mozilla :: #developers

17 Apr 2017
00:29pulsebotCheck-in: - Mantaroh Yoshinaga - servo: Merge #16451 - Make perspective origin animatable (from mantaroh:animate_perspective_origin); r=hiro,emilio
01:18pulsebotCheck-in: - Mike Hommey - Bug 1356931 - Remove gcc-b2g.manifest. r=sfink
01:48pulsebotCheck-in: - Fischer.json - Bug 1355795 - about:preferences#privacy is broken with browser.cache.offline.enable=false, r=jaws
01:51pulsebotCheck-in: - vincentliu - Bug 1350677 - Add Labeling for gfxFontCache. r=jfkthame
01:59pulsebotCheck-in: - Tobias Schneider - Bug 1313972 - List intersectionobserver_cross_domain_iframe.html as support file in mochitest.ini, r=jet
02:00pulsebotCheck-in: - Tobias Schneider - Bug 1324135 - Wait 300ms after the first IntersectionObserver notification to prevent intermittent test failure, r=jet
02:00pulsebotCheck-in: - Tobias Schneider - Bug 1354163 - Wait until there are intersection records to observe. Prevents an intermittent test failure, r=jet
02:02pulsebotCheck-in: - Cameron McCormack - servo: Merge #16488 - style: add Gecko-only support for sideways-{lr,rl} and deprecated SVG 1.1 writing-mode values (from heycam:wm); r=Manishearth
02:08pulsebotCheck-in: - Oriol - Bug 1356660 - Do not warn when pasting code into the console on unofficial branding, r=bgrins
02:08pulsebotCheck-in: - Paul Bignier - Bug 1356797 - Update ThirdPartyPaths.txt. r=sylvestre
02:27pulsebotCheck-in: - John Lin - Bug 1353459 - handle queueInputBuffer exceptions. r=esawin
02:38pulsebotCheck-in: - Matthew - servo: Merge #16347 - Removing recursion from ComplexSelector (from tictakk:ticbranch); r=emilio
02:38pulsebotCheck-in: - Servo VCS Sync - No bug - Revendor rust dependencies
02:45pulsebotCheck-in: - maliu - Bug 1354973 - Remove view before add to new ViewGroup, r=sebastian
02:52pulsebotCheck-in: - Cameron McCormack - Bug 1356510 - stylo: Update test expectations. r=manishearth
02:59pulsebotCheck-in: - Phil Ringnalda - Backed out changeset 245952e7a2d5 (bug 1355795) for failures in browser_bug795764_cachedisabled.js
03:09pulsebotCheck-in: - Yoshi Huang - Bug 1356165 - Log origin attributes for cookie. r=jduell
03:28pulsebotCheck-in: - Aneesh Agrawal - servo: Merge #16490 - Remove windows gnu part 2 (from aneeshusa:remove-windows-gnu-part-2); r=metajack
03:32pulsebotCheck-in: - Makoto Kato - Bug 1356496 - Don't use nsIDOM* in ConfirmSelectionInBody. r=masayuki
06:09pulsebotCheck-in: - Mantaroh Yoshinaga - Bug 1355344 part 2 - Enable w-p-t of perspective-origin animation. r=hiro
06:48pulsebotCheck-in: - Nevin Chen - Bug 1352004 - Fix doorhangers missing background. r=sebastian
07:07pulsebotCheck-in: - Shawn Huang - Bug 1286717 - Part 1: Expose persist/persisted to StorageManager, r=janv, r=baku
07:07pulsebotCheck-in: - Shawn Huang - Bug 1286717 - Part 2: Add mochitests for persist/persisted functions, r=janv
07:07pulsebotCheck-in: - Shawn Huang - Bug 1286717 - Part 3: Add wpt test cases for persist/persisted functions, r=janv
07:07pulsebotCheck-in: - Shawn Huang - Bug 1286717 - Part 4: Resolve a promise for persist() in private browsing mode, r=janv
07:07pulsebotCheck-in: - Shawn Huang - Bug 1348874 - If origin is an opaque origin, reject promise with a TypeError, r=janv
07:34pulsebotCheck-in: - JW Wang - Bug 1355756. P1 - change the type of MediaData::mDuration to TimeUnit. r=gerald
07:34pulsebotCheck-in: - JW Wang - Bug 1355756. P2 - let VideoData::UpdateDuration() take TimeUnit. r=gerald
