mozilla :: #developers

17 Jul 2017
00:07pulsebotCheck-in: https://hg.mozilla.org/mozilla-central/rev/aff336ac161d - Nicholas Nethercote - Backed out changeset cb1548354e0c (bug 1352559, part 3) for causing *many* crashes in Nightly. r=backout
00:08njnok, backout done
00:08njnif anyone knows how to retrigger Nightly builds, please let me know
01:18* njn has never seen IRC this dead before on a Monday
01:33kanruI thought one need special power to spin nightly.
01:45pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/050a2bc83e57 - Kartikaya Gupta - Bug 1381110 - Remove the unused ContainerLayerParameters::mInLowPrecisionDisplayPort field. r=mattwoodrow
01:46njnkanru: apparently sheriffs can do it. I asked in #sheriffs, but nobody is around :/
02:31pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/974788a46fa3 - JW Wang - Bug 1376004. P1 - include http response code in the error.message field so the UI has more details to display a proper error message. r=cpearce
02:33pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/2dcd3f591a14 - Emilio Cobos lvarez - servo: Merge #17751 - style: Fix a typo that makes us resolve styles for the whole spin to the root (from emilio:embarrasing-typo); r=heycam
04:46pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/a31143203b44 - Kearwood "Kip" Gilbert - Bug 1381165 - Ensure that only the first VRDisplay.submitFrame call per VRDisplay RAF callback renders and kicks off the next VR RAF callback. r=daoshengmu
04:49pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/d97dc1dbc289 - Kearwood "Kip" Gilbert - Bug 1321275 - Fix reference counting of PVRLayerChild when GPU process has been terminated r=daoshengmu
05:42pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/2fee42d106b3 - vincentliu - Bug 1367689 - Add a preference to make Wrench work with Gecko. r=jrmuize, jerry
05:47pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/b27e73bea124 - Luke Chang - Bug 1352324 - (Part 1) Retrieve PhoneNumber utility JSMs from ad87ed1bbed5. r=MattN
05:47pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/5ebb632fa7bb - Luke Chang - Bug 1352324 - (Part 2) Update the metadata for PhoneNumber utility to v8.4.1. r=MattN
05:47pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/e27d693ddcc4 - Luke Chang - Bug 1352324 - (Part 3) Update PhoneNumber utility to fulfill Form Autofill. r=MattN
05:47pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/27971d8110df - Luke Chang - Bug 1352324 - (Part 4) Fix eslint errors of PhoneNumber utility. r=MattN
05:47pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/b69772478582 - Luke Chang - Bug 1352324 - (Part 5) Add unit tests for PhoneNumber.jsm and PhoneNumberNormalizer.jsm. r=MattN
05:47pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/7aec8351b1db - Luke Chang - Bug 1352324 - (Part 6) Store tel in E.164 format and calculate its components while reading. r=MattN
05:56dbaronnjn: hmmm, it didn't *used* to need special power... I think it was doable from the BuildAPI interface -- but not sure if that's changed with taskcluster
05:57dbaronnjn: at least the UI is still there on BuildAPI (which you can get to from the [v] next to any revision in treeherder)
05:57njndbaron: at this point we're only ~4 hours from the next scheduled build anyway, probably not worth the effort
05:58njndbaron: interesting, the Build API page has "Create new nightly builds on mozilla-central revision"
05:59* dbaron is of course still in a different state of Monday-ness than njn
06:00njnnot-Monday?
06:00dbaron:-)
06:00dbaronstill 59 minutes away
06:01globslacker :)
06:02njnglob: you're setting yourself up for ridicule by Kiwis
06:02dbaronyou should see how quiet things are at 4pm on a Friday afternoon here..
06:03globnjn: dammit
06:10pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/5b29bf0b577c - David Anderson - Enable Advanced Layers for Nightly users on Windows 7. (bug 1379731, r=milan)
06:17pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/1c1a86e1296b - Jessica Jong - Bug 1366361 - Part 1: .action/formAction should return the document's URL if @action/formaction is missing or empty. r=smaug
06:17pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/8f7663ac2068 - Jessica Jong - Bug 1366361 - Part 2: Update reflectURL() and web platform tests. r=smaug
06:34dbaronnjn, glob: we should hire somebody on Kiritimati just for the timezone... :-P
06:36heycamdbaron: huh, I didn't know there were two Christmas Islands.
