problems. WebAssembly will not distinguish signed/unsigned vector types, so reusing the wasm compiler types for validating asm.js is tricky.
- Proposed that asm.js won't allow unsigned SIMD types in function signatures and global variables. This will simplify interoperation with wasm code.
SIMD support in nightly:
SIMD.* SM Ion Asm Int8x16 Y Int16x8 Y Int32x4 Y Y Y Uint8x16 Y Uint16x8 Y Uint32x4 Y Y Float32x4 Y Y Y Float64x2 Y Bool8x16 Y Bool16x8 Y Bool32x4 Y Y Y Bool64x2 Y
Uint32x4 support for Odin/Baldr
firstname.lastname@example.org 5 days ago
- Started integrating Jonas' and my code for FlyWeb. See bug 1245963.
- Spoke with CD "Magnet" project team about potential for collaboration.
- Wrote a report of progress for Nicole to provide to CD.
- Bunch of other coordination stuff.
- Keep on with bug 1245963, land.
- Keep following Jit Coach progress, wait for right time to post blog.
- Speak with Brian Mastenbrook of Wearable, exploring prototyping of use cases and opportunity for collaboration.
email@example.com 5 days ago
- rebased e10s SW interception changes and got initial tests passing
- reviewed more Fetch tests
- reviewed network error PR
- reviewed compiletest PR
- reviewed file load cancellation PR
- reviewed mozbrowser securitychange PR
- reviewed image load PR
- addressed review comments for iframe load blocking
- addressed review comments for lazy compilation
- rebased and addressed review comments for media element
- Finalize NCSU projects
- Fix test failures for e10s SW interception redesign
firstname.lastname@example.org 5 days ago
- Created first version of longitudinal dataset (https://mail.mozilla.org/pipermail/fhr-dev/2016-February/000792.html)
- Tried different tech to provide SQL access to our data
- Various hiring related stuff
- UCOSP intern mentoring
- Reviewed various e10s analyses (https://github.com/vitillo/e10s_analyses/pull/17)
- Spark support
- Looked at luigi, airflow and drake for ETL scheduler.
- Churn analysis review
- Interviewed Brandon Wolfgang
- Bug 1245489 - Add more metrics to longitudinal dataset - filed, reviewed
- Bug 1245492 - Sort longitudinal dataset by subsessionStartDate and profileSubsessionCounter - filed
- Bug 1245490 - Add crash pings to longitudinal dataset - filed
- Bug 1245494 - De-dupe pings in longitudinal dataset - filed
- Bug 1245514 - Histograms.json should adhere to a schema - filed
- Bug 1245569 - Deploy Zeppelin notebook server on Spark clusters - filed
- Bug 1191789 - Add extended telemetry flag as dimension on S3 - followed up
- Bug 1245818 - Trim Telemetry aggregates - filed
- Bug 1219505 - Telemetry experiment for unified urlbar-searchbar with one-off search buttons - support Marco’s analyses
- Bug 1182637 - [meta][e10s][telemetry] e10s is jankier than non-e10s ('gecko_hangs_per_minute', Telemetry experiments) - followed up, reviewed
- Bug 1246137 - Add nightly and aurora channels to longitudinal datasets. - filed
- Bug 1246137 - Add nightly and aurora channels to longitudinal datasets.
- Bug 1246408 - Update EMR release to 4.3.0 - filed
- Bug 1246425 - Import Parquet datasets in Hive metastore. - filed
- Bug 1246426 - Make Parquet dataset queryable with Presto - filed
- Bug 1245536 - Update derived stream for Beta 45 experiment - closed
- Bug 1246420 - Paquet partitions should use underscore_casing instead of camelCasing - filed, closed
- azhang project discussion meeting
- Loasis meeting
- Empty histograms meeting
- Discussion about new team goals
- Metrics analysis meeting
- A.t.m.o reboot meeting
- Add-on hangs discussion
email@example.com 5 days ago
- Landed bug 1243851.
- Wrote a patch for 1205298 (will probably flag for review today).
- Filed bug 1245877.
- Created several WIP patches for bug 932080, all of which captured the easy bit (actually supporting shorthand destructuring with defaults "cover initialized name" but fail on some edge case or another. This head against wall was a good enough learning experience that I should have a 100% working patch this week.
firstname.lastname@example.org 5 days ago
- Write a tab switching probe + histogram - https://bugzilla.mozilla.org/show_bug.cgi?id=1242013
- Fix https://bugzilla.mozilla.org/show_bug.cgi?id=1245846
- Statuser stuff:
- New README.
- Add child hang reporting mode and parent only mode.
- Various other fixes and feedback for Statuser, like warning banner tweaks and switching modes not clearing the count.
- Add it to the perf wiki page.
- Setup and test telemetry-batch-view on a Spark cluster.
- Fix JSON ping schema: https://github.com/SamPenrose/data-pipeline/pull/1
- Fix https://bugzilla.mozilla.org/show_bug.cgi?id=1245114.
