mozilla :: #developers

7 Sep 2017
00:25pulsebotCheck-in: - 47 changesets - Merge m-c to inbound, a=merge
00:25pulsebotCheck-in: - 4 changesets - Merge m-c to autoland, a=merge
00:27pboneAre there any perfherder people awake in my timezone?
00:48pulsebotCheck-in: - Ricky Chien - Bug 1395807 - Waiting for "sync-pane-loaded" event for fixing browser_notification_open_settings intermittent r=jaws
00:53pulsebotCheck-in: - Ricky Chien - Bug 1396583 - Fix wrong tooltip position by wrapping hbox r=jaws
00:59pulsebotCheck-in: - Ricky Chien - Bug 1397121 - Some of Preferences subdialog's font size are too huge r=jaws
01:34pulsebotCheck-in: - Makoto Kato - Bug 1395443 - DictionaryFetcher::Fetch should initializate nsIContentPrefService2 by idle thread. r=masayuki
02:05pulsebotCheck-in: - - Bug 1359055 - Part 1: PBO offset for WebGL compressedTexImage; r=baku,jgilbert
02:05pulsebotCheck-in: - Daosheng Mu - Bug 1359055 - Part 2: Enable compressed texture unpack buffer tests; r=jgilbert
02:05pulsebotCheck-in: - Daosheng Mu - Bug 1359055 - Part 3: Fail-if webgl-compressed-texture-size-limit tests on Win 7; r=jgilbert
02:10pulsebotCheck-in: - Daosheng Mu - Bug 1392476 - Add dropframe telemetry for WebVR; r=francois,kip
02:35dietrichanyone know about cc/gc edge logs? my phone filled up with 8gb of them.
02:35dietrichare they ours? known issue?
02:36dietrichah, yes, they're ours.
02:37dietrichhm, what component to file bug under...
02:38araion desktop, "Save ..." buttons in about:memory generates them. not sure about mobile
02:57philortnikkel: would it be okay to land ? I'm crashing there every few minutes, getting a little tired of it
03:01pulsebotCheck-in: - Bevis Tseng - Bug 1390076 - Label MainThreadInvoker using SystemGroup. r=aklotz
03:02tnikkelphilor: sure go ahead
03:05pulsebotCheck-in: - Zibi Braniecki - Bug 1395871 - Open toolbar menus on mousedown, rather than oncommand.
03:06gandalfugh, forgot to manually add my reviewer while pushing to inbound. Sorry for that!
03:06gandalfgot used to autoland
03:07philorbleah, conflicts
03:11tnikkelphilor: based on top of the patch in bug 1397235 maybe?
03:17pulsebotCheck-in: - Andrew Osmond - Bug 1397235 - Ensure that we reset factor-of-2 mode for an ImageSurfaceCache when it becomes empty. r=tnikkel
03:17pulsebotCheck-in: - Andrew Osmond - Bug 1397223 - Gracefully handle failures in SurfaceCacheImpl::StartTracking. r=tnikkel
03:27pulsebotCheck-in: - Aaron Klotz - Bug 1383501: Assert that com proxy is non-null before calling SendPDocAccessibleConstructor; r=yzen
03:40pulsebotCheck-in: - Nicholas Nethercote - Bug 1393642 (follow-up) - Fix potential leak in HTMLEditor methods. r=masayuki.
03:41philorManishearth: assertions aplenty from
03:53heycamnjn: I think we should measure @import rules (in a followup) since they have their whole imported style sheet hanging off them
03:54njnheycam: sure. Where do they live?
03:55heycamnjn: CssRule::Import(rule), rule is an ImportRule
03:55heycamnjn: and it hangs off the ImportRule's stylesheet field
03:56heycam(that field is actually a ::gecko::data::GeckoStyleSheet, which means is a newtype for a C++ ServoStyleSheet object
03:57njnheycam: hmm, that would require a Rust-to-C++ call?