07:34pulsebotCheck-in: - JW Wang - Bug 1355756. P3 - let CreateAndCopyData() and its friends take TimeUnit for duration. r=gerald
07:54pulsebotCheck-in: - vincentliu - Bug 1350330 - Make sure IPC channel still Open before sending IPC message. r=dvander
08:08pulsebotCheck-in: - Ting-Yu Lin - Bug 1356971 - Use StyleSetHandle::BackendType in PresShell::Init(). r=bholley
09:03pulsebotCheck-in: - Mantaroh Yoshinaga - servo: Merge #16494 - Make column-count/column-gap/column-width animatable (from mantaroh:animate_column); r=hiro
09:03pulsebotCheck-in: - Michael Brennan - Bug 1350233 - Add MDN links for netmonitor performance analysis. r=gasolin ui-r=bwinton
10:43pulsebotCheck-in: - Boris Chiou - servo: Merge #16496 - stylo: Bug 1341372 - Detect new transitions and let it run (from BorisChiou:stylo/transition/trigger); r=heycam
10:45pulsebotCheck-in: - Boris Chiou - Bug 1341372 - Part 1: Let animation-only restyle include css-transition. r=heycam,hiro
10:45pulsebotCheck-in: - Boris Chiou - Bug 1341372 - Part 2: Add one FFI for TElement::has_css_transitions. r=heycam
10:45pulsebotCheck-in: - Boris Chiou - Bug 1341372 - Part 3: Add FFIs for Triggering transitions. r=birtles,heycam,hiro
10:45pulsebotCheck-in: - Boris Chiou - Bug 1341372 - Part 4: Add a utility method to get AnimationCollection by Element and nsIAtom. r=hiro
10:45pulsebotCheck-in: - Boris Chiou - Bug 1341372 - Part 5: Update mochitest expectations. r=xidorn
10:45pulsebotCheck-in: - Boris Chiou - Bug 1341372 - Part 6: Update css-transitions/reftest-stylo.list. r=hiro
11:34pulsebotCheck-in: - Ted Mielczarek - bug 1355099 - Disable sccache on clang-cl builds. r=mshal
11:34pulsebotCheck-in: - Ted Mielczarek - bug 1354048 - update sccache to rev 7a3847276d05deb564cb84a16b8c551b690aaa3e. r=mshal
11:34pulsebotCheck-in: - Ted Mielczarek - bug 1355961 - add some sccache stats to build_metrics. r=gps
11:46pulsebotCheck-in: - bjorn3 - servo: Merge #16497 - Fix indentation of a `}` (from bjorn3:patch-1); r=nox
13:09RyanVMfroydnj: RE: --enable-debug-rust: "This configure option is not represented in any of our automation configs" is specifically referencing --disable-debug + --enable-debug-rust?
13:18pulsebotCheck-in: - Locke Chen - Bug 1327756 - Netmonitor divides js files into 2 groups when I sort by type. r=Honza
13:34pulsebotCheck-in: - Ryan VanderMeulen - Bug 1357069 - Update pdf.js to version 1.8.229. r=bdahl
13:38pulsebotCheck-in: - 12 changesets (bug 1290951)
13:47pulsebotCheck-in: - Alex Gaynor - Bug 1294641 - whitelist reads from the .app directory in the macOS sandbox r=froydnj,haik
13:52pulsebotCheck-in: - Jon Coppeard - Bug 1354294 - Don't clear extra slots when nuking a proxy to avoid reviving dead compartments. r=billm, a=jcristau
13:52pulsebotCheck-in: - Shu-yu Guo - Bug 1354294 - Preserve IsCallable and IsConstructor when nuking wrappers. r=evilpie, a=lizzard
14:13froydnjRyanVM: right
14:16pulsebotCheck-in: - Tom Ritter - Bug 1355584 Make NSIS an optional requirement for MINGW builds r=glandium
14:19pulsebotCheck-in: - 17 changesets (bugs 16388, 15822, 16466, 1356898, 16451 and 10 other bugs)
14:21pulsebotCheck-in: - 21 changesets - merge mozilla-inbound to mozilla-central. r=merge a=merge
14:24AryxAlex_Gaynor: hi, bustage from 1294641:
14:26Alex_GaynorAryx: This is my first time breaking... well, everything... I assume it gets rolled-back (or un-autolanded, not sure how that works), and then I go fix it?