- Write a thing to list histograms that Fennec users have submitted - https://pastebin.mozilla.org/8858787.
- Finish adding additional metrics to the longitudinal datasets - https://bugzilla.mozilla.org/show_bug.cgi?id=1245489
- Requirements + specification for a.t.m.o project.
- Follow up on a few open bugs.
email@example.com 5 days ago
- Released v0.1.0 of the (renamed) parsell parser combinator library (https://crates.io/crates/parsell).
- Pushed 137 commits to parsell (https://github.com/asajeffrey/parsell).
- Finished tidying up and documenting the Parsell API (http://asajeffrey.github.io/parsell).
- Added myself to the servo review pool (https://github.com/servo/saltfs/pull/213).
- Reviewed https://github.com/servo/string-cache/pull/141, https://github.com/servo/servo/pull/9525 and https://github.com/servo/servo/pull/9276.
- Get back to wasm, no really for real this time.
- Reviewing PRs.
firstname.lastname@example.org 5 days ago
- Upgraded Mac development environment, to much gnashing of teeth. Still not done, as I think Xcode helpfully overwrote my 10.6 SDK.
- Worked on standalone
libc++for OS X builds.
- Tried to help lsalzman deal with Skia's whole-hearted embrace of C++11 stdlib constructs.
- Made cross-compile builds for OS X work with Rust.
- Played with Ralph Giles's Rust build scripts.
- Attempted to make cross-compile builds for Android work with Rust.
- Wrote patch to make some forms of leak checking at shutdown more debuggable.
- Discovered more cases of duplicated refcounts, worked on fixing those.
- Worked on building multiple Rust crates in the build system (bug 1163224). Need to figure out how to handle inter-crate dependencies.
- Looked at the Gecko profiler a bit for surfacing more stuff from the event loop.
- 1:1 with Nick.
libc++builds on the Mac.
- Rust build system work.
- Need to see about re-supporting (?)
i686-linux-androidtarget for Rust.
- Will probably wind up doing another Gecko/C++ onboarding session.
email@example.com 5 days ago
- Rebuilt servo-mac1
- Evaluated different price & performance points for EC2 build machines
- WAY too many infrastructure crises. Hopefully better now.
- Spun up and investigated a mac pro instead of mac mini for Servo builds
- Updated cargo (after ~2 months of poking around, alex and I finally found the cargo bug and were able to update!)
- Moved us to reserved linux EC2 instances
- Lots of interviews
- Made two RA offers, working on the rest of them (including legal/paperwork for a third)
- Investigated headless OSX testing to speed up OSX builders
- Revisions for the final version of the ICSE paper
- Either get headless OSX testing work or try to split OSX CSS & WPT across more machines
- Test gw's WebRender PR on Windows
- Submit final ICSE paper
- More drafts of Servo arch. overview, both for us and the EuroPar submission
- Windows & ARM CI support (requires homu additions)
- Android startup perf
- Oxidation planning
firstname.lastname@example.org 5 days ago
- Bug 1240517 - Display a message in the chat history when a user unexpectedly leaves a co-browsing session
- Bug 1244621 - Change link clicker UI favicon to Firefox Hello logo
- Bug 1245147 - Inform the link clicker when tab sharing is paused
- Bug 1245149 - Inform link clickers on mobile that they should use a desktop
- Bug 1245149 - Inform link clickers on mobile that they should use a desktop
- IOS Menu: https://docs.google.com/document/d/1b-icFn4Km5jacRUIhgW4OqXLMZmV9bRhI1YGvqNGFWw/edit#heading=h.cgya9ohwk6bq
- Feedback from use of the latest Nightly build (FF46): https://docs.google.com/document/d/1b7asULNfm-CLJw8NR4doeupQ_9iOB--H_zC8ZDGWoVU/edit
- Malware, Uncommon and Unwanted Downloads in Firefox Desktop Download Panel: https://docs.google.com/document/d/1P0vP0UJhkXF-jDjTc_lskprhV_uMuqIx_DcdebG8JCw/edit#heading=h.qvogb1tunyod
- Aha! HAO-46 Created by Romain Testard on Nov 17, 2015 Terminating the web sharing session prompts the user to assign a name to the room
- New user journey link clicker UI strings: https://docs.google.com/document/d/1M8rW_BpDI1iXZ55hwaMaelnrZutPk6j_73laxLttPtc/edit
- (Filed) Bug 1245534 - Request Plugins Doorhanger is poorly formatted
email@example.com 5 days ago
firstname.lastname@example.org 5 days ago
- Worked with the onboarding teams to get started
- Work with ckprice and dolske on defining goals for the QX team
- Wrote and revised a job post
- Broke down tab center into job stories https://docs.google.com/spreadsheets/d/1PwEWPWy8NfqguExQyJWjrULWK2J9pSP9i84SUjm1i2I/edit#gid=0
- Work with recruiting/HR to get the interaction designer job post out
- Prioritize backlog for QX
email@example.com 5 days ago
- Bug 1245050 - Convert from CRLF to LF for WindowsUIUtils.cpp
- Bug 1245038 - On screen keyboard isn't visible after inputing character via IME on command prompt
- Bug 1201904 - [x86_64][Win] IME is disabled on Flash Player
- Bug 1238326 - erroneous Korean output in Google Docs
- Investigating fix
- Bug 1243268 - Support ImmSetCandidateWindow(CFS_EXCLUDE) on plugin process
- Discussion for plugin shortcut support
- Flash issue such as IME
- Analyze Google docs bugs.