06:37dbaronheycam: https://en.wikipedia.org/wiki/Christmas_Island_(disambiguation)
06:37dbaron(you were asking for it!)
06:37heycamok, more then two!
06:37globinteresting
06:38dbaronhonestly, I'm surprised the wikipedia disambiguation page is that short
06:38heycamso not only does Australia have a Christmas Island as an external territory, but it's also an island that's part of Tasmania
06:58pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/93af9545a6f4 - cku - Bug 1377329 - Enlarge the default capacity of nsDisplayBuilfer::mFramesMarkedForDisplayList. r=mattwoodrow
06:59pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/fa924e4258af - Emilio Cobos lvarez - Bug 1381323: Account for text frames in ExpectedOwnerForFrame. r=heycam
07:18whimbootnguyen: hi. you cannot land your patches via mozreview on autoland?
07:20pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/8e5252bec8f7 - JW Wang - Bug 1380569. P1 - rename CannotDecryptWaitingForKey() to NotifyWaitingForKey() and add the function to MediaDecoderOwner. r=gerald
07:20pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/3e0ca8c329ac - JW Wang - Bug 1380569. P2 - plumb the 'waitingForKey' event into the media element. r=gerald
07:20pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/4046631bac69 - JW Wang - Bug 1380569. P3 - remove AbstractMediaDecoder::NotifyWaitingForKey() and its callers. r=gerald
07:21pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/994d25c6390b - JW Wang - Bug 1380574 - remove the call to MediaDecoderOwner::DecodeWarning() from MediaFormatReader. r=gerald
07:26pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/9395f52d60ac - Thomas Nguyen - Bug 1374268 - Increase waiting time for all safebrowsing files being downloaded r=francois
07:27whimbootnguyen: so i pushed it for you
07:48pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/d6286a7f8094 - Alfredo Yang - Bug 1380468 - don't fallback when table size is too large. r=kinetik
07:48pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/294935222edd - Alfredo Yang - Bug 1380468 - update rust parser. r=kinetik
08:01pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/fdf32120816a - Paul Rouget - servo: Merge #17684 - allow_navigation: use channel to send response (from paulrouget:fix_allow_navigation); r=paulrouget
08:16igoldanhi
08:16igoldanhave there been any changes to the Windows builds?
08:16igoldanspecifically the PGO builds?
08:16pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/bc861f344831 - Paul Adenot - Bug 1312623 - Don't attempt to shutdown a thread in the dtor of a GraphDriver, since it might be called during stable state. r=smaug
08:17igoldanI'm seeing performance regressions on Windows7 PGO
08:17igoldanwhich are totally unrelated to other platforms
08:37tnguyenwhimboo: oh, thanks you
08:37whimbootnguyen: usually you should be able to push via mozreview or don't you have level 3 access?
08:38tnguyenI don't have level 3 access
08:56pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/4da0146f568a - Raajit Raj - Bug 1368713 - Add test coverage for parsing historical Histograms.json formats r=Dexter
08:59PikeI'm failing to find a way to just run this one test: https://dxr.mozilla.org/mozilla-central/source/xpcom/tests/gtest/TestTextFormatter.cpp
09:20pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/164b77a5a9d9 - Do Gottwald - Bug 1380966 - Increase the size of the tab overflow indicator. r=daleharvey
09:21marcokanru: ping
09:25pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/eb92b29f5500 - Jon Coppeard - Bug 1380387 - Only repeat shutdown GCs for removed roots r=sfink
09:25pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/baa7949eb81f - Jon Coppeard - Bug 1380967 - Add a separate reason code for minor GC performed at the start of major GC r=sfink