14:27Aryxyes, will back it out and have to do a central to autoland merge after it, so you should be able to reland that later
14:29Alex_Gaynorcheers, thanks
14:30pulsebotCheck-in: - Sebastian Hengst - Backed out changeset 225683fed1d1 (bug 1294641) for breaking crashtests, reftests and Windows builds. r=backout on a CLOSED TREE
14:31pulsebotCheck-in: - 22 changesets - merge mozilla-central to autoland. r=merge a=merge
14:32pulsebotCheck-in: - 19 changesets - merge mozilla-central to mozilla-inbound. r=merge a=merge
14:39pulsebotCheck-in: - Ricky Chien - Bug 1356786 - Add waitUntil in browser_webconsole_shows_reqs_in_netmonitor.js. r=ntim, a=test-only
14:39pulsebotCheck-in: - L. David Baron - Bug 1351359 - Make nsGridContainerFrame call ReflowInput::SetBResize(true) because of auto-block-size swapping between measuring reflows and regular reflows. r=mats, a=gchang
14:40pulsebotCheck-in: - Masayuki Nakano - Bug 1347835 - NativeKey should dispatch keypress events even if WM_KEYDOWN is processed by IME but followed by printable WM_(SYS)CHAR messages. r=m_kato, a=gchang
14:40pulsebotCheck-in: - Nico Grunbaum - Bug 1241066 - Fix mozRtt always 0 or 1. r=jib, a=gchang
14:40pulsebotCheck-in: - Marco Bonardo - Bug 1355414 - places.sqlite schema migration fails if an application has never used the bookmarks service. r=past, a=gchang
14:40pulsebotCheck-in: - Do Gottwald - Bug 1356183 - Let MousePosTracker use event.clientX instead of event.screenX - window.mozInnerScreenX to avoid flushing layout. r=florian, a=gchang
14:40pulsebotCheck-in: - Shih-Chiang Chien - Bug 1355680 - Set progressMax to -1 when HTTP response head is not available. r=dragana, r=mayhemer, a=gchang
14:40pulsebotCheck-in: - Ehsan Akhgari - Bug 1356755 - Don't pass a null pointer to NotifyDocumentTree(). r=qdot, a=gchang
14:42jimmdoes canada have the day off today?
14:43RyanVMi believe so
14:43RyanVMor not
14:43RyanVMjust last friday
14:44nigelbwoah, RyanVM is sheriffing?
14:44RyanVMlet's not go that far now
14:44jimmthey do not
14:44Aryx has it under "Statutory holidays for federal employees"
14:44jimmaccording to mana they don't
14:47nigelbRyanVM: :)
14:47RyanVMnigelb: long time no see
14:48RyanVMstill at RH?