firstname.lastname@example.org 5 days ago
- Shared memory spec work following TC39
- Travel, maintenance
- Two PTO days
- Shared memory spec, test suite, bugs, ...
- Review backlog
email@example.com 5 days ago
- Random triaging
- Bug 1242342: wasm if/if_else control flow opcodes (+ assertion in bug 1246136)
- Bug 1243810.
- Opened bug 1245023, about storage of enum classes.
- Bug 1242835, bug 1243374: oom fuzzbug
- Bug 1245048: static analysis bug
- Bug 1245421: remove dead code in jit/TypeAnalysis
- Bug 1244254: initial investigation and work, but most bigger than expected -- has been taken over by Jakob.
- Started work on bug 1246116 (loop / break opcodes in wasm).
- 18 reviews (Bug 1240796, Bug 1244403, Bug 1244405, Bug 1244828).
- bug 1246116, loop / break opcodes in wasm
- more reviews
Ms2ger@gmail.com 5 days ago
- 2016-02-01: filed heapsize issue #28 No need for experimental gate for HashMap
- 2016-02-01: wrote gaol PR #17 Setup up the minimum possible CI on Travis.
- 2016-02-01: wrote web-platform-tests PR #2541 Fix "shoud" typos in HTML tests
- 2016-02-01: wrote web-platform-tests PR #2542 Fix "shoud" typo in XHR test
- 2016-02-01: wrote web-platform-tests PR #2543 Fix "shoud" typos in WebGL tests
- 2016-02-01: filed gaol issue #18 Run tests on Travis.
- 2016-02-01: wrote servo PR #9485 Update references to pcwalton's repositories.
- 2016-02-01: wrote rust-fnv PR #7 Notify homu.
- 2016-02-01: wrote testharness.js PR #181 Fix typo in the readme.
- 2016-02-01: wrote gaol PR #19 Compile tests on Travis.
- 2016-02-01: wrote saltfs PR #206 Add gaol and rust-fnv to homu.
- 2016-02-01: wrote gaol PR #20 Avoid some unnecessary allocations in exec().
- 2016-02-01: filed gaol issue #21 Stop using OsStr::as_bytes().
- 2016-02-01: updated https://github.com/servo/servo/wiki/Adding-a-repo-to-Homu
- 2016-02-01: enabled homu for gaol and rust-fnv
- 2016-02-01: wrote web-platform-tests PR #2547 Fix some nits in embed-document.html.
- 2016-02-01: investigated bug 1243110
- 2016-02-01: filed heapsize issue #31 Consider relicensing
- 2016-02-01: fixed comment on r/6144 Various improvements to websocket.js.
- 2016-02-01: addressed a comment on web-platform-tests PR #2522 Fix #2411: Merge DOMTokenList/DOMSettableTokenList
- 2016-02-01: wrote rust-url PR #164 Enable doctests and correct nearby documentation.
- 2016-02-01: filed web-platform-tests issue #2548 Add a test for textarea.value and child text nodes
- 2016-02-01: filed rust-mozjs issue #237 Conversion to vector should throw a TypeError
- 2016-02-01: filed servo issue #9489 Pass the wpt XMLHttpRequest (XHR) test suite
- 2016-02-01: filed servo issue #9490 Implement XMLHttpRequest.send(Document)
- 2016-02-01: wrote servo PR #9491 Disable send-entity-body-document.htm.
- 2016-02-01: finished web-platform-tests PR #2530 Various improvements to websocket.js.
- 2016-02-01: filed bug 1244751
- 2016-02-01: rebased servo PR #9470 Bump Rust to 2016-01-31 nightly
- 2016-02-01: wrote servo PR #9493 Use the select macro on the paint thread.
- 2016-02-01: investigated heapsize issue #31 Consider relicensing
- 2016-02-02: wrote webidl PR #89 Remove references to the removed platform arrays.
- 2016-02-02: filed FileAPI issue #30 Consider using USVString for the Blob constructor
- 2016-02-02: wrote web-platform-tests PR #2553 Remove a pointless stylesheet inclusion.
- 2016-02-02: worked on getting Travis to test with multiple --features arguments
- 2016-02-02: wrote euclid PR #119 Introduce an 'unstable' feature, test on all channels.
- 2016-02-02: wrote euclid PR #120 Bump version.