09:25pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/19fd4dedb303 - Emilio Cobos lvarez - servo: Merge #17731 - style: Respect calc for percentages (from emilio:percentage-calc); r=nox
09:26kanrumarco: pong
09:27marcokanru: I was wondering if there's a way to have sync messages from parent to child
09:27marcokanru: for the code coverage stuff, so only during testing
09:27marco(so it can't be a perf problem)
09:32kanrumarco: you can try 'intr' message
09:34kanrumarco: that will change the impl function name to Call* and Answer* instead of Send*/Recv*
09:35kanruI'm not sure if you can use that with PContent though
09:47marcokanru: okok, thanks, I'll try
09:57pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/4d1a13d4e94f - Jon Coppeard - Bug 1381058 - Split full store buffer reasons into a separate reason for each store buffer r=sfink
09:57pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/d0c73b22611c - Jon Coppeard - Bug 1380778 - Fixup shape pointers into the nursery using a dedicated list r=sfink
10:01pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/fbe95017106a - Jon Coppeard - Bug 1380778 - Fix review comment I missed r=sfink
10:21pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/81572281c073 - Emilio Cobos lvarez - Bug 1380918: Update expectations. r=me
10:26pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/59252c9ed421 - Henrik Skupin - Bug 1381339 - Re-enable test_navigation.py TestBackForwardNavigation.test_frameset. r=ato
10:56pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/f32a1cf344aa - Mayank Srivastav - Bug 1354032 - Remove Places analyze call and use PRAGMA optimize(0x02) for optimizations. r=mak
11:12pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/522ed11f6fc1 - Mark Banner - Bug 1380801 - Add eslint-plugin-mozilla's node_modules to the ignore path. r=ahal
11:48pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/8df92cb52eb4 - Lie Ryan - Bug 1361855 - Refactor browser/modules/test/browser/head.js:getHistogram() and getKeyedHistogram(); r=Gijs
11:48pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/14d02c0af528 - Lie Ryan - Bug 1361855 - Added TAB_COUNT to Histograms.json; r=bsmedberg
11:48pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/4397e17fc5fc - Lie Ryan - Bug 1361855 - Add helper function getTabCount() in BrowserUsageTelemetry.jsm; r=Gijs
11:48pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/6bfc9384ed64 - Lie Ryan - Bug 1361855 - Add helper function BrowserUsageTelemetry:_recordTabCount(); r=Gijs
11:48pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/2d6877c0b84e - Lie Ryan - Bug 1361855 - Record TAB_COUNT histogram during TabOpen event; r=Gijs
11:48pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/06dbf8fc7da3 - Lie Ryan - Bug 1361855 - Update tab count on page load and other URL changes; r=Gijs
11:48pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/7e9ac96978ad - Lie Ryan - Bug 1361855 - Filter TAB_COUNT histogram to record only on 5 minute intervals; r=Gijs
11:48pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/ede5862f0950 - Gijs Kruitbosch - Bug 1361855 - fix eslint issue, rs=eslint,me,firebot,etc
11:49pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/e4dc716c401c - Nicolas Silva - Bug 1377595 - Avoid using blob images for items that must be painted on the content side. r=jnicol
11:49pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/59479f0872ec - Nicolas Silva - Bug 1377595 - Track the valid region in WebRenderPaintedLayer. r=jrmuizel
11:49pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/ea065f577c1c - Nicolas Silva - Bug 1377595 - fuzz a test. r=jrmuizel
11:49pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/e8ee58818e46 - Nicolas Silva - Bug 1377595 - Mark four tests as passing with webrender.