14:49nigelbI finish a year in 2 weeks
14:49RyanVMnice, congrats :)
14:50pulsebotCheck-in: - Martin Robinson - servo: Merge #16463 - Fix various issues with overflow clipping (from mrobinson:various-overflow-fixes); r=emilio
14:54pulsebotCheck-in: - ffxbld - No bug, Automated HSTS preload list update from host bld-linux64-spot-311 - a=hsts-update
14:54pulsebotCheck-in: - ffxbld - No bug, Automated HPKP preload list update from host bld-linux64-spot-311 - a=hpkp-update
14:54pulsebotCheck-in: - ffxbld - No bug, Automated HSTS preload list update from host bld-linux64-spot-361 - a=hsts-update
14:54pulsebotCheck-in: - ffxbld - No bug, Automated HPKP preload list update from host bld-linux64-spot-361 - a=hpkp-update
14:57pulsebotCheck-in: - ffxbld - No bug, Automated HSTS preload list update from host bld-linux64-spot-342 - a=hsts-update
14:57pulsebotCheck-in: - ffxbld - No bug, Automated HPKP preload list update from host bld-linux64-spot-342 - a=hpkp-update
14:59pulsebotCheck-in: - Lee Salzman - Bug 1357009 - add missing UnscaledFontGDI::GetFontInstanceData. r=jrmuizel
15:08pulsebotCheck-in: - Bob Silverberg - Bug 1342207 - chrome.tabs.onActivated does not fire for new windows, r=kmag
15:11froydnjnigelb: RyanVM never *really* stopped sheriffing
15:13nigelbWell once you're a sheriff, you're always a sheriff.
15:14nigelbThis is why I tend to not visit treeherder.
15:14nigelbThe temptation to star is too strong :)
15:21ehsanmstange: you at the office today?
15:21mstangeehsan: I'm about to go to the office
15:21mstangeehsan: so "yes, starting around noon"
15:21ehsanmstange: great, if you're free over lunch, we should talk about how to create these diagrams that I emailed you about :)
15:21ehsanmstange: assuming you'll get here in 40 mins ;)
15:21mstangeehsan: ok!
15:22ehsani'll let you go now
15:22mstangeprobably not quite, but I'll try
15:22ehsanmstange: I'll wait :)
15:27froydnj"...and alleviate time pressure to ship to reduce the 400-600 patches currently uplifted to Aurora"
15:27froydnjI will believe that when I see it
15:28RyanVMunder the current plan, we're only cutting two weeks off the release cycle
15:29RyanVMso other than "It's Beta and we're more risk-averse" I don't expect much difference either
15:29pulsebotCheck-in: - Masatoshi Kimura - Bug 1356218 - Make screen.colorDepth multi-monitor aware. r=jimm
15:29pulsebotCheck-in: - Masatoshi Kimura - Bug 1356218 - Fix nsDeviceContext::GetDepth to use the information from the correct monitor. r=jfkthame
15:31philoroh, it's a longer beta cycle?
15:31RyanVMphilor: current incarnation is 8w
15:31RyanVMwould like to eventually get it down to 6w
15:32philorI never really look at the dates, just copy-paste them into scary bug summaries, didn't notice that part
15:32RyanVMit's not exact either as some dates had to get moved around for holidays and such
15:35pulsebotCheck-in: - Jeff Muizelaar - Bug 1356689. Add pref to highlight painted layers. r=mchang
15:52pulsebotCheck-in: - Zibi Braniecki - Bug 1356718 - Query gtk settings schemas to avoid crashing when schema is missing. r=jfkthame
15:56pulsebotCheck-in: - Kartikaya Gupta - Bug 1356695 - Don't expose DOM touch APIs with autodetection if APZ is disabled. r=smaug
15:56pulsebotCheck-in: - Kartikaya Gupta - Bug 1356695 - Ensure gfxPrefs is initialized if gfxPlatform::AsyncPanZoomEnabled is called early in startup. r=milan
16:30pulsebotCheck-in: - Paul Rouget - servo: Merge #16498 - Properly modify scroll_location (from paulrouget:scroll_lock); r=jdm
17:02pulsebotCheck-in: - David Keeler - bug 1352262 - make OCSP timeout values configurable r=Cykesiopka,jcj
17:06dmajorgps: watchman always times out the first few times I use it after reboot (on windows, if it matters). is there anything I can do to avoid it?
17:09gpsdmajor: `hg help -e fsmonitor` documents a timeout you can increase
17:09gpstimeout on first use is expected: it needs to walk the filesystem to get a consistent state
17:09gpsunless you want to prime watchman on startup somehow, i just deal with the timeout
17:10mconleysheppy: ping
17:11dmajorgps: thanks
17:15sheppymconley: pong
17:15mconleysheppy: hey - do you know if it's possible to embed an AirMo video in an MDN page?