- 2016-02-02: wrote https://github.com/servo/app_units/pull/13
- 2016-02-02: wrote html PR #610 Use the concept of a 'closed' Blob from FileAPI
- 2016-02-02: filed dlib issue #3 Macro trouble
- 2016-02-03: wrote web-platform-tests PR #2556 Correct an assert_array_equals call.
- 2016-02-03: wrote skia PR #88 Update euclid.
- 2016-02-03: wrote rust-azure PR #214 Update euclid.
- 2016-02-03: wrote io-surface-rs PR #49 Update euclid.
- 2016-02-03: wrote skia PR #89 Update io-surface.
- 2016-02-03: wrote rust-azure PR #215 Update skia.
- 2016-02-03: wrote https://github.com/ecoal95/rust-offscreen-rendering-context/pull/49
- 2016-02-03: wrote servo PR #9519 Update Euclid.
- 2016-02-03: filed web-platform-tests issue #2558 Make it trivial to have a test run in window and worker scopes
- 2016-02-04: wrote servo PR #9536 Remove a dead_code annotation from a live method.
- 2016-02-05: filed homu issue #12 Consider showing the Travis status in the homu queue
- 2016-02-05: wrote servo PR #9554 Remove some unused dependencies from the util crate.
- 2016-02-01: reviewed heapsize PR #27 Change the bound for HashMap to use BuildHasher
- 2016-02-01: reviewed web-platform-tests PR #2546 Fix "eigth" typos in HTML li[type] tests
- 2016-02-01: reviewed rust-cssparser PR #92 Allow heapsize 0.2.z
- 2016-02-01: reviewed heapsize PR #29 Support char
- 2016-02-01: reviewed heapsize PR #30 Bump to 0.2.2
- 2016-02-01: reviewed r/6138 Fix #2411: Merge DOMTokenList/DOMSettableTokenList
- 2016-02-01: reviewed servo PR #9470 Bump Rust to 2016-01-31 nightly
- 2016-02-02: reviewed string-cache PR #137 Allow heapsize 0.2
- 2016-02-02: reviewed heapsize PR #32 Support Atomic types
- 2016-02-02: reviewed rust-selectors PR #66 Allow heapsize 0.2
- 2016-02-02: reviewed servo PR #9509 Deduplicate the cocoa crate.
- 2016-02-03: reviewed heapsize PR #33 Implement HeapSizeOf for () and Result<T, E>
- 2016-02-04: reviewed heapsize PR #35 Implement HeapSizeOf for &'static str
- 2016-02-04: reviewed servo PR #9383 Clean up XHR API on Reviewable
- 2016-02-04: reviewed r/6166 improved testing of sync open() and events, fix #2559
- 2016-02-04: reviewed r/6166 improved testing of sync open() and events, fix #2559
- 2016-02-04: reviewed r/6167 Only drop into pdb on exception if --pdb is passed on the command line
- 2016-02-04: reviewed web-platform-tests PR #2563 Test sync and async in one file - #2559 follow-up
- 2016-02-04: reviewed web-platform-tests PR #2562 Add [PutForwards=value] to HTMLAreaElement#relList
- 2016-02-04: reviewed servo PR #9521 Make websockets work in a worker scope
- 2016-02-05: reviewed servo PR #6677 Make iframes block the enclosing document's load event
- 2016-02-05: reviewed a bit of servo PR #8491 Compile raw inline event handlers lazily. Resolves #8489. on Reviewable
- 2016-02-05: reviewed a bit of servo PR #8491 Compile raw inline event handlers lazily. Resolves #8489. on Reviewable
firstname.lastname@example.org 5 days ago
- 2016-2-1 3:53 Misc analysis, notably GMail.
- 2016-2-2 9:39 Triage, mail, follow up clipboard stuff, GMail and Facebook. 1:1 chat, mail.
- 2016-2-3 3:39 Compatipede run, mail, misc.
- 2016-2-4 5:53 Mail (bugmail, Clipboard follow-up), some XHR test refining. More bugmail. More clipboard thinking, this time in bug 860857.
- 2016-2-5 0:57 .
- 2016-2-6 8:59 Facebook scrolling.
TOTAL: (1980min): 33:0
Noteworthy: finished these tasks:
- Look at Compatipede 2 test data, ponder heuristics
- Finish the "use QueryParams" pull request for webcompat.com
- ZeroClipboard HTML5
- Compatipede 2, test CSS analysis and Phantom timeout
- Plugins for C 2
- Write more test cases for C 2
- How can we switch SlimerJS binary on the virtual machine to test Nightly C 2-style?
- Blog about webcompat 1984
- More video testing - handle the open, underanalysed video-related bugs
- Analyse data from latest SlimerJS test run
- bug mail catch up, bugzilla triage
- webcompat.com needsanalysis
- Finish webcompat.com E-mail subscription backend work
- Refactor Marionette docs on https://developer.mozilla.org/en-US/docs/Mozilla/QA/Marionette
- XHR test meta data
- Work on test automation - needed for getting an overview of implementation status
- Clipboard - writing data formats and security - review what the spec says, decide on RTF
WHEN PHANTOM FIX IS RELEASED ( https://github.com/ariya/phantomjs/issues/10158 ):
- Experiment with -webkit- style discovery from JS running in Phantom
- And explore phantom.onError functionality
- Check status of Clojure bugs
- See if we can use JS error logging on C2 to test those Clojure JS bugs (if any are still live?)