03:57heycamnjn: yep
03:57njnhaven't had any of those yet; gruesome
03:57njntrouble with this stuff is that almost every new reporter I've added has required some kind of new infrastructure
03:58heycammaybe the imported sheets hang off the ServoStyleSheet children list
03:58heycamI forget how this works
03:59heycamyeah, forget all that, we will find it through the child sheets
04:00heycam(that is the case for Stylo, but I'm not sure for Servo, actually)
04:08pulsebotCheck-in: - Phil Ringnalda - Backed out changeset 118a2b0b07c2 for assertion failures
04:19Manishearthphilor: will look at when i get home
04:20philorManishearth: 'k, backed out, I assume the magic of backing out notifies you in some place that's not a bug that it was backed out
04:21philorI might need to work on the habit of more explicit backout commit messages, I'm used to having a bug to expand on them in
04:31philorglandium: seems too frequent
04:32globphilor: normally it would (a servo PR would be created). unfortunately in this case the PR creation failed due to a merge conflict. urgh
04:36glandiumphilor: backout, I guess
04:37pulsebotCheck-in: - Nicholas Nethercote - Bug 1397579 - Report stylesheet memory usage under "layout/style-sheets". r=heycam.
04:38* philor should really do something about his constant "warning: Watchman unavailable"
04:38philorit's a wonder I don't get midaired more, as often as I have to wait those out
04:39pulsebotCheck-in: - Phil Ringnalda - Backed out 2 changesets (bug 1396723) for !ElementAccess::Get(mHead).mPrev assertion failures
04:46pulsebotCheck-in: - Nicholas Nethercote - Bug 1276488 - Add diagnostics about Prefs startup failures. r=froydnj.
04:49pulsebotCheck-in: - Chris Peterson - Bug 1394603 - Replace fails-if(!styloVsGecko) to fails. r=jryans
04:49pulsebotCheck-in: - Chris Peterson - Bug 1394603 - Replace random-if(styloVsGecko) with skip-if. r=jryans
04:49pulsebotCheck-in: - Chris Peterson - Bug 1394603 - Infer asserts-if(styloVsGecko, X*2) from asserts(X). r=jryans
04:52glandiumroc: fwiw, the rr 5.0 release notes aren't lying about it being faster.
04:53glandiumroc: compared to 4.5 it's night and day for what I was recording this morning
04:53rocif it made your workload a lot faster, maybe I should look at your workload in case we can make it even faster again
04:57pulsebotCheck-in: - Bobby Holley - Bug 1397052 - Don't update |result| after the acquire fence in nsThreadSafeRefCnt::operator--. r=froydnj
04:57glandiumroc: could it be related to the fact that my recorded process had a very large log being written to?
04:58rocI wouldn't think so
04:58roctracee writes are basically ignored by rr
04:59glandiumroc: well then I wasn't doing anything special except running firefox with stylo enabled, forced with 3 styling threads and opening the wikipedia page for obama
04:59glandiumroc: with 4.5 it was a) slow b) unsuccessful
05:00glandium(as in, the content process would crash after a long while)
05:00rocduring recording?
05:00rocthen I dunno
05:00glandiumwith 5.0, it worked like a breeze
05:00rocI just fixed a really crazy bug that was affecting long Firefox replays
05:00glandiumarguably, 4.5 is really old
05:00rocbut recording ... I dunno. We've fixed a lot of bugs since February
05:01rocmy fault for not doing more frequent releases
05:02db48xrunning the tests still takes ages :)
05:02rocI added just support for accelerating getpid and setsockopt syscalls that were showing up in some Firefox traces (since glibc stopped caching getpid) but that shouldn't have made a huge difference
05:02rocdb48x: that's because we keep adding more :-)
05:02db48xwell, no wonder
05:03rocanyway, my work here is done, time to go home
05:06db48xenjoy your evening
05:22db48xalas, 5 test failures
05:27db48x4 of them cpuid related, which is interested but not very critical for me
05:43philorcpeterson: what have you done?
