mozilla :: #developers

20 May 2017
01:10pulsebotCheck-in: - Simon Sapin - servo: Merge #16954 - Avoid returning / passing around a huge ParsedDeclaration type (from servo:arrayvec); r=emilio
01:10pulsebotCheck-in: - Servo VCS Sync - No bug - Revendor rust dependencies
01:16pulsebotCheck-in: - Steve Fink - Bug 1291954 - Update tsan failures to current set and move SM(tsan) to tier 1, r=jonco
01:44jesupdmajor: thanks!!!
01:44jesupit is really, really weird
02:18jesupdmajor: works now with private: virtual void DummyVS2015Fix() {}
02:37pulsebotCheck-in: - Gregory Terzian - servo: Merge #16861 - Use microtasks to await a stable state (from gterzian:use_microtask_to_await_stable_state); r=jdm
03:15pulsebotCheck-in: - Ricky Chien - Bug 1351867 - Show search tooltip on top of button r=mconley
03:58philorsfink: having trouble deciding the right thing to do with that tsan patch, since the move to tier-1 was massively premature and it doesn't remotely qualify for tier 1, but then, you didn't remove the run-on-projects: [] so it doesn't actually run
04:00philora 20% intermittent-failure rate plus only reporting failures through a near-invisible TinderboxPrint isn't even on the same planet as tier-1
04:24pulsebotCheck-in: - Phil Ringnalda - Bug 1291954 followup, revert the part of changeset 2db29775e18f which prematurely declared SM(tsan) tier-1
04:46philorhuh, what did add that requires something that's disabled on release_or_beta?
04:47philorsurely that's not the first need for dom.animations-api.core.enabled in the test, is it?
04:50philorguess so, since it doesn't set it
04:59pulsebotCheck-in: - Ehsan Akhgari - Bug 1362791 - Enable testing permissions using URIs without having to mint principals; r=mystor
05:41philorgetting really really bored with filing "permaorange when gecko 55 merges to beta on 2017-06-12" bugs
05:49philormrbkap: that's inconvenient, apparently the process in question in is a nightly-only process
06:12pulsebotCheck-in: - Manish Goregaokar - servo: Merge #16914 - Update app_units to 0.4.1 (from Manishearth:update_au); r=jryans
06:20Manishearth needs to be applied to make stylo green after the above merge
06:27pulsebotCheck-in: - Hiroyuki Ikezoe - Bug 1341102 - Update test expecrations. r=me
06:27hiroManishearth: done ^
06:29Manishearth:) thanks
06:38philorhere, let me just close autoland while we have the usual discussion about who it is that remembers how to manually revendor
06:47* heycam tries
06:56pulsebotCheck-in: - Cameron McCormack - Bug 1366454 - Revendor app_units manually. r=me (CLOSED TREE)
06:57heycamphilor: ^ hopefully that works
07:00philorstarting to think taskcluster's broken, though, between those decision tasks and the builds and tests all around that have been running for hundreds of minutes
07:18philorand with that, my work here is done
07:19xidornwow all trees are closed
07:20philoreh, I left Try open
07:20philorand comm-*, and nss which I probably shouldn't have but it's not my department
07:23philorand if I was really good, I would have left esr45 open, since it doesn't have a taskcluster dependency, but it's not like anyone pushes there anyway
10:19mrbkapphilor: that's indeed unfortunate.
10:20mrbkapphilor: I have an idea of how to fix it.
11:08emiliois autoland open?
11:08mrgigglesemilio: autoland: closed (Bug 1366456 - taskcluster jobs starting and then not doing anything at all)
11:08emiliosounds fun
15:11borisMy servo patches are landed, but autoland is closed. Should I push the button of the interdependent Gecko patches now? OR just wait?
15:44Aryxboris: hi, autoland is now open
15:45pulsebotCheck-in: - Shih-Chiang Chien - Bug 1363421 - Part 1, lazy load HTTP protocol handler to improve start-up performance. r=mcmanus
15:45pulsebotCheck-in: - Shih-Chiang Chien - Bug 1363421 - Part 2, delay the initialization of UserAgentOverrides.jsm until first nsHttpChannel is created. r=mcmanus
15:46pulsebotCheck-in: - Boris Chiou - Bug 1366441 - Enable dom.animations-api.core.enabled for test_animation_observers_sync.html. r=hiro
15:48pulsebotCheck-in: - Emilio Cobos lvarez - Bug 1364871: Factor out some code from UpdateStyleOfChildAnonBox into UpdateStyleOfOwnedChildFrame. r=heycam
15:48pulsebotCheck-in: - Emilio Cobos lvarez - Bug 1364871: Add a function to update frame pseudo-element styles during the post-traversal, and restyle bullet frames. r=heycam
15:48pulsebotCheck-in: - Emilio Cobos lvarez - Bug 1364871: Restyle ::backdrop too. r=heycam
15:48pulsebotCheck-in: - Emilio Cobos lvarez - Bug 1364871: Revert workaround added in bug 1328319. r=heycam
15:50pulsebotCheck-in: - Mantaroh Yoshinaga - servo: Merge #16942 - Make font-stretch animatable (from mantaroh:font-stretch); r=hiro
15:50pulsebotCheck-in: - Boris Chiou - servo: Merge #16963 - stylo: Bug 1334036 - Enable animations running on compositor (from BorisChiou:stylo/animation/omta); r=emilio
15:52pulsebotCheck-in: - 12 changesets (bug 1334036)
15:56pulsebotCheck-in: - Emilio Cobos lvarez - servo: Merge #16961 - style: Move some bloom filter code outside of (from emilio:bloom-simplify); r=heycam
16:46pulsebotCheck-in: - 345 changesets (bugs 1351010, 1359011, 1356244, 1365901, 1365643 and 13 other bugs)
16:49philor345? is that all you could find?