- Look into Responsive mode faking META viewport support
- Talk again to Jet et al about using the memory data generated by SlimerJS test runs (but SlimerJS test runs are at some point going to be obsoleted by Compatipede 2, which doesn't as of now do memory stats)
- Keep tweaking the script to extract "data points" from existing site compat tests
- AWCY: Make the "data from compat-entomology" feature an on-demand thing
- Check if Compatipede-2 can run on PaaS
- fixlist sites testing - build DB of data points
- Maybe stuff all current regression tests into a compat-entomology table and make test scripts running regression tests get their "sitedata.js" directly from the compat-entomology server?
- (Also remove all regression tests that can be replaced by "data point" monitoring)
- compat-entomology auth (for submits)
- static AWCY
- Figure out if it's possible to make perf (mem/cpu time) measurements on Flame hw while playing URLs through Marionette
- Make synced browsing output test data in webcomptest JSON format
DON'T FORGET ABOUT (longer term):
- test bug 891247 fix
- Test other stuff regarding clipboard implementation (and the clipboard tests test suite)
email@example.com 5 days ago
- Fix a number of browser.html + WR bugs.
- Rebase and tidy up the WR servo branch.
- Add support for texture atlas borders for GPU raster ops.
- Ensure that text nodes don't inherit opacity.
- Investigate, report and test ipc-channel deadlock on Linux.
- Fix a few more browser.html + WR bugs.
- Open WR servo PR to merge to master.
- Work on fixing some of the remaining WR reftest failures.
firstname.lastname@example.org 5 days ago
- Continue familiarizing myself with Spark; look through issues and pick one to work on
email@example.com 5 days ago
- Bug 1187137, bug 1187151, bug 1187188, bug 1181444 - Finally finished nsBaseHashtable enumeration-to-iteration conversions! The old enumeration functions are now gone. A Q1 goal completed.
- Bug 1244606 - Remove PL_DHASH_FASTCALL. A small clean-up.
- Bug 1245233 - Avoid unnecessary RegExpObject clones in CopyScript().
- Bug 1244982, bug 1244992, bug 1245393, bug 1245761 - Small improvements in memory reporting coverage and accuracy.
More profiling and optimization of memory usage with multiple content processes.
firstname.lastname@example.org last week
- Merged pull request for JS client_id check
- Updated README with updated install and jupyter_notebook_config.py info
- Did initial cleanup of Python code according to PEP8 and submitted pull request
- Start working on the keyboard shortcut for listing all running Spark jobs
email@example.com last week
- fixed: bug 1239889 - KeyframeEffectReadOnly constructor crashes when the target element does not have the parent?
- reviewed: bug 1174575 - Implement a (CSS)PseudoElement interface
- reviewed: bug 1211783 - add the KeyframeEffect interface
- reviewed: bug 1226047 - Implement AnimationEffectTiming interface
- reviewed: bug 1244080 - Add an automation test for bug 1235478
- reviewed: bug 1244586 - Implement KeyframeEffect constructor
- reviewed: bug 1244595
- ni: bug 1242872 - nsAnimationManager::CheckAnimationRule should not create temporary animations
- triage: bug 1141468 - HTML based hamburger menu slide-out/in animation is choppy
- triage: bug 1245601 - Firefox nightly hangs forever on zoomcare.com
- investigate: bug 1216842 - implement effect-level easing
- filed: bug 1245000 - Ship Element.animate
- filed: bug 1245748 - Store specified values in KeyframeEffectReadOnly
- filed and debugged: bug 1246046 - Pausing and seeking a play-pending animation does not render as expected
- Edited CSS Animations to make animation-delay live
- Made tests for CSS test suite for dynamic changes to animation-delay
- Various edits to SVG spec
- CSS WG F2F
- SVG WG F2F
- Meet with Google to address Web Animations interop issues
- fix: bug 1246046 - Pausing and seeking a play-pending animation does not render as expected
- 1 day PTO
- 1 day public holiday
- 1 day travel
firstname.lastname@example.org last week
- Organized Bug Program Pilots for Hello and Core::DOM
- Hello starts 2/8
- Core::DOM starts 2/16 (Overholt on PTO next week)
- Set up Worksheets for Hello and Core::DOM Triage
- Start triaging new bugs in Hello
- Get Developer Tools on-boarded and triaging
email@example.com last week
- Update the TabCenterTest add-on to use the light theme by default.
- Talk to vng about making Doorhangers in addons.
- Talk to Matt Grimes about Shield and Test Pilot.
- Help vng access browser-chrome from his add-on.