05:47* philor fails to see it
05:49pulsebotCheck-in: - Phil Ringnalda - Backed out 3 changesets (bug 1394603) for reftest syntax error
05:51Manishearthphilor: ping
05:51philorManishearth: pong
05:51Manishearthphilor: i don't see any assertions? just known intermittents?
05:52Manishearthand then leakcheck fails which usually happen in the case of weird intermittent
05:53philorManishearth: and
05:55philorthe crashtest ones weren't just QuantumRender, but that's apparently the only debug crashtest platform we run every push
05:56Manishearthphilor: ah, i see. wasn't showing up on the link you sent
05:57philoroh, I probably used a "hey, look at these things that were unstarred at the very moment I pasted the link" link
05:57Manishearthyeah these look like ok assertions
05:58Manishearthwe have a bunch of layout assertions that switch around as you change rounding behaviors
05:58Manishearthi'll reland
06:05Manishearthphilor: uh.
06:05Manishearthphilor: so it seems the backout never happened
06:05Manishearthphilor: if I close that PR , can you undo the backout?
06:05heycamglob: ^ can you say if that's a good or bad idea
06:05philorManishearth: undo the backout, you mean, "recause the permaorange"?
06:06Manishearthphilor: yes but I'll be ready with patches to fix the orange
06:06Manishearthit's just an expectations file update
06:06Manishearththese are known crashes
06:06Manishearththe problem is that if you change how floats are handled the exact crashtests that hit these crashes change
06:06globManishearth: "undoing backouts" isn't something that's ever been requested to be supported by servo-vcs-sync
06:06Manishearthglob: the PR at has stalled already
06:07Manishearthso we can just close it
06:07Manishearthinstead of having to wait two servo CI cycles
06:07pulsebotCheck-in: - Kris Maglione - Bug 1396856: Part 1 - Remove spread call fallback overhead in event dispatch. r=zombie
06:07pulsebotCheck-in: - Kris Maglione - Bug 1396856: Part 2 - Add top outer window ID to LoadInfo. r=ehsan
06:07globManishearth: how long are you online for?
06:07pulsebotCheck-in: - Kris Maglione - Bug 1396856: Part 3 - Add a WebIDL wrapper class for necko channels. r=ehsan,mixedpuppy
06:07pulsebotCheck-in: - Kris Maglione - Bug 1396856: Part 4 - Update WebRequest.jsm to use ChannelWrapper bindings. r=mixedpuppy
06:07pulsebotCheck-in: - Kris Maglione - Bug 1397536: Avoid newURI overhead for MatchPattern. r=ehsan,mixedpuppy
06:08Manishearthglob: one servo cycle, not two :)
06:08globManishearth: so 45 minutes?
06:09globManishearth: in theory that should work, as long as the autoland repo is put back in sync, and as long as the servo tree remains closed while this is happening. ie. when the servo tree reopens the repos should be in sync
06:09globManishearth: my preference is to let the normal process continue
06:09globManishearth: you're on the hook to fix it up if it goes pear shaped
06:09Manishearthglob: lol
06:10Manishearthglob: it's already backed out on the other side, no?
06:10philorI'm around for about 5 minutes, fwiw
06:10pulsebotCheck-in: - JW Wang - Bug 1395802. P1 - add AbstractMainThread() to MediaResourceCallback. r=cpearce
06:10pulsebotCheck-in: - JW Wang - Bug 1395802. P2 - assert ChannelMediaResource::GetDownloadRate() runs on the main thread. r=cpearce
06:10pulsebotCheck-in: - JW Wang - Bug 1395802. P3 - ensure mChannelStatistics.AddBytes() to happen on the main thread. r=cpearce
06:10pulsebotCheck-in: - JW Wang - Bug 1395802. P4 - we don't need lock since mChannelStatistics is always accessed on the main thread. r=cpearce
06:10globManishearth: yes, it's already backed out of the autoland repo. the backout pr is created as a result of that backout
06:11philorand to the best of my knowledge, the next paid sheriff starts in maybe 10 hours
06:11Manishearthglob: hmm
06:12Manishearthglob: so pushing a revert of that commit to autoland and closing the backout PR in theory should work
06:13globManishearth: as long as servo remains closed until after the pr is cancelled, yes, in theory that should work
06:13Manishearthglob: "until after the PR is cancelled"?