17:15sheppymconley: It is.
17:15mconleyI know we can do that with YouTube via the {{EmbedYouTube}} thinger
17:16sheppymconley: I haven't done it recently, but yeah, it can be done.
17:16sheppyJust has to be done more manually.
17:17mconleysheppy: is there an example I can look at that you're aware of?
17:17sheppymconley: Good question, let me see if I can think of one
17:21pulsebotCheck-in: - Josh Matthews - servo: Merge #16489 - Show HTTP response headers in debug output (from jdm:http-response); r=asajeffrey
17:22sheppymconley: All you need to do is edit the source and put in a <video> element pointed at the video.
17:22sheppyThat&#39;s the easiest way. We should add a macro for it.
17:23sheppyHm, that should work... but doesn&#39;t seem to be.
17:23sheppyThere we go
17:23mconleysheppy: how&#39;d you do it?
17:24sheppyJust clicked the source button in the editor and inserted something like <video src=&quot;url&quot; controls&quot;. Obviously you probably want to use <source> and offer different formats, but the basics work there.
17:33pulsebotCheck-in: - Kris Maglione - Bug 1357137: Make sure all content scripts for the same page load into the same sandbox. r=mixedpuppy
17:34* John-Galt uses HTML mode by default on MDN.
17:45katsKWierso: could you do a merge from graphics to m-c?
17:45KWiersosounds like something I can do
17:45katsKWierso: thanks :) graphics tip should be pushable directly to m-c
17:48pulsebotCheck-in: - 13 changesets - Merge m-c to graphics
17:48katsonly 13 changesets? man we&#39;ve been slacking
18:16* froydnj watches hang trying to get profiler data :(
18:20pulsebotCheck-in: - Aaron Klotz - Bug 1322532: Add support for live regions to the com handler dll; r=tbsaunde
18:20pulsebotCheck-in: - Aaron Klotz - Bug 1322532: Move a11y retrieval of native window handle to DocAccessibleChild; r=yzen
18:20pulsebotCheck-in: - Aaron Klotz - Bug 1322532: Add sync text event to PDocAccessible; r=tbsaunde, r=billm
18:21pulsebotCheck-in: - Aaron Klotz - Bug 1322532: Platform a11y changes to enable handler-based live regions; r=tbsaunde
18:25pulsebotCheck-in: - Mason Chang - Bug 1351733. Reuse D2D Source Surfaces with Draw Target Dual. r=bas
18:36froydnjmconley: re: ,
18:36froydnj(SFW, FYI)
18:36mconleyfroydnj: you said it
18:36* mconley adds to repetoire
18:36mconleyer, repertoire
18:40pulsebotCheck-in: - Mike Conley - Bug 1350324 - Be more rigorous about ignoring the initial about:blank load for tab progress listeners. r=Gijs
18:41rhuntfroydnj: ping
18:41froydnjrhunt: pong
18:42rhuntfroydnj: I&#39;m looking into integrating a rust tool for generating c bindings for a rust library into the build process for webrender
18:43rhuntfroydnj: and it seems like it&#39;s possible to vendor it into the tree and compile it as a standalone, but I don&#39;t see a way to run it to generate headers at the right time
18:43RyanVMmconley: what&#39;s the urgency on ^? Does it need to hit 54 before the uplift to Beta tomorrow?
18:44rhuntfroydnj: I don&#39;t know the build system very well, do you have any ideas on how to accomplish this?
18:44mconleyRyanVM: which ^? Bug 1350324? Bug 1357169?
18:44RyanVMmconley: you&#39;re too busy :P Bug 1357169
18:44mconleyRyanVM: I don&#39;t think it needs to hit 54 before uplift, no.
18:46sfinkrhunt: fitzgen is doing something similar to this to generate bindings for spidermonkey. You might ask him.
18:47froydnjrhunt: my suggestion would be to make all the bindgen machinery callable from a build script; that&#39;s what happens for the Servo bindgen-y stuff
18:47froydnjrhunt: oh, wait, you probably want to include these generated headers in the .cpp files for building Gecko, don&#39;t you...?