12:05pulsebotCheck-in: https://hg.mozilla.org/mozilla-central/pushloghtml?startID=32188&endID=32189 - 22 changesets - merge autoland to mozilla-central. r=merge a=merge
12:06pulsebotCheck-in: https://hg.mozilla.org/mozilla-central/pushloghtml?startID=32189&endID=32190 - 7 changesets - merge mozilla-inbound to mozilla-central. r=merge a=merge
12:07pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/pushloghtml?startID=46568&endID=46569 - 12 changesets - merge mozilla-central to autoland. r=merge a=merge
12:08pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?startID=92621&endID=92622 - 27 changesets - merge mozilla-central to mozilla-inbound. r=merge a=merge
12:14pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/71526d1e9db8 - Mark Banner - Bug 1380956 - Build the node_modules directory for ESLint on the docker image for the worker, rather than downloading it each time. r=ahal
12:14pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/576eaf084c0b - Mark Banner - Bug 1380956 - Add a minimum version requirement for npm in ESLint setup and improve the version handling. r=ahal
12:14pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/74d387e1e050 - Mark Banner - Bug 1380956 - Don't manually install our local eslint plugins as npm will do that for us. r=ahal
12:14pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/23a1032b57c9 - Mark Banner - Bug 1380956 - Use exact matches of versions in eslint-plugin-mozilla for better tracking. r=ahal
12:15pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/962c95ec8881 - Mark Banner - Bug 1380956 - List only the directly required dependencies for ESLint in the top-level package.json. r=ahal
12:33pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/5d9a6384bf51 - Bobby Holley - Bug 1378207 - Stop bypassing the Xray layer when walking the prototype chain. v2 r=krizsa
13:37pulsebotCheck-in: https://hg.mozilla.org/releases/mozilla-beta/pushloghtml?startID=7513&endID=7514 - 17 changesets (bugs 1364329, 1365791, 1381340, 1378061, 1377738 and 9 other bugs)
13:38pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/8843631f2926 - Andrew Osmond - Bug 1380341 - Fallback on the document's channel to check private browsing status if the load group has no load context. r=baku
13:54pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/5a280a5cc9a0 - Jean-Yves Avenard - Bug 1355048: P1. Have WebrtcMediaDataDecoder placeholder. r=jesup
13:54pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/24bcb56d3213 - Jean-Yves Avenard - Bug 1355048: P2. Remove OMX h264 decoder. r=jesup
13:54pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/212bf3d3538c - Jean-Yves Avenard - Bug 1355048: P3. Remove ImageHandle. r=jesup
13:54pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/e7f1a5ab2282 - Jean-Yves Avenard - Bug 1355048: P4. Remove unnecessary test, and make better use of monitor. r=jesup
13:54pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/655f161c7d17 - Jean-Yves Avenard - Bug 1355048: P5. Remove RenderVideoFrame virtual method. r=jesup
13:54pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/d0237f00a291 - Jean-Yves Avenard - Bug 1355048: P6. Don't use unnecessary reentrant monitor. r=jesup
13:54pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/49ea8c43becf - Jean-Yves Avenard - Bug 1355048: P7. Use Image's dimensions when available. r=jesup
13:54pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/6fa1a719ca74 - Jean-Yves Avenard - Bug 1355048: P8. Implement WebrtcMediaDataDecoderCodec. r=jesup
13:55pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/f763d7564c56 - Jan de Mooij - Bug 1380961 - Fix SimpleArithOperand to return false if the TypeSet is empty. r=nbp
14:10pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/8321474ca566 - Simon Sapin - servo: Merge #17727 - Upgrade to rustc 1.20.0-nightly (ab91c70cc 2017-07-14), use non-"alt" std (from servo:rustup); r=nox
14:18pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/86ee390b7596 - Mark Banner - Bug 1292426 - Rewrite browser_visituri_privatebrowsing_perwindowpb.js to use modern async facilities and ensure we await on the final check to avoid intermittents. r=mak
14:20Gijshow do I get a stacktrace for a content process crash on my local build?
14:20Gijson windows?
14:20GijsI've tried using msvc with entrian attach, but it attached to the gpu process as well and that crashes on startup...
14:22Gijsand attaching to a debug build after startup seems to show a "firefox.exe has triggered a breakpoint" modal dialog every half-second or so, so it's unusable. :|
14:28padenotI think if you do continue continuen continue for some time it will work
14:30pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/e73020a1e023 - Jan de Mooij - Bug 1381438 - Fix boxDouble to not clobber the source register on x86 CPUs without SSE 4.1. r=evilpie
14:39Gijspadenot: looks like that particular issue was caused by ohnoreflow
14:39Gijspadenot: which apparently breaks into the debugger for every sync reflow, or something?
14:43Gijsalso looks like content process assert stacks just go to /dev/null on windows?
15:02sheppyOK... Nightly is using 516 GB of virtual memory.
15:02sheppyMemory size is reported as 92.01 GB and climbing in activity monitor.
15:04pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/3da4ad542ed3 - Micah Tigley - Bug 1379715 - Grid Inspector Area Names are giant when they overflow. r=gl
15:11sheppyI even have 48 GB of real memory now instead of just 16 and still nightly is able to use up all available space and leave me with all my apps paused.
15:12bzsheppy: are you in a position to use about:memory to see what's going on?
15:12bzsheppy: if so, can you pastebin the result, perhaps?
15:13sheppybz: I have in fact grabbed logs from about:memory and a GC verbose log to share with erahm and mccr8 shortly.
