5dd47ac3e2157e8d709eaa1e239050d74f567503: Bug 1367631 - Round the dest rect of the image layer to match the color layer behavior. r=mattwoodrow
Ethan Lin <ethlin@mozilla.com> - Wed, 31 May 2017 14:28:50 +0800 - rev 361522
Push 43803 by ryanvm@gmail.com at Wed, 31 May 2017 18:34:20 +0000
Bug 1367631 - Round the dest rect of the image layer to match the color layer behavior. r=mattwoodrow
14a2665a64f6e78b7c470a3eb0635e491660955a: Bug 1366984 - Fix image container check for background image layer. r=mattwoodrow
Ethan Lin <ethlin@mozilla.com> - Wed, 31 May 2017 14:01:28 +0800 - rev 361521
Push 43803 by ryanvm@gmail.com at Wed, 31 May 2017 18:34:20 +0000
Bug 1366984 - Fix image container check for background image layer. r=mattwoodrow
fd22c45d4adaf340a733457443a686bd136012d5: Bug 1366316 - Separate thread for IPCBlobInputStream actors - part 5 - WorkerHolder, r=smaug
Andrea Marchesini <amarchesini@mozilla.com> - Wed, 31 May 2017 07:41:11 +0200 - rev 361520
Push 43803 by ryanvm@gmail.com at Wed, 31 May 2017 18:34:20 +0000
Bug 1366316 - Separate thread for IPCBlobInputStream actors - part 5 - WorkerHolder, r=smaug WorkerHolder must be released after the migration.
3965a5c72ea35aede1a3db86556713fbc85b6d84: Bug 1366316 - Separate thread for IPCBlobInputStream actors - part 4 - monitoring child processes, r=smaug
Andrea Marchesini <amarchesini@mozilla.com> - Wed, 31 May 2017 07:41:11 +0200 - rev 361519
Push 43803 by ryanvm@gmail.com at Wed, 31 May 2017 18:34:20 +0000
Bug 1366316 - Separate thread for IPCBlobInputStream actors - part 4 - monitoring child processes, r=smaug
c1b48496a0f39423b97304c01d855f0f361969b3: Bug 1366316 - Separate thread for IPCBlobInputStream actors - part 3 - tests, r=smaug
Andrea Marchesini <amarchesini@mozilla.com> - Wed, 31 May 2017 07:41:10 +0200 - rev 361518
Push 43803 by ryanvm@gmail.com at Wed, 31 May 2017 18:34:20 +0000
Bug 1366316 - Separate thread for IPCBlobInputStream actors - part 3 - tests, r=smaug
432077ba9b7f109d0e123ae4852de30de9e8578b: Bug 1366316 - Separate thread for IPCBlobInputStream actors - part 2 - callback stored, r=smaug
Andrea Marchesini <amarchesini@mozilla.com> - Wed, 31 May 2017 07:41:10 +0200 - rev 361517
Push 43803 by ryanvm@gmail.com at Wed, 31 May 2017 18:34:20 +0000
Bug 1366316 - Separate thread for IPCBlobInputStream actors - part 2 - callback stored, r=smaug
59c54a18bf9532ccb206494646b59af56012df51: Bug 1366316 - Separate thread for IPCBlobInputStream actors - part 1 - actor migration, r=smaug
Andrea Marchesini <amarchesini@mozilla.com> - Wed, 31 May 2017 07:41:10 +0200 - rev 361516
Push 43803 by ryanvm@gmail.com at Wed, 31 May 2017 18:34:20 +0000
Bug 1366316 - Separate thread for IPCBlobInputStream actors - part 1 - actor migration, r=smaug
4c2698925268c7cc51ac8aa83a38115c8ad06164: Bug 1366316 - Separate thread for IPCBlobInputStream actors - part 0 - comments, r=smaug
Andrea Marchesini <amarchesini@mozilla.com> - Wed, 31 May 2017 07:41:10 +0200 - rev 361515
Push 43803 by ryanvm@gmail.com at Wed, 31 May 2017 18:34:20 +0000
Bug 1366316 - Separate thread for IPCBlobInputStream actors - part 0 - comments, r=smaug
e011d885b00b28124c89ce74fca5924e0aafbf57: Bug 1367629 - IonMonkey: MIPS: Fix sync instruction encoding for Loongson. r=h4writer
Heiher <r@hev.cc> - Wed, 31 May 2017 13:21:31 +0800 - rev 361514
Push 43803 by ryanvm@gmail.com at Wed, 31 May 2017 18:34:20 +0000
Bug 1367629 - IonMonkey: MIPS: Fix sync instruction encoding for Loongson. r=h4writer --- js/src/jit/mips-shared/Assembler-mips-shared.cpp | 2 ++ 1 file changed, 2 insertions(+)
8429360d3050f174202a1a8133489b6f7291348f: Bug 1347274 (part 3) - Make some hot profiler functions lockless. r=mstange.