18:48rhuntfroydnj: yeah, the headers will be used in cpp files in Gecko :)
18:48froydnjbleh, ok
18:49* froydnj thinks
18:49fitzgenrhunt: out of curiosity, why use cbindgen instead of rusty-cheddar?
18:51froydnjrhunt: the quick-and-dirty solution would be to make this generated header a GENERATED_FILES thing, and then write a short python script that would call cbindgen (and maybe compile cbindgen, ugh)
18:51pulsebotCheck-in: - Thomas Wisniewski - Bug 1334266 - Add support for WebExtension APIs browser.sessions.{forgetClosedTab|forgetClosedWindow}. r=bsilverberg, r=mixedpuppy
18:52katsfitzgen: rusty-cheddar didn&#39;t work for us
18:53froydnjkats: why not?
18:54katsfroydnj: i don&#39;t recall the specifics. but we do a bunch of stuff with passing pointers and generics and things that it didn&#39;t seem to handle
18:55rhuntfroydnj, fitzgen: i didn&#39;t start the project so i don&#39;t know the exact answer either, but it doesn&#39;t look like rusty-cheddar support searching extern crates or mods which is a requirement for us too
18:56rhuntfroydnj: okay, I think that would work
18:58pulsebotCheck-in: - Andrew McCreight - Bug 1356799, part 1 - Use one line per arg for methods with many args in subscript loader. r=mrbkap
18:58pulsebotCheck-in: - Andrew McCreight - Bug 1356799, part 2 - Don&#39;t use JS:: for handle types in subscript loader. r=mrbkap
18:58pulsebotCheck-in: - Andrew McCreight - Bug 1356799, part 3 - Handlify the target argument to ReadScript{,Async}. r=mrbkap
18:58pulsebotCheck-in: - Andrew McCreight - Bug 1356799, part 4 - Remove unused argument to ReadScript. r=mrbkap
18:59* froydnj wonders if we shouldn&#39;t have some dummy rust library, built during export or somesuch, that we can stuff random things into its
18:59froydnjted: ^ scrollback may be of interest to you
18:59pulsebotCheck-in: - Michael Layzell - Bug 1168169 - Mark nsThreadShutdownContext::joiningThread as MOZ_UNSAFE_REF, r=froydnj
18:59pulsebotCheck-in: - Michael Layzell - Bug 1352281 - Include KIND_OTHER types in DataTransfer.types, r=baku
19:03tedfroydnj: oh, mystor was looking at doing .rs file generation via GENERATED_FILES
19:04tedi think we determined it worked, but the build system needed a tweak to put .rs file generation in export
19:04tedfor things that are gecko-specific that should be fine
19:04froydnjted: so this is the other direction, generating .h files from Rust code =/
19:04froydnjted: requiring running Rust
19:05froydnjthat was my reaction too!
19:05tedthen yeah, we certainly don&#39;t have any sensible way to do that right now
19:05mystorfroydnj: I don&#39;t think that a dummy crate is the right way
19:05mystorfroydnj: We definitely don&#39;t need the full build script infrastructure IMO
19:05mystorfroydnj: I would probably have some sort of `export` crate which we run during the export phase which can perform codegen
19:07mystorfroydnj: Perhaps allow for to declare GENERATED_FILES with a Cargo.toml as its target, where that Cargo.toml has a binary target. We would build that binary and run it during the export phase before running C++ code with some environment variables set like they would be in a build script?
19:07tedmystor: i feel like you&#39;d just wind up reinventing all the stuff eventually
19:07mystorted: You&#39;d reinvent some of it, but basic `` stuff isn&#39;t really all that useful?