15:13pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/07976bd4a36c - Benjamin Smedberg - Bug 1379254 - Add a @note to the documentation of MOZ_CRASH_UNSAFE_* to make clear data collection risk and requirements, r=erahm
15:13sheppyIf you'd like to see the about:memory info, yeah, I can pastebin that. One sec.
15:15sheppybz: https://pastebin.mozilla.org/9027353
15:16bzsheppy: hmm, 136 tiny windows totaling 1674MB
15:16bzso averaging >10MB each
15:16sheppyTiny windows?
15:16bzIf you expand that, what do those look like?
15:16bz++ (136 tiny)
15:16bzUnder explicit > window-objects
15:17sheppyHuh. They're hidden windows then, because I only have 8 open windows listed.
15:17sheppy(in the UX)
15:17bzwell, can you expand that twisty and re-pastebin?
15:17pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/7eceab6da875 - Josh Matthews - servo: Merge #17739 - Suppress CSS parser errors for vendor-prefixed properties (from jdm:no-vendor-prefixed-errors); r=emilio
15:17sheppyOh - yeah, sorry
15:17bz(note that "window" in this case means "toplevel tab")
15:18bzthe 90MB google docs is about as expected....
15:18sheppyAh, okay, so tabs.
15:18bz"window-object" means Window in the JS sense
15:18bzNot in the operating system sense. ;)
15:18sheppyOh - uh... as far as I can tell everything is open. Let me do a redo with verbose on.
15:20bzI mean, that "++" in "++ (136 tiny)" means it can be expanded, if you click on it
15:20sheppyHeh - too long for pastebin now
15:20pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/6d6a00964ace - Michael Kaply - Bug 1380846 - Add coccoc to Vietnamese desktop build. r=flod
15:20bzawesome
15:20bzPastebin the first 10 and last 10 from the "tiny" list?
15:21sheppyOK, hang on
15:21* bz wants some idea of what we're considering "tiny" at the top and how low we can go on the bottom
15:21bzThat said, this is all showin about 4.5GB of RAM, not 92GB....
15:22bz(but resident-peak is 26.5GB, which is quite something)
15:23sheppybz: that's pretty weird then, because activity monitor is saying 95.45 GB now.
15:23bzIn which column?
15:24bz"Real Mem" or something else?
15:24sheppy"Memory"
15:24bzAh, so including random mmapped stuff and whatnot
15:25bzSo "Real Mem" in Activity Monitor is actual allocated memory
15:25bz"Memory" includes files mapped as memory and so forth
15:25* bz checks some numbers
15:25bzso my nightly shows 3.4GB "Real Mem" and "15.99GB" under "Memory"
15:26sheppybz: https://www.dropbox.com/s/nb6l3598ig11f75/memory-report-verbose-2017-07-17b.txt?dl=0
15:26sheppyComplete memory report with everything opened up
15:27* bz looks
15:31bzok
15:31bzSo you have a bunch of about:blank at about 225KB each
15:31bzThose must be still-unloaded restored tabs...
15:31sheppybz: Yeah, I have a bunch of those
15:32bzThat's reasonable
15:32bzAlso some extension compartments
15:32pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/a14a6b94f2ed - Thomas Wisniewski - Bug 1348911 - Add lastAccessed to tabs.Tab; r=mixedpuppy,zombie
15:32bzbut not terribly big
15:33bzAnyway, nothing here looks weird for a largish browsing session
15:33sheppySomething's wrong though, because it's still using all available memory.
15:33pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/ecfa13e75024 - bechen - Bug 1372854 - Enable test_seek_nosrc.html on Android. r=alwu
15:33bzExcept for whatever mmapping is going on to give that 92GB value
15:33bzsheppy: how are you determining this last?
15:34bzsheppy: This report says we're using a bit over 4GB of actual allocated memory
15:34sheppybz: Because my Mac has a box up telling me that all application memory has been used and Nightly is what's at the top of the memory usage list.
15:34bzAh
15:34pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/7514d624a816 - Michael Brennan - Bug 1367125 - Record search cohort using the new annotation API. r=gfritzsche
15:34pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/39dc330b3560 - btian - Bug 1379925 - Part 1: Set PaymentRequestUpdateEvent.mRequest only when event target is a PaymentRequest. r=baku
15:34bzSo...