16:49pulsebotCheck-in: - 23 changesets - merge mozilla-inbound to mozilla-central. r=merge a=merge
16:54pulsebotCheck-in: - 24 changesets - merge mozilla-central to autoland. r=merge a=merge
16:55pulsebotCheck-in: - 346 changesets - merge mozilla-inbound to mozilla-central. r=merge a=merge
17:12Aryxemilio: hi, can you update the test expectations, please? started failing with 1364871
17:13emilioAryx: whoops, sorry, yeah, missed it since autoland was closed, will do ASAP, thanks!
17:18pulsebotCheck-in: - L. David Baron - Bug 1365449 - Reflow absolutely positioned children when they need to be repaginated. r=mats
17:22philoroh wpt, I hate you so much
17:22philornewly added test will fail on beta because PerformanceObserver isn't defined, fine, I'll find what pref the surviving ones are setting
17:22philoroh, they all fail on the trunk too, that's how they survive
17:22philorexcept this one...
17:39pulsebotCheck-in: - Emilio Cobos lvarez - Bug 1364871: Update expectations. r=me
17:39emilioAryx: ^, sorry for the lag
17:39Aryxno worries, thanks
18:39pulsebotCheck-in: - Jonathan Guillotte-Blouin - Bug 1365068 - Cleanup getListSuggestions. r=MattN
18:39pulsebotCheck-in: - Jonathan Guillotte-Blouin - Bug 1365068 - Added missing brackets & refactored some conditional statements. r=MattN
18:39pulsebotCheck-in: - Jonathan Guillotte-Blouin - Bug 1365068 - normalize indentation to 2 spaces. r=MattN
18:39pulsebotCheck-in: - Jonathan Guillotte-Blouin - Bug 1365068 - import with destructuring, switch to shorthand function declarations and use array.includes. r=MattN
18:40pulsebotCheck-in: - Tobias Schneider - Bug 1358666 - (intersection-observer) Import web-platform tests. r=jet
18:41pulsebotCheck-in: - Will Wang - Bug 1356605 - Part 1: Add an enumerator for getting session cookies only. r=valentin
18:41pulsebotCheck-in: - Will Wang - Bug 1356605 - Part 2: Use the new `sessionEnumerator` to iterate only the cookies we need. r=mikedeboer
18:41pulsebotCheck-in: - Geoff Lankow - Bug 1361899 - Allow extensions to specify provider when calling topSites.get. r=mixedpuppy
18:42pulsebotCheck-in: - Andreas Farre - Bug 1355311 - Set default values for throttling background timeouts. r=ehsan
18:45pulsebotCheck-in: - Anthony Ramine - servo: Merge #16960 - Derive HasViewportPercentage (from servo:derive-all-the-things); r=emilio
18:45pulsebotCheck-in: - Servo VCS Sync - No bug - Revendor rust dependencies
18:46* philor wonders whether we actually fixed the problems that caused everything to explode from the last wineglass commit message
19:49pulsebotCheck-in: - Manish Goregaokar - Bug 1362599 - Guard against threadsafety issues in lang group stuff. r=heycam
19:49pulsebotCheck-in: - Manish Goregaokar - Bug 1362599 - Remember which languages were used and force-cache when reset; r=heycam
20:00pulsebotCheck-in: - Bobby Holley - Bug 1329876 - Stop accepting aTreeMatchContext in the ServoStyleSet methods. r=bz
20:00pulsebotCheck-in: - Bobby Holley - Bug 1329876 - Don't maintain a TreeMatchContext for Servo. r=bz
20:03noxIf the glass of wine breaks everything again, I'll send a box of bottles to whoever have to clean up my mess again.
20:05pulsebotCheck-in: - Emilio Cobos lvarez - servo: Merge #16967 - Bug 1366144: Correctly diff ::before and ::after pseudo-element styles if there's no generated content. r=heycam (from emilio:after); r=heycam,emilio
20:21pulsebotCheck-in: - Michael Kohler - Bug 1354884 - convert uses of 'defer' to 'new Promise' in client/webaudioeditor r=nchevobbe
20:23pulsebotCheck-in: - Emilio Cobos lvarez - Bug 1366144: Test. r=heycam
20:24philorhave I mentioned how much I hate wpt in the last three hours?