19:07froydnjmystor: my idea was to at least consolidate all the crates that you might need to build, so you wouldn&#39;t build them multiple times
19:07mystorted: It&#39;s just a rust binary which is run by cargo with specific environment variables set
19:08froydnjI guess if you stuffed everything into this one binary it has the same effect
19:08mystorfroydnj: For that you can use the same trick which we use for libgkrust and libgkrust-gtest
19:08mystorfroydnj: Which would be to have the different binaries share a target dir
19:08* ted bbl
19:09mystor(you could also have the crates which you declare in the file be libraries which are pulled in as dependencies of an implicit export-phase-rust-code crate)
19:09froydnjmystor: eh, that means their compilation has to be serialized =/ (which may or may not be a problem)
19:10mystorfroydnj: true - I kinda like the idea that you just define a crate which exports a &quot;build()&quot; method which we invoke for you
19:10mystorfroydnj: from a generated crate
19:11froydnjmystor: true. otoh, there&#39;s some value in adhering to the wider conventions of the Rust ecosystem
19:11mystorfroydnj: is not really intended to generate C header files
19:12mystorfroydnj: It&#39;s usually intended to preprocess or generate code which will be used for the compilation of one specific crate
19:12mystorfroydnj: It seems a bit silly to create one/multiple dummy crate(s) which won&#39;t actually be used, but will have their scripts invoked
19:15pulsebotCheck-in: - Michael Layzell - Bug 1356277 - Part 1: Respect ALLOW_FIRST_PARTY_ONLY and LIMIT_THIRD_PARTY permissions in StorageAllowedFor*, r=ehsan
19:15pulsebotCheck-in: - Michael Layzell - Bug 1356277 - Part 2: Add a test for LIMIT_THIRD_PARTY permission with sessionStorage, r=ehsan
19:17froydnjkats: maybe moz-cheddar has fixes that you would need
19:21katsfroydnj: i was not aware of moz-cheddar
19:22* kats read backscroll
19:24katswell, the &quot;where to generate bindings&quot; problem will exist regardless of which tool we&#39;re using to generate the bindings
19:24katsso perhaps in hindsight we might have been able to use moz-cheddar, but we have cbindgen working well now
19:38pulsebotCheck-in: - Kris Maglione - Bug 1356891: Get rid of getAPILevelForWindow. r=mixedpuppy
19:41pulsebotCheck-in: - 24 changesets - Merge autoland to central, a=merge
19:44pulsebotCheck-in: - Michael Layzell - backout bug 1356277 due to failing test_localStorageSessionPrefOverride.html on try
19:56pulsebotCheck-in: - 22 changesets - Merge inbound to central, a=merge
20:01pulsebotCheck-in: - Kearwood Gilbert - Bug 1356387 - Add constructor to VRHMDSensorState
20:04pulsebotCheck-in: - 39 changesets - Merge m-c to inbound, a=merge
20:05pulsebotCheck-in: - 37 changesets - Merge m-c to autoland, a=merge
20:07RyanVMright on, that&#39;s a message that can&#39;t be harped on often enough :)
20:15rhuntfroydnj, mystor: is there a consensus on a preferred approach? if there&#39;s a better way to implement this and it&#39;s not much work, I think I&#39;d like to go that route
20:16mystorrhunt: What&#39;re you currently thinking of as your approach?
20:17rhuntmystor: I&#39;ve been prototyping vendoring the code in, using GENERATED_FILE to run a python script which will build the tool and run it
20:17rhuntwhich I think will work
20:18rhuntI&#39;m running into a problem with mach vendor rust and multiple versions of packages, but I think it&#39;s unrelated
20:18pulsebotCheck-in: - Kirk Steuber - Bug 1356156 - Increase test timeout for Flash block tests r=francois
20:19mystorrhunt: Off of the top of my head that might be the easiest way to implement it right now - but I would probably want to do some build system work to get better tooling for this.
20:21pulsebotCheck-in: - Imanol Fernandez - servo: Merge #16499 - Update servo-glutin (from MortimerGoro:update_glutin); r=Wafflespeanut
20:21rhuntmystor: ok, that&#39;s fine with me, thanks!
20:39pulsebotCheck-in: - Andrew Swan - Bug 1356027 Add pref to disable non-MPC extensions r=mossop
20:49mstangefroydnj: are there plans for adding an rwlock to mfbt?