15:34pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/10677a3988fc - btian - Bug 1379925 - Part 2: Revise method |updateWith| based on spec change. r=baku
15:34sheppy95.7 GB of memory used and still climbing
15:34bzThat number isn't "memory" in the sense that "application memory" is meant
15:34bzWhat are the memory stats on kernel_task for you?
15:35pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/f4aacb756bac - gasolin - Bug 1374717 - show alternative message when firefox already the default browser;r=flod,mossop
15:35bzAt least I don't think that number is "memory" in the sense it's meant
15:35bzThough I suppose it could be....
15:35bzIt might be interesting to see what system calls nightly is making to allocate virtual memory
15:35sheppyReal Memory is listed at 7.84 GB
15:36sheppyCompressed Memory is 88.05 GB
15:36bzsheppy: for kernel_task?
15:36sheppykernel_task is 3.32 GB
15:36bzsheppy: e.g. if we're doing mmap (and if so with what args) or what
15:36pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/356bbf472e75 - Scott Wu - Bug 1374960 - [Form Autofill] Make small refinements for Manage Profiles and Edit Profile dialogs. r=lchang
15:37bzSo in my experience I get that "application memory" thing when kernel_task gets very bloated and starts being unable to allocate its own data structures
15:37bzIndependently of what's going on with actual memory usage. :(
15:38pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/6393471737b6 - steveck-chung - Bug 1377636 - Fix doorhanger intermittent error by using synthesizeMouseAtCenter and waiting for popup hidden. r=lchang
15:38sheppyAll I know is that the system thinks Nightly has allocated 95.75 GB of memory which it owns, 88 GB of it has been idle long enough to get compressed to disk.
15:38bzNot allocated
15:38bzmapped
15:38bznot the same thing
15:39sheppyWell, fine, true.
15:39sheppyRegardless, it's an awful lot. :)
15:39bzIt's a 64-bit system, which means it can support up to 48 bits of memory being mapped
15:39sheppyTrue.
15:39bzAnyway
15:39bzI agree it would be interesting to see what all those memory mappings are and wy
15:39bzer, why
15:39sheppyHuh. OK.
15:39pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/9c869c35cb7b - Ray Lin - Bug 1378072 - Update entity name of the warning note and its string as the meaning is slightly changed in the latest UX spec. r=lchang
15:40* bz wonders how hard it would be for about:memory to show that info.
15:40sheppyI did an "info" grab on Nightly...
15:40sheppyReal memory: 7.94 GB. Virtual Memory used: 529.7 GB. Shared memory size: 203.6 MB. Private memory size: 5.85 GB
15:41RyanVMwasn't there a recent OSX nightly regression along those lines?
15:41RyanVMrelated to a jemalloc patch that landed
15:41* RyanVM thought that got fixed last week though
15:41* sheppy is intrigued
15:41bzhmm
15:41sheppyLet's see, which nightly is this?
15:41bzSo what does that "92GB" number mean?
15:41sheppyIt'll be from late last week...
15:41* sheppy doesn't know. Let's see if I can find out
15:41pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/90497fbb54a6 - bechen - Bug 1378285 - Last frame reftest for bipbop.mp4 . r=alwu
15:41bzIf it's not the same as the (insane, btw!) 529.7GB number
15:41pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/820a85e09ac3 - bechen - Bug 1378285 - Modify the generateREF.html to resolve the output limit of console. r=alwu
15:42pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/299ad6c757d7 - Evan Tseng - Bug 1380205 - Check preferences version for some openPreferences calls need to direct to different pane in different version. r=mconley
15:43RyanVMsheppy: bug 1379139 is where the fix landed I was thinking of
15:43RyanVMyou can see the deps/dupes involved
15:44sheppyYeah, that could be me
15:44pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/c94795e84153 - Fischer.json - Bug 1381010 - After resizing the window, open the onboarding overlay but the tours would not appear. r=mossop
15:44sheppyYeah, I bet that's me. I'm using a July 7 build.
15:44sheppyI don't want to update and restart until I'm sure I've gathered all info that people want.