06:13Manishearthyeah, that's the plan
06:13Manishearthok, I'll try this
06:14ManishearthI'm going to keep the servo tree closed (it already is)
06:14globi'm cautious mostly because i just spent a few hours fixing servo-vcs-sync :)
06:14ManishearthI'll be around for an hourish to keep an eye on things
06:15globthanks; i'mm off in about 1.5 hours too
06:15globManishearth: do you need philor for anything?
06:15Manishearthglob: maybe to do the revert of the revert
06:16* glob is mindful he has to go in -1 minutes
06:16Manishearthphilor: okay, I'm unable to push to servo/, could you undo that backout PR?
06:17Manishearther, backout commit
06:17cpetersonphilor: sorry. thanks for backing me out. my change was hit by autoland conflicts twice yesterday before getting a chance to land tonight. I probably should have run on Try again after two autoland conflicts... :|
06:19pulsebotCheck-in: - Manish Goregaokar - servo: Merge #18399 - stylo: Overflow fixes (from Manishearth:stylo-fuzzfix); r=emilio
06:19philorhg-oops++, don't think I've ever backed out a backout, didn't realize it was smart enough to handle the recommit message
06:19Manishearththat's nifty
06:20Manishearthok, so now I can close the PR on the servo side
06:20Manishearthwell, wait, let me land the orange fixes first
06:20Manishearthleave servo closed for a bit
06:21philornice, I figured out where gfx is constantly crashing me. not so nice, it's because I'm obsessively rechecking
06:25pulsebotCheck-in: - Manish Goregaokar - Bug 1397363 - stylo: Add crashtest; r=emilio
06:25pulsebotCheck-in: - Manish Goregaokar - Bug 1397439 - stylo: Add crashtest; r=emilio
06:25pulsebotCheck-in: - Manish Goregaokar - Bug 1397363 - stylo: Update crashtest expectations; r=orange
06:26Manishearththere we go
06:28Manishearthglob: should be ok to reopen servo tree, yes?
06:28Manishearththe backout has been undone, with orange fixes landed, and the servo PR closed
06:28* glob looks
06:28pulsebotCheck-in: - Nicholas Nethercote - Bug 1396694 (part 1) - Make nsIAtom non-scriptable. r=froydnj.
06:29pulsebotCheck-in: - Nicholas Nethercote - Bug 1396694 (part 2) - Replace nsIAtom.idl with nsIAtom.h. r=froydnj.
06:29globManishearth: looks sane
06:35Manishearthalright, reopened tree
06:35Manishearthlet's see what happens
06:51njnxidorn: good idea about moving None into Image, and removing the Either
06:53xidornnjn: I really hope that Rust can be able to do that kind of flatterning as soon as possible...
06:54njnxidorn: sounds complicated; I wouldn't get your hopes up
06:54xidornnjn: I think someone in rust team has idea how to do
06:54njnthe field reordering is nice
06:54xidornnjn: although I guess the impl hasn't started...
06:55xidornnjn: there are lots of places in servo which use deep nested enum, so that should help a lot
06:55xidornand there are people who want to make them even deeper...
07:03pulsebotCheck-in: - Paolo Amadini - Bug 1387416 - Place the search bar in the customization palette for new profiles. r=Gijs
07:18njnheycam: what top-level structure (style sheets, style sets, etc) do the PropertyDeclarations live under?
07:18njnheycam: i.e. where will they show up in about:memory?