20:49pulsebotCheck-in: - Gregory Moore - Bug 1180799 - Support momentum scrolling after two-fingered pans on pages that don&#39;t allow zooming. r=botond,kats
20:49mstangethe profiler could really use one
20:49mstange(or do we already have one?)
20:50mstangeI see one in nspr, in WebRTC, in Skia, and in a bunch more 3rd party code
20:50froydnjmstange: there are no plans that I know of (it&#39;d probably go in mozglue if we added one)
20:51froydnjmstange: Manishearth was just complaining the other day that we didn&#39;t have such basic things, though :)
20:51mstangefroydnj: ah
20:55pulsebotCheck-in: - Alex Gaynor - Bug 1294641 - whitelist reads from the .app directory in the macOS sandbox r=froydnj,haik
21:02pulsebotCheck-in: - Felipe Gomes - Bug 1351933 - Remove support for whitelisting addons for the e10s rollout. r=krizsa
21:04billmjchen: ping
21:04jchenbillm: pong
21:05billmjchen: hi. I have a leak I&#39;m trying to track down on android. somehow a patch I wrote is leaking an nsIAndroidEventListener object. it looks like these are stored in widget/android/EventDispatcher.cpp.
21:05billmjchen: who calls into that code? I&#39;m trying to figure out how stuff gets unregistered.
21:09jchenbillm: quite a bit of js. e.g. if you search &quot;EventDispatcher.registerListener&quot; or &quot;EventDispatcher.instance.registerListener&quot;
21:10jchenbillm: if you can find the event type obviously that&#39;d be a lot easier
21:10billmjchen: how would I get that?
21:11billmjchen: oh, I guess the hashtable here would have it?
21:11jchenbillm: yeah. or log calls to EventDispatcher::RegisterListener
21:11billmjchen: ok, thanks. let me try.
21:15pulsebotCheck-in: - Sam Foster - Bug 1334642 - Cache luminance values for each toolbar in ToolbarIconColor. r=dao,mconley
21:25billmjchen: I think it&#39;s this one:
21:25billmjchen: how would that normally get unregistered?
21:27jchenbillm: we don&#39;t unregister that currently. you can try unregistering it on xpcom shutdown or something?
21:28billmjchen: I&#39;m just wondering why we weren&#39;t leaking before my patch. is there anything that frees EventDispatchers on shutdown or something?
21:28jchenbillm: what does your patch do?
21:29billmjchen: it refactors a bunch of things in XPConnect
21:29billmjchen: the &quot;leak&quot; is actually an assertion firing because an nsXPCWrappedJS object exists at shutdown when we don&#39;t expect it to
21:30billmjchen: and, oddly, this happens in a content process that we spin up during mochitests
21:32jchenbillm: i think we should limit that sessionstore code to the main process
21:32billmjchen: yeah, that makes sense. but I really only care about how my patch could possibly be changing this. it seems really odd.
21:33billmjchen: maybe I&#39;ll do a build without my patch and see where this stuff gets deleted
21:33billmjchen: is there any way to get a stack track on android?
21:36jchenbillm: it&#39;s hard if you&#39;re in c++
21:36billmjchen: is there any sort of C++ debugger?
21:36jchenbillm: you can use gdb
21:36billmjchen: how do I set it up to talk to them emulator?
21:38jcheni think mach run --debug should work to a certain degree
21:38billmjchen: unfortunately I need it to talk to the content process
21:39billmjchen: I guess it would be much better if I could just dump out a stack trace. there&#39;s no way to do that?
21:39jchennot really...
21:39pulsebotCheck-in: - Martin Robinson - servo: Merge #16336 - Used shared clips for overflow:hidden and CSS clip (from mrobinson:overflow-hidden); r=glennw
21:40jchenif you moz_crash without breakpad, android may print out a stack trace
21:40jchenbillm: also seems like that sessionstore code always gets called in content processes. maybe your patch let it run successfully where it failed before?
21:41billmjchen: hmm, that&#39;s possible
21:41billmjchen: I can get a backtrace with