Nicholas Nethercote <nnethercote@mozilla.com> - Tue, 30 May 2017 13:36:40 +1000 - rev 361513
Push 43803 by ryanvm@gmail.com at Wed, 31 May 2017 18:34:20 +0000
Bug 1347274 (part 3) - Make some hot profiler functions lockless. r=mstange. The functions are profiler_is_active(), profiler_feature_active(), profiler_tracing(), profiler_log(), profiler_add_marker(). This is achieved by adding RacyFeatures, which duplicates the features stored in ActivePS in a way that can be accessed locklessly. There's a chance of imprecision, but that doesn't matter for the way these functions are used; the worst that can happen is we might very occasionally miss adding a marker, or add one that we shouldn't.
79af4e3b4d718a368449a0a442c67ca3c1a7d3fd: Bug 1347274 (part 2) - Make CorePS::ProcessStartTime() lockless. r=mstange.
Nicholas Nethercote <nnethercote@mozilla.com> - Tue, 30 May 2017 11:02:59 +1000 - rev 361512
Push 43803 by ryanvm@gmail.com at Wed, 31 May 2017 18:34:20 +0000
Bug 1347274 (part 2) - Make CorePS::ProcessStartTime() lockless. r=mstange. It's immutable. This also means that profiler_time() is now lockless, though that's a very rarely used function so it's not a performance win.
ad10343345034bf5f0f1c31e529af3c042427573: Bug 1347274 (part 1) - Rename some profiler parameters. r=mstange.
Nicholas Nethercote <nnethercote@mozilla.com> - Tue, 30 May 2017 10:31:17 +1000 - rev 361511
Push 43803 by ryanvm@gmail.com at Wed, 31 May 2017 18:34:20 +0000
Bug 1347274 (part 1) - Rename some profiler parameters. r=mstange. The profiler is inconsistent about how it names some things, mostly parameters. This patch fixes them as follows. - A marker's identifier is now consistently |aMarkerName| (matching ProfilerMarker::mMarkerName), not |aMarker| or |aMarkerStr| or |aInfo|. - Subclasses of ProfilerMarkerPayload are now consistently |payload|, not |marker|. - The first string in a PseudoStack entry is now consistently |aLabel| (matching ProfileEntry::label_), not |aInfo|.
cda03201b522c7fd3e7b10827b04bf5483e32816: Bug 1368910 - Update cubeb from upstream to 3428c2b. r=kinetik
Alex Chronopoulos <achronop@gmail.com> - Wed, 31 May 2017 07:59:22 +0300 - rev 361510
Push 43803 by ryanvm@gmail.com at Wed, 31 May 2017 18:34:20 +0000
Bug 1368910 - Update cubeb from upstream to 3428c2b. r=kinetik MozReview-Commit-ID: FIUMIri84Wr
2645d067b1e3f4064f18a7d1241ec96b33402d07: Backed out changeset 970698e1bb62 (bug 1361755) for Android bustage at /mobile/android/base/resources/layout/preference_set_homepage.xml:20 (unbound prefix). r=backout
Sebastian Hengst <archaeopteryx@coole-files.de> - Wed, 31 May 2017 16:02:50 +0200 - rev 361509
Push 43802 by archaeopteryx@coole-files.de at Wed, 31 May 2017 14:03:06 +0000
Backed out changeset 970698e1bb62 (bug 1361755) for Android bustage at /mobile/android/base/resources/layout/preference_set_homepage.xml:20 (unbound prefix). r=backout
82c13bf361192b7c9a93d20a0af5c4aaf48f4cff: Bug 1368940 - Avoid potential race condition in adjustTabstrip. r=mconley
Dão Gottwald <dao@mozilla.com> - Wed, 31 May 2017 09:31:44 +0200 - rev 361508
Push 43801 by dgottwald@mozilla.com at Wed, 31 May 2017 13:47:52 +0000