07:18heycamnjn: style sheets
07:19njnk, thx
07:20xidornnjn: dom tree can own some of them for style attribute
07:20njnxidorn: ok
07:21heycamand where do mapped attributes live?
07:21heycamI guess they're off the Element somewhere too
07:21njnheycam: my latest patch completely measures the record in, plus a bunch of smaller (though still sizeable) ones
07:21njnheycam: thanks for explaining where that data was
07:21heycamnjn: great, thanks!
07:24xidornheycam: in nsMappedAttributes it seems
07:25freddybq cr
07:25heycamthey might not be as big a deal
07:25freddybeeerr wrong window :-)
07:25heycam(though perhaps for SVG documents they are...)
07:25xidornheycam: so it's owned by dom tree as well
07:36pulsebotCheck-in: - Christoph Kerschbaumer - Bug 1396320: Fix CSP sandbox regression for allow-scripts. r=dveditz
07:45pulsebotCheck-in: - Gijs Kruitbosch - Bug 1396966 - fix dnd in a test to work in small screens, r=mystor
07:50* db48x grumbles
07:50db48x0x00007efe3100a333 in NS_ABORT_OOM (size=30807835312) at /home/db48x/gecko-dev/xpcom/base/nsDebugImpl.cpp:573
07:50db48x(rr) p size
07:50db48x$1 = 30807835312
07:53AryxManishearth: hi, the Rs4 on linux x64 debug needs a fix or expectation update
07:53pulsebotCheck-in: - Nicholas Nethercote - servo: Merge #18400 - Measure selectors (from nnethercote:measure-selectors); r=heycam
07:59pulsebotCheck-in: - Brian Birtles - Bug 1365855 - Drop comment about needing to update SMIL animation styles in ServoRestyleManager::UpdateOnlyAnimationStyles; r=hiro; DONTBUILD
08:07pulsebotCheck-in: - Andrea Marchesini - Bug 791440 - nsGlobalWindow::SetInitialPrincipalToSubject should check the existence of mDoc, r=qdot
08:09Aryxemilio: ^^
08:18pulsebotCheck-in: - Luke Chang - Bug 1394854 - [Form Autofill] Identify country code from "country" attribute as well. r=MattN,steveck
08:25pulsebotCheck-in: - Andreas Farre - Bug 1396368 - Update mTopInnerWindow before unlinking it. r=bkelly
08:25emilioAryx: will look as soon as I arrive to the office, thanks!
08:28xidornemilio: I guess it should be asserts-if(styloVsGecko,3-4)...
08:29xidornit has asserts(1-2)... so actually 2-4 is all possible I suppose
08:34emilioxidorn: can you push it? Not at my laptop r/n
08:34xidornemilio: ok
08:43pulsebotCheck-in: - Xidorn Quan - Bug 1341102 - Bump assertion count on 582146-1.html for styloVsGecko.