- Add radio buttons to the Style Guide.
- Add checkboxes to the Style Guide.
- Talk to Mike Hoye about getting community contributions to the Style Guide.
- Add Inputs content.
- Add Panels content.
- Port the "GitHub Canned Responses" Chrome add-on to a WebExtension!
- Help notwaldorf get the "GitHub Canned Responses" WebExtension up on AMO.
- Email Ghostery developer about upgrading to a more-supported options API.
- Help Triage WebExtensions bugs for Version 1!.
- Make a shallow-dive about the various data sources for Monday's meeting.
- Send Rebecca Weiss the Data Sources info I collected.
- Make Whimsy e10s compatible.
firstname.lastname@example.org last week
- Synced tabs contingency plans
- FxA eye password experiment https://github.com/mozilla/fxa/blob/rfeeley/eye-password-experiment/features/FxA-80-eye-password-experiment/README.md#treatment-ux
- FxA reset password improvement https://github.com/mozilla/fxa/blob/rfeeley/reset-password/features/FxA-72-reset-password/README.md
- Notifications on received tabs https://bugzilla.mozilla.org/show_bug.cgi?id=1244597
- Synced Tabs button introduction for multi-device syncers
- UX support for AMO to FxA migration
email@example.com last week
- Worked on script loader changes to support modules. Try run looking promising. (bug 1240072)
- Worked on off thread compilation of modules loaded by the browser
- Implement the getOwnPropertyDescriptor hook for module environments (bug 1245518)
- Added a missing OOM check while doing structured cloning (bug 1245520)
Jury service. No fixed end date but probably until 22nd Feb.
firstname.lastname@example.org last week
- Windows perf: Found minorgc issue on raytrace (bug 1245485) which terrence fixed and should land soonish
- SM: Investigating fallout from shared stubs (bug 1243757). Tried poking with try, but unable to find cause yet.
- SM: thought about LICM last iteration unrolling. Which I thought Sinking meant.
- SM: Add documentation to binary arith optimization annotations and extra fixes (bug 1242578)
- AWFY: some maintenance. Get windows 10 working again. Edge is updated. Update webaudio to newer version. Get Unity working more reliable.
- SM: move and/or knowledge in Ionbuilder, bug 1246141
email@example.com last week
Getting back on webcompat.com dev a bit.
firstname.lastname@example.org last week
- Fixed a ton of TSan bugs.
- Removed Traceable. Any class with a trace method can now be used with [Persistent]Rooted.
- Removed the restrictions on PersistentRooted so that you can now store any kind of GC pointer in one.
- Understood, simplified, and rewrote the comment about wrapper rooting.
- Reviewed intern application test results.
- Read and better understood browser tracing.
- Converted more kinds of roots over to using RootLists.
- Check talos results to see if we can remove RootLists from JSContext.
- Make ExposeToActiveJS a barrier.
email@example.com last week
Long overdue update!
- Attended Mozlando
- Completed Exec Report powered by Unified Telemetry
- Redshift credentials are available in the metadata, see example
- Fixed telemetry-server to deploy latest histogram_tools.py (and stop the annoying alerts)
- Fixed a problem with old telemetry stemming from https://bugzilla.mozilla.org/show_bug.cgi?id=1201492
- Squeeze in some Holiday PTO
- Help vladikoff with Spark Analysis
- Several discussions about longitudinal analysis and schemas / validation.
- Fixed Telemetry scheduled job failure Bug 1238016
- Split 'core' pings out of OTHER https://bugzilla.mozilla.org/show_bug.cgi?id=1242258
- Verified that core pings are flowing in
- Fix a.t.m.o failure https://bugzilla.mozilla.org/show_bug.cgi?id=1242433
- Met and discussed with Dave Z, Chris M, Josephine T
- Initial proof of concept based on raw UT data ("active" signal only)
- Derived dataset (parquet) using telemetry-batch-view framework
- Schedule a job to keep Churn dataset up to date
- Backfill churn data to Nov 1
- Revise PoC using derived dataset (active only)
- Add "inactive user" signal
- Fix sort+dropDuplicates, use Window Function instead
- Generate sample data with active and inactive signal
- Begin backfilling aggregate data to Nov 1, 2015 (Up to week of 20151206 to 20151212 so far)
- Continue backfilling / validating churn aggregates
- Package jupyter-notebook-gist on PyPI
- Help with longitudinal analysis where possible
firstname.lastname@example.org last week
FileList and Directory:
- Bug 1173320 - Allow FileList objects to contain Directory objects
- Bug 1244782 - Removed non-used variables in Directory.cpp
- Bug 1243881 - Consider unshipping translateTime
- Bug 1244995 - Crash [@ JSAutoCompartment::JSAutoCompartment | mozilla::dom::Console::ProcessCallData] with gczeal
- Bug 1245184 - CookieManager should remove cookies only if they match the userContextId
email@example.com last week
- dealing with web compat fallout with block-scoped functions
- some progress on implementing raising DebuggeeWouldRun
- probably more web compat fallout
- evaluate jit coach UI
- wasm baseline debugging
firstname.lastname@example.org last week
This post includes this 2 weeks work:
- FIXED: Bug 1052947 - WidgetMouseEvent::e4thButtonFlag and e5thButtonFlag or windows WM_XBUTTONDOWN / XBUTTON1 mouse events are always handeled as Browser: Forward and Browser:Back and there is no option to disable this functionality in about:config or anywhere else
- FIXED: Bug 1237582 - [TSF][e10s] Browser hangs after bug1234422 with (hidden) Microsoft New Changjie or New Quick
- FIXED: Bug 1237216: [TSF] Unnecessary composition events are raised when typing Korean characters