Bug 1368940 - Avoid potential race condition in adjustTabstrip. r=mconley The early this.getAttribute("overflow") == "true" check + requestAnimationFrame + _closeButtonsUpdatePending can make us skip an adjustTabstrip call from the overflow event that would remove the overflow attribute. MozReview-Commit-ID: 4cCdotvWmIa
9e738d40b62dd65294529c666bd7505ba9ece8d4: Bug 1319226 - Process JARManifests in the tup backend; r=chmanchester
Mike Shal <mshal@mozilla.com> - Fri, 26 May 2017 10:43:31 -0400 - rev 361507
Push 43800 by mshal@mozilla.com at Wed, 31 May 2017 13:46:38 +0000
Bug 1319226 - Process JARManifests in the tup backend; r=chmanchester The JARManifests are relatively straightforward since they can be converted into FinalTargetFiles by _consume_jar_manifest(). Since these can contain RenamedSourcePaths, we no longer assert an error for those. The main difference is that some JARManifests can result in preprocessing .css files, which use a different marker than the default. The two odd cases are bookmarks.html.in and buildconfig.html. The first relies on an l10n define in a Makefile, while the latter needs CXXFLAGS. The faster make backend works around these by defining BOOKMARKS_INCLUDE_DIR to the en-US version, and skips the CXXFLAGS case by defining BUILD_FASTER=1, so that is what the tup backend does for now as well. MozReview-Commit-ID: LlmNjQ2g40j
9bce19249d9b7fd841f8c85c7974b0e08d0f4440: Bug 1319226 - allow symlinks to point to files with a different basename; r=chmanchester
Mike Shal <mshal@mozilla.com> - Wed, 24 May 2017 15:41:40 -0400 - rev 361506
Push 43800 by mshal@mozilla.com at Wed, 31 May 2017 13:46:38 +0000
Bug 1319226 - allow symlinks to point to files with a different basename; r=chmanchester Some jar.mn files create symlinks with a different name than the file they point to. The tup backend should support this by using the object's target_basename as the destination filename. MozReview-Commit-ID: IB4moXA7KWH
970698e1bb620df117a6da3964b9e4d9a7e7f7c2: Bug 1361755 - Allow distributions to specify a homepage option in prefs. r=sebastian
Michael Kaply <mozilla@kaply.com> - Wed, 03 May 2017 15:47:09 -0500 - rev 361505
Push 43799 by mozilla@kaply.com at Wed, 31 May 2017 13:30:28 +0000
Bug 1361755 - Allow distributions to specify a homepage option in prefs. r=sebastian MozReview-Commit-ID: KjdZvHNG6VR
703fc5a4a92c63151f7d1d070ec1c765ee610830: Bug 1368502 - Remove unused import in about:telemetry r=chutten
flyingrub <flyinggrub@gmail.com> - Mon, 29 May 2017 17:42:25 +0200 - rev 361504
Push 43798 by ryanvm@gmail.com at Wed, 31 May 2017 13:29:38 +0000
Bug 1368502 - Remove unused import in about:telemetry r=chutten MozReview-Commit-ID: 3NK85cXtuLP
8f2b4efc5f0d3191a7e80d9324933621e111b44a: Bug 1361699 - Add buffer when writing hashstore to file r=gcp,mcmanus
Thomas Nguyen <tnguyen@mozilla.com> - Fri, 19 May 2017 13:17:07 +0800 - rev 361503
Push 43797 by ryanvm@gmail.com at Wed, 31 May 2017 13:28:15 +0000
Bug 1361699 - Add buffer when writing hashstore to file r=gcp,mcmanus We write a lot of 4-bytes prefixes to file which call many system calls. We should use a buffer and only write to file if the buffer is full or finish writing. nsIBufferedOutputStream is a good candidate to do that MozReview-Commit-ID: CzGOd7iXVTv
(0) -300000 -100000 -30000 -10000 -3000 -1000 -300 -100 -50 -20 +20 +50 +100 +300 +1000 +3000 +10000 +30000 +100000 tip