08:44xidornAryx: ^
08:48pulsebotCheck-in: - Jared Wein - Bug 1394680 - Apply the white fill color to animatable boxes inside of toolbars that have bright text. We were already doing this for compact themes but missed doing it for lightweight themes. r=ntim
08:55pulsebotCheck-in: - Do Gottwald - Backed out changeset b440aaae26a1 (bug 1383051) for causing bug 1396281
08:55pulsebotCheck-in: - Do Gottwald - Backed out changeset 633d738656d5 (bug 1383051) for causing bug 1396281
08:59pulsebotCheck-in: - Sebastian Kaspari - Bug 1396604 - Remove content notifications code. r=JanH,nechen
09:00pulsebotCheck-in: - L10n Bumper Bot - no bug - Bumping Fennec l10n changesets r=release a=l10n-bump
09:18pulsebotCheck-in: - Jean-Yves Avenard - Bug 1392143 - P1. Change video size used for testing if h264 decoder is working. r=mattwoodrow
09:18pulsebotCheck-in: - Jean-Yves Avenard - Bug 1392143 - P2. Disable hardware decoding for small videos. r=mattwoodrow
09:30pulsebotCheck-in: - Jan de Mooij - Bug 1397026 - Make sure JSOP_INIT* IC behavior matches [[DefineOwnProperty]] instead of [[Set]]. r=anba
09:33pulsebotCheck-in: - Marco Bonardo - Bug 1392081 - Use a document fragment to populate toolbar and menu Places views. r=Gijs
09:33pulsebotCheck-in: - Marco Bonardo - Bug 1392081 - Set the image attribute on bookmarks toolbar buttons only when they are visible. r=Gijs
09:33pulsebotCheck-in: - Marco Bonardo - Bug 1392081 - Only build a subset of the buttons that may become visible on the bookmarks toolbar. r=Gijs
09:33pulsebotCheck-in: - Marco Bonardo - Bug 1392081 - Mochitest browser test for bookmarks toolbar overflow. r=Gijs
09:33pulsebotCheck-in: - Marco Bonardo - Bug 1392081 - Reset the Places toolbar view when the toolbar is collapsed. r=Gijs
09:41pulsebotCheck-in: - Julian Descottes - Bug 1388412 - update devtools addon install rdf to use toolkit em:id;r=ochameau
09:44Aryxxidorn: thank you both
09:45Aryxfarre: hi, windows bustage for bug 1396368
09:45farreAryx: looking
09:46Aryxthanks, use CLOSED TREE if you want to push a folow-up
09:52farredoing it now
09:55pulsebotCheck-in: - Andreas Farre - Bug 1396368 - Follow up fix: Fixed Windows build error. CLOSED TREE
09:55farreAryx: pushed
09:56* farre starts checking docs for how to catch adding new warnings
10:03xidornAryx: you can always ping me or heycam when it is still too early for europe
10:07emilioAryx: do you know when inbound is going to merge to central?
10:07emilioxidorn: thanks for pushing the expectation change before :)
10:10xidornemilio: actually I have a feeling that I haven't pushed expectation update for quite a while
10:27Honzachutten: ping
10:31pulsebotCheck-in: - 97 changesets (bugs 18391, 1394381, 1395227, 1396894, 1396618 and 55 other bugs)
10:39pulsebotCheck-in: - 140 changesets - merge mozilla-inbound to mozilla-central. r=merge a=merge
10:41pulsebotCheck-in: - 141 changesets - merge mozilla-central to autoland. r=merge
10:43pulsebotCheck-in: - 99 changesets - merge mozilla-central to mozilla-inbound. r=merge a=merge
10:45Aryxemilio: done
10:45emilioAryx: awesome, thanks a lot!
10:52pulsebotCheck-in: - Jan de Mooij - Bug 1395919 - Don't copy the unknown-properties flag in AddPropertyTypesAfterProtoChange. r=bhackett
10:55pulsebotCheck-in: - Olli Pettay - Bug 1378092 - Consider to remove MaybeGC from AutoEntryScript::~AutoEntryScript(), r=jonco
10:58jandemsmaug: can we remove the empty dtor now, or do you want to do that as followup if it sticks? :)
11:00Aryxsebastian: hi, linting failure for 1396694
11:03pulsebotCheck-in: - Jan Odvarko - Bug 1397169 - Properly implement openLink; r=rickychien
11:04pulsebotCheck-in: - Jan Odvarko - Bug 1397228 - Properly destroy input field; r=rickychien
11:07whimboomstange: hi. is bug 1395504 also something you could have a look at? I'm not that familiar with the activity monitor output
11:07whimboomstange: it completely locks up the webcontnet process during shutdown, and firefox doesn't quit
11:12mstangewhimboo: the first dump looks like a deadlock between different processes to me
11:12mstangewhimboo: the second dump is confusing; it's taking over two seconds to load a shared library
11:13mstangeI wouldn't trust the second dump, unless you can reproduce this hang stack
11:13whimboomstange: i checked and it looks like it varies a lot
11:13whimbooi'm trying to find a regression range
11:14whimboobut given that everything looks different makes me wonder if I'm still on the right path
11:14mstangea regression range would be the most useful piece of information
11:14whimbooso I will go back that far as I can reproduce the infinite hang
11:14mstangeyes, that sounds reasonable
11:14whimboohowever the stack looks like
11:16sebastianAryx: Argh. I wonder why my try run didn't catch that. I'll fix that.