- FIXED: Bug 1240336 - At setting either <textarea>.value or <input>.value, we shouldn't commit composition if the value won't be changed
- FIXED: Bug 1153156 - [e10s] Mousewheel scroll distance does not match non-e10s (with apz disabled)
- FIXED: Bug 1242895 - [non-e10s][TSF] Forcibly committing composition during TSF locking document fails to commit composition and makes "hidden" composition string
- FIXED: Bug 1242331 - [e10s][TSF] When typing fast, IME composition may be committed unexpectedly and input won't cause text input after that
- FIXED: Bug 1243657 - [IMM] When there is only composition string in editor, canceling the composition causes painting caret at odd position (right-most of the editor)
- Reviewed: Bug 1226148 - On-screen keyboard should stay open (rather than hiding and reshowing or just plain hiding) when switching focus between two inputs/textareas/contenteditables
- Reviewed: Bug 1236058 - virtual (on screen) keyboard should not be opened when a bluetooth keyboard is connected
- Reviewed: Bug 1241408 - Google docs commits composition text per input with TSF and non-e10s mode
- Reviewed: Bug 1110030 - Routing hardware key events to keyboard app when an input field is focusing.
- Reviewed: Bug 1243268: Support ImmSetCandidateWindow(CFS_EXCLUDE) on plugin process
- Reviewed: Bug 1243656 - Use async method for RequestCommitOrCancel
- Reviewed: Bug 1240208 - Give a hint to on-screen keyboards that the awesomebar accepts URLs (e.g. to display '/' and TLDs)
- Answered to ni?: Bug 1203059 - [e10s] Default shortcuts are not supressed when overridden by web-application
- Answered to ni?: Bug 1244240 - ABORT: constructor for actor failed: file ipc/ipdl/PLayerTransactionChild.cpp | ABORT: file ipc/chromium/src/base/shared_memory_posix.cc
- Answered to: Bug 354358 - keydown/keyup events should be dispatched even during composition (but keypress shouldn't be so)
- Investigated: Bug 1244607 - Custom system mouse scroll settings are ignored
- Pinged to reporter for confirming if already fixed: Bug 717087 - compositionupdate events are fired when the device orientation changes
- Pinged to reporter for confirming if already fixed: Bug 717088 - compositionupdate events aren't being fired when text is deleted
- Filed: Bug 1244647 - "Share this page" button doesn't open its popup due to ReferenceError: DOMParser is not defined in microformat-shiv.js
- Bug 1137572 - Add some useful methods to TextEventDispatcher for native key and IME handlers
- Bug 1137561 - Make KeyboardLayout, nsIMM32Handler and nsTextStore should use TextEventDispatcher
- Bug 1137563 - Make TextInputHandler use TextEventDispatcher
- Bug 1137565 - Make KeymapWrapper and nsGtkIMModule use TextEventDispatcher
email@example.com last week
- ongoing work on a blog post about typed arrays
- make array/typed array indexOf/lastIndexOf methods never return -0 (bug 1242043)
- make ArrayBuffer "neutering" (old ES terminology) be "detached" in code/comments (bug 1079844)
- NOTE: "neuter" is still an HTML term of art, applying to things more generally than just ArrayBuffer, so it's still the right term for some things (just not applied to ArrayBuffers specifically)
- minor Gecko refactoring to enable Intl.DateTimeFormat.prototype.formatToParts in b2g certified apps, only (bug 1216150)
- inline the PossiblyWrappedTypedArrayLength intrinsic (bug 1239068)
- make CompartmentOptions& non-optional to JS_NewGlobalObject (bug 1239666)
- add an import script for importing fdlibm source for our use (bug 933257)
- slim down Date instances some (bug 1240283)
- eliminate an assertion-botch in OOM testing of Vector (bug 1238582)
- fix stringifying of class constructors to include "use strict" in them (bug 1243596)
- various reading of a DOM security spec for window/location objects, from the cross-global-access POV
- met with Facebook peoples, about perf/other issues on their side; filed bug to expose why code doesn't syntax-parse (bug 1242087) as one useful result
- tried to make structured cloning throw when transferring a detached ArrayBuffer (bug 919259): hit complexity, pondering whether to fix it or punt it
- attended TC39 meeting to co-present (with littledan) a proposal to lock down the global object's prototype chain (Object.prototype and the global object, to be precise), followup to bug 1052139
- well-received and accepted
- adding a MOP operation deferred til proposal/spec text written up (but tentatively interested)
- rest of the meeting was of questionable value (shocker)
- review queue
- finish up for-loop stuff (bug 449811)
- blog post about typed array bugfixes -- bumping this upward, feeling itchy to write this
- get cracking on a few goal-ish things
- write proposal for a setImmutablePrototype MOP operation before a (March?) TC39 meeting
- do %TypedArray%.prototype.set perf-testing, enable the self-hosted version (bug 1140752)
- investigate our use of various atomics operations for correctness -- particularly all Relaxed cases, and interrupt flags cases especially
- combinatorically exhaustive .caller tests (bug 1057208)
- update update-test262.sh to pull from github (and work at all again)
- realloc semantics discussions followup (bug 1035001, bug 1035002)
- arguments object sealing stuff (bug 937922)
- eliminate an in-source comment pointing to bug 934889 as still needing a fix (it no longer does)
- move style checking into the build (bug 1063728) (landed but backed out for various inanities, sigh)
- tinyid/shortid stuff that busted Windows (bug 985687)
- ES6 environment introduction into the JSAPI
firstname.lastname@example.org last week
- Fixed Linux graphics breakage caused by double buffering accidentally disabled in glutin GLX back-end (tomaka/glutin#715).