15:45pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/ec0dc3ef3bf7 - Ed Lee - Bug 1381335 - Increase waitForCondition time for initial onboarding-overlay check. f=Fischer r=dmose
15:48pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/7fc8b6cff9a4 - Sebastian Hengst - Backed out changeset e4dc716c401c (bug 1377595) on suspicion of causing reftest failures on Android "(waiting for MozAfterPaint)". r=backout
15:53Aryxgabor: hi, 1378207 seems to break a devtools test: https://treeherder.mozilla.org/#/jobs?repo=mozilla-inbound&revision=5d9a6384bf513139ca33f54825db4f035ebdbdab&filter-resultStatus=testfailed&filter-resultStatus=busted&filter-resultStatus=exception&filter-resultStatus=retry&filter-resultStatus=runnable&filter-resultStatus=success&filter-resultStatus=pending&filter-resultStatus=running&filter-searchStr=windows+M(dt4)
15:56gaborAryx: thanks, let's just back it out for now... seems like I missed it on my try run because of the intermittent bug probably
15:56Aryxok
15:59pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/e55488c64a0b - Sebastian Hengst - Backed out changeset 5d9a6384bf51 (bug 1378207) for frequently timing out in devtools' browser_addons_debug_webextension_popup.js and browser_addons_remove.js on Windows. r=backout
16:00sheppyHuh.
16:00sheppyHow can my July 7 build of Nightly show an uptime of 24 days?
16:00Aryxgabor: can't access the bug, so please keep track of it
16:01gaborAryx: alright, I'll take care of it
16:01sheppy24 days ago is the day I got this new iMac. Hah. It's showing my cumulative Nightly uptime since the day I first set up the machine. That's weird.
16:02sheppybz: So looks like at this time I have no tabs that are actually abusing memory then? Guess this is indeed a case of discarded memory not being released properly at some level.
16:14bzsheppy: Well, it's still not quite clear to me what this 90GB number is actually measuring...
16:14bzsheppy: So I have a hard time saying anything useful about it
16:15sheppybz: fair enough
16:15sheppyGoogling around a bit, it sounds like nobody really knows :)
16:18pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/5333e9eede4c - Fausto Nez Alberro - servo: Merge #16508 - Properly set origin of fetch requests (from brainlessdeveloper:fetch-set-origin); r=asajeffrey
16:22pulsebotCheck-in: https://hg.mozilla.org/releases/mozilla-esr52/rev/0b83a9dd14dc - ffxbld - No bug, Automated HSTS preload list update from host bld-linux64-spot-302 - a=hsts-update
16:22pulsebotCheck-in: https://hg.mozilla.org/releases/mozilla-esr52/rev/d6a351872b8f - ffxbld - No bug, Automated HPKP preload list update from host bld-linux64-spot-302 - a=hpkp-update
16:23pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/bccbac6fcf20 - Do Gottwald - Bug 1380268 - Let the new search textbox styling graduate from in-content/common.inc.css to textbox.css. r=nhnt11
16:25pulsebotCheck-in: https://hg.mozilla.org/mozilla-central/rev/6adf5aba89cd - ffxbld - No bug, Automated HSTS preload list update from host bld-linux64-spot-303 - a=hsts-update
16:25pulsebotCheck-in: https://hg.mozilla.org/mozilla-central/rev/e0b0865639ce - ffxbld - No bug, Automated HPKP preload list update from host bld-linux64-spot-303 - a=hpkp-update
16:26pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/565499a7bd59 - Benjamin Bouvier - Bug 1381467: Extend JS_AOT_USAGE telemetry probe to Firefox 60; r=luke, data-review=bsmedberg
16:28pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/f9c4ffc88192 - Brian Hackett - Bug 1375404 - Remove bogus assert, r=jandem.
16:41AryxGijs: hi, new failure for the tab count telemetry: https://treeherder.mozilla.org/logviewer.html#?job_id=114884732&repo=mozilla-inbound
16:44GijsAryx: is this intermittent or not?