11:16pulsebotCheck-in: - Drew Willcoxon - Bug 1397609 - Fix a11y of bookmark star and Pocket urlbar buttons. r=Gijs
11:17pulsebotCheck-in: - Mark Banner - Bug 1397545 - Don't update the recently bookmarks item list when it is hidden and a bookmark is deleted. r=mak
11:47pulsebotCheck-in: - Bob Silverberg - Bug 1397081 - Update browserSettings to report the current home page and new tab page in all cases, r=mixedpuppy
11:48pulsebotCheck-in: - Do Gottwald - Bug 1394767 - Log an error and return in unexpected situations rather than calling NS_ASSERT and letting the subsequent code fail. r=mikedeboer
11:51pulsebotCheck-in: - James Graham - Bug 1397215 - Update web-platform-tests to revision 7abe12a95a33cf1906182c1e56da37a32effce3f, a=testonly
11:51pulsebotCheck-in: - Andreas Farre - Bug 1397215 - Update wpt metadata, a=testonly
11:51pulsebotCheck-in: - James Graham - Bug 1397215 - Fix --exclude and --include arguments to wptrunner, r=me
11:51pulsebotCheck-in: - James Graham - Bug 1397217 - Disable WebRTC ontonechange test for instability, a=testonly
11:51jgrahamHaha, oops
11:51pulsebotCheck-in: - James Graham - Bug 1397218,1397219 - Disable unstable WebDriver tests, a=testonly
11:52jgrahamfarre: Apologies, I managed to screw up the commit data for that wpt update so it looks like you did it :)
11:52farrejgraham: noticed, as long as it's awesome I'm claiming it :)
11:52jgrahamfarre: You are going to be so disappointed :p
11:55froydnjfarre: the only problem is that people might start thinking you are the wpt contact point :)
11:55farre... jgraham, you fix that commit right now! ;)
11:55pulsebotCheck-in: - Julian Seward - servo: Merge #18397 - Add fallible append APIs for Vec and SmallVec (from julian-seward1:master); r=Manishearth
11:56pulsebotCheck-in: - Servo VCS Sync - No bug - Revendor rust dependencies
11:56jgrahamfarre: I'll fix it if I get backed out. So pretty good odds I think :)
12:00pulsebotCheck-in: - Marco Castelluccio - Bug 1397505 - Only support dumping/resetting coverage counters with GCC builds. r=froydnj
12:14pulsebotCheck-in: - Olli Pettay - Bug 1368972 - Increase slice time for long-running incremental GCs when CC has tried to run, r=mccr8,jonco
12:20pulsebotCheck-in: - Nathan Froyd - Bug 1396892 - turn off -Wtautological-pointer-compare in memory/build/ for Android; r=njn
12:27pulsebotCheck-in: - 15 changesets (bug 1290481)
12:48Aryxjya: hi, 1392153 triggers an unexpected pass in the gl suite: Will back it out, can you fix it and reland?
12:49jyaAryx: that would be unfortunate...
12:49jyathere's nothing to "fix" per say.