- Still investigating intermittent test failures; not much forward progress yet.
- Some work on std::collections::BTreeSet perf and docs.
- Got rid of the obsolete
- Continue the intermittent hunt
- Work on pinch zoom and WebRender
email@example.com last week
- Closed some old bugs
- Made a successful warnings-as-errors compile with gcc 4.9.3, which should allow upgrading at least mulet (the original impetus) and hopefully desktop firefox
- created two new taskcluster builds: b2g hazards and browser hazards. Still not completely working
- more whining on mailing lists, trying to get people to understand what I see as the underlying UX issues in mozreview. ("I'm right! Look at me!")
- Further investigation of memory usage/recursion on bad site bug 1234973
- uncovered more badness. Need to split out into 3 bugs.
- Finish up the two TC builds
- some GC stuff would be nice
- Fix the functions that are getting rejected by sixgill. Turns out it hit a rather important one, though fortunately it doesn't matter for the rooting analysis. Inconvenient for other analyses, though.
firstname.lastname@example.org last week
- TC39. Killed a proposal which would have been hard for us to implement performantly, and had weirdly thought out semantics. A direct win.
- REVIEWS (Oh god, you can't just take a week off of reviews...)
- Punt on JSOP_*PROP_SUPER to get classes jit intergration back moving
email@example.com last week
- bug 1234355: Remove "Don't ask me again" footgun option from the "enable DRM" info bar notification - Reviewed patches.
- bug 394984: Unable to update on mac if admin user is not the same admin user as the person who installed firefox - Addressed review feedback, fixed newly discovered issues, sent patches for additional review.
- Update orphaning dashboard - Revised wiki page based on feedback. Continued implementation.
- bug 394984: Unable to update on mac if admin user is not the same admin user as the person who installed firefox
- Continue work on the proposal for the update orphaning dashboard as well as the spark job and the actual dashboard.
firstname.lastname@example.org last week
- Cont'd with Bookmarks/RL updates
- Onboarding V2 follow-up
- Notifications exploration
- Address Bookmarks/RL feedback
- Subscribe-to-page ideas
- Misc bugs to follow up on
- Notifications exploration
email@example.com last week
- Workflows: The ESM/Diary Study launched last week and has been dominating my time. We have 32 people participating a 7 day Paco experiment. Many participants have had to be replaced, so we are extending it by 2 days for replacement participants. Between troubleshooting, German language translation, and pretty disappointing recruitment team, it's made for extremely long days.
- Workflows: Legal and Risk Analysis questions about ESM. Went through a "Rapid Risk Analysis" for the first time :)
- Public Speaking: Gave a 3-hour workshop to about 75 librarians at the LACONi Continuing Education program at the Skokie Library on Thursday morning.
- Recruiting: Phone screened 1 candidate.
- Recruiting: Met with Tim to go over on-sites and some other issues.
- Content Strategy: Reviewed Michelle's draft, gave feedback.
- User Testing: Met with Sevaan to discuss copy testing and iOS research.
- Mobile UR Office Hours: Presented my Research 2015 Overview to the team.
- Workflows: Provide workflows survey to ESM participants. Translating email copy today.
- Public Speaking: Giving a talk at the DePaul HCI School tonight on Contextual Inquiry for UX Research.
- Workflows: ESM/Diary Study analysis.
- Workflows: Interview Guide drafting. I've gathered input from the Mobile team and I will start to work on incorporating relevant questions into the guide this week. *Recruiting: Remotely observing a candidate presentation this week.