16:45Aryxintermittent, found 1 hit so far. just a warning that if it turns out as frequent intermittent, that it would get backed out
16:46Gijsthere's more than 1
16:46Gijshttps://treeherder.mozilla.org/#/jobs?repo=mozilla-inbound&filter-searchStr=mochitest-browser-chrome-e10s&fromchange=2fee42d106b3bf5ea490154137ce5a65a4d60065&selectedJob=114823062
16:47GijsAryx: at least 3, across windows + linux debug
16:47GijsAryx: back it out and needinfo Lie on the bug, please? :)
16:48Aryxok, thanks for looking into this
16:51pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/44e71f8196d5 - Sebastian Hengst - Backed out changeset ede5862f0950 (bug 1361855)
16:51pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/c6588b951845 - Sebastian Hengst - Backed out changeset 7e9ac96978ad (bug 1361855)
16:51pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/18b3cca1c770 - Sebastian Hengst - Backed out changeset 06dbf8fc7da3 (bug 1361855)
16:51pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/3325a300f82f - Sebastian Hengst - Backed out changeset 2d6877c0b84e (bug 1361855)
16:51pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/4473acefd1b7 - Sebastian Hengst - Backed out changeset 6bfc9384ed64 (bug 1361855)
16:51pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/afc6dfe5af8d - Sebastian Hengst - Backed out changeset 4397e17fc5fc (bug 1361855)
16:51pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/d13067985d25 - Sebastian Hengst - Backed out changeset 14d02c0af528 (bug 1361855)
16:51pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/cde0022cc6e6 - Sebastian Hengst - Backed out changeset 8df92cb52eb4 (bug 1361855) for frequently failing browser_UsageTelemetry.js | TAB_COUNT telemetry. r=backout
16:57pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/c135a38a76a8 - Aaron Klotz - Bug 1380471: Move follow-up initialization for emulated windows into a callback invoked by the emulated window's WM_CREATE handler; r=yzen
17:04pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/a02a6cb9c94c - Nicolas B. Pierron - Bug 1364908 - Move moveValue into the generic MacroAssembler. r=jandem
17:04pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/60809bf55571 - Nicolas B. Pierron - Bug 1364908 - Add MacroAssembler::branchToComputedAddress. r=jandem
17:04pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/b6be7c2be50e - Nicolas B. Pierron - Bug 1364908 - IonMonkey: Add LoadElementFromSate to support argument[x] in inlined functions. r=jandem
17:05pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/a249596785aa - Nicolas B. Pierron - Bug 1364908 - Fix ion/bug1284485.js unreported OOM. r=bbouvier
17:14Aryxnbp: hi, bustage https://treeherder.mozilla.org/#/jobs?repo=mozilla-inbound&revision=a249596785aa58973113cf830a55a54d0a5257de&filter-resultStatus=testfailed&filter-resultStatus=busted&filter-resultStatus=exception&filter-resultStatus=retry&filter-resultStatus=runnable
17:19sheppySheesh... Activity Monitor continues to show my nightly memory use climbing. "Memory" is now 110.43 GB and real memory used is 22.79 GB)
17:29pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/227ff88b9848 - Brian Grinstead - Bug 1326937 - Provide HTML page when running new console frontend; r=bgrins
17:29pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/57701c371c8d - Jan Odvarko - Bug 1326937 - Patch react-dom to avoid double load r=bgrins
17:29pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/d99559a7adc7 - Jan Odvarko - Bug 1326937 - Fix CSS; r=bgrins
17:30pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/0b9f81280f19 - Chris Peterson - Bug 1367226 - Remove unused pref "plugins.navigator_hide_disabled_flash". r=bsmedberg
17:31pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/f4a2bab0b86a - Chris AtLee - Bug 1381526: switch OSX DE builds to cross-compile r=kmoir
17:39pulsebotCheck-in: https://hg.mozilla.org/integration/autoland/rev/33a29b909d86 - Mark Banner - Bug 1336806 - In browser_bookmark_popup.js' hover tests, place the mouse in the middle of the panel to try and avoid intermittents. r=mak
17:41nbpAryx: looking
17:41nbpAryx: can you back it out.
17:42Aryxnbp: ok
17:42nbpAryx: I don't think I would have time to fix it quickly.
17:44nbpAryx: This sounds like a missing include, I will fix and push it back tomorrow, thanks.
17:44pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/a51672e13a1b - Sebastian Hengst - Backed out changeset a249596785aa (bug 1364908)
17:44pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/9d84a411ea6d - Sebastian Hengst - Backed out changeset b6be7c2be50e (bug 1364908)
17:45pulsebotCheck-in: https://hg.mozilla.org/integration/mozilla-inbound/rev/5913aca49150 - Sebastian Hengst