20:26emiliophilor: how's that?
20:28philorsomebody spent some time writing, and we imported it without actually looking at it, and now we spend time and money to run it on every push on every platform expecting it to time out, because it's totally unowned in our world so nobody will ever look at why it
20:28philortimes out
20:29pulsebotCheck-in: - Cameron McCormack - Bug 1289868 - Part 1: stylo: Add an outparam to Gecko_CalcStyleDifference that returns whether any style data changed. r=emilio
20:29pulsebotCheck-in: - Cameron McCormack - Bug 1289868 - Part 2: stylo: Compare all structs in CalcStyleDifference so that Servo can accurately determine whether to stop traversal. r=emilio
20:33philorif I set browser.storageManager.enabled in it, that should get it back to timing out on beta like it does on trunk, tempting to just stop there
20:36emiliophilor: hmm... chromium whitelists directories people want to sport fwiw
20:37pulsebotCheck-in: - Cameron McCormack - Bug 1364845 - stylo: Adjust test expectations. r=emilio
20:38emiliophilor: err, to import, that is
20:39emiliophilor: and they ensure their tests are owned by someone, perhaps we should do something like that
20:47noxemilio: Can't they be skipped like in Servo?
20:47emilionox: perhaps
20:47emilionox: not sure how much of a win is it though
20:48noxemilio: Well, no time and money spent at least.
20:48* nox would love it if every test was in wpt.
20:49emilionox: oh, sure, that's the other way around though
20:49emilionox: philor was annoyed at some imported test
20:49noxemilio: I understood that.
20:49noxemilio: I was zeroing philor's hate with my own love. :P
20:50philorI think our theory is that by running and expecting failure, then when someone causes them to not fail they will have to mark them as passing, where skipping stays skipped forever
20:52philorwholesale moving everything to wpt would in fact solve my problem, since it would require us to own the wpt heirarchy despite it not matching our modules
20:54philorwe can't usefully say "every annotation of failure in a wpt import has to be matched with a bug" since the bulk of the bugs would be uselessly filed in core::general since enormous swathes of wpt are covered by between zero and five or six bug components
20:56philorthis one I'm going to fix without bothering to file a bug, in /server-timing/ - wtf is /server-timing/?
21:04philorthe answer appears to be "it's an unowned misplaced bit of performance-timeline, which is owned but misplaced dom/performance"
21:09philorso a great big ball of w3c and thus no doubt also whatwg politics means I waste half an hour finding out why this dom/performance test is in a place where it doesn't already have a __dir__.ini setting dom.enable_performance_observer
21:14noxThat sounds complicated indeed.
21:16philorour politics is clear and simple and understandable, because I've spent 13 years studying it; their politics is obscure and convoluted and pointless, because I have no desire to study it
21:22pulsebotCheck-in: - Sylvestre Ledru - Bug 1366497 - disable strict-aliasing for gcc in js/src/gdb & js/src/jsapi-tests r=sfink
21:26sfinkphilor: r=you for deleting the run-on-projects and resetting to tier 3 for now?
21:27philorsfink: yes on both
21:30pulsebotCheck-in: - Steve Fink - Bug 1366522 - Drop SM(tsan) back to tier 3, r=philor
21:33sfinkhm, I guess running this as tier 3 for a week or two should work out well for me anyway, since mrgiggles is paying attention to it and gathering all new failures
21:35sfinkoh, you'd already dropped it to tier 3. I didn't see that.
21:35sfinkI guess the rebase just happily merged them
21:39philorthe commit message will ensure confusion for our enemies, though, so it's a win
21:45philoremilio: heap write hazard,
21:46philoroh, you pushed something by that notoriously hazardous heycam, you'll regret that
21:47emiliophilor: errr, that's heycam's yeah, but I think it's a bool outparam, so should be easy to whitelist
21:47* emilio needs to find out how though
21:47emiliophilor: thanks for the ping
21:50sfinkemilio: add it to
21:51emiliosfink: thanks! will do :)
22:03emiliophilor: should I push the hazard fix even though the tree is closed?
22:03emiliophilor: or should I just upload it to a bug and enqueue it?
22:04philoremilio: yes, push it, it's only buildbot that's broken and we're only interested in taskcluster results from the fix
22:05emiliophilor: thanks, done
22:05pulsebotCheck-in: - Emilio Cobos lvarez - Bug 1289868: Whitelist outparam in heap write hazard analysis on a CLOSED TREE. r=me
22:55pulsebotCheck-in: - Imanol Fernandez - servo: Merge #16933 - Implement MutationObserver childList mutations (from MortimerGoro:mutation_children); r=jdm
21 May 2017
No messages
Last message: 1 day and 17 hours ago