12:50Aryxthe test expectation
12:50jyabut if you want to back it out... Ill change the buggy expectations
12:51ochameaujmaher: hi, it looks like talos not longer run (at least g2-e10s on linux64 try builds)
12:51jyainteresting that HW decoding was also disabled on try machines because of a buggy test
12:51ochameaujmaher: see this from yesterday, is still not executed:
12:52pulsebotCheck-in: - Sebastian Hengst - Backed out changeset acf2ccf34d75 (bug 1392143)
12:52Aryxochameau: Bug 1397674
12:52pulsebotCheck-in: - Sebastian Hengst - Backed out changeset 6c13864a1bee (bug 1392143) for unexpectedly passing gl suite's dom/canvas/test/webgl-mochitest/test_video_fastpath_mp4.html, at least on Windows 7 debug. r=backout
12:52jmaherochameau: yes, I have a few try pushes from yesterday pending on linux talos jobs; it appears our hardware machines for linux are not able to boot up and accept jobs- the problem started somewhere about 21 hours ago
12:53jmaherthanks Aryx
12:54ochameauwill this require new pushes once it is back to normal or will it try to resume the old pending jobs?
12:54jmaherochameau: my understanding is that it will pick up the existing jobs and run them
12:54jmaherochameau: they are just queued up right now
12:55ochameauok, I'll follow that bug anyway, thanks for the update!
12:55jmaher2,022 jobs in the queue
12:55jmaherthanks for asking and for your patience- it looks promising that this will be fixed soon and hopefully the queue cleared out in the next 8-10 hours or so
12:59pulsebotCheck-in: - Jan de Mooij - Bug 1394835 part 4 - Split js::DefineProperty/DefineElement in separate accessor vs data functions. r=anba
13:01pulsebotCheck-in: - Nicolas B. Pierron - Bug 966743 - Inline Array.prototype.push with more than one argument. r=jandem
13:07RyanVMmstange: ping
13:07mstangeRyanVM: pong
13:07RyanVMhey, have you had a chance to look at that OSX logspam by chance?
13:07RyanVMit's showing up on literally *every* OSX test run
13:08mstangeRyanVM: wow
13:08mstangeRyanVM: I was just about to look into that
13:08mstangeI'm currently looking at the other regression from the same patch
13:08RyanVM for example
13:08RyanVMsee how far down that pushes the actual failures from that run
13:11Aryxcan also push it out of the log summary
13:13chuttenHonza: pong
13:14Honzachutten: I had a question related to the new telemetry in DevTools, but I think I solved it and new patch is attached, bug 1396811
13:15pulsebotCheck-in: - Valentin Gosu - Bug 1395938 - Prevent crash when we are missing an IOThread during shutdown r=michal
13:22pulsebotCheck-in: - Stone Shih - Bug 1390044 - Enable the input event queue on nightly. r=smaug
13:22chuttenHonza: Excellent. If you have any further questions, let me know
13:22pulsebotCheck-in: - Chung-Sheng Fu - Bug 1382533 - Disable Presentation API when privacy.resistFingerprinting = true r=arthuredelstein,smaug
13:22pulsebotCheck-in: - Chung-Sheng Fu - Bug 1382533 - Add tests for Presentation API fingerprinting resistance. r=schien,smaug
13:22pulsebotCheck-in: - Ethan - Bug 1396468 - Spoof navigator.oscpu as 'Windows NT 6.1; Win64; x64' when resisting fingerprinting is enabled. r=ehsan
13:23pulsebotCheck-in: - Will Wang - Bug 1357114 - Move the taskbar refresh timer to idle dispatch. r=florian
13:23pulsebotCheck-in: - Tom Ritter - Bug 1394013 Fix forbidden string constant assignment r=froydnj
13:24pulsebotCheck-in: - Tom Ritter - Bug 1395418 Revert a previous patch that disabled std::thread on MinGW because we didn't have it. Now we do. r=cpearce
13:24pulsebotCheck-in: - Eric Rahm - Bug 1393230 - Part 1: Remove remaining string forward declarations. r=njn
13:24pulsebotCheck-in: - Eric Rahm - Bug 1393230 - Part 2: Fix more improper string usages. r=njn
13:24pulsebotCheck-in: - Eric Rahm - Bug 1393230 - Part 3: Convert the xpcom string classes to be templated on char type. r=njn, r=fitzgen, r=sfink