c03edb23edaa0f8e8fc5585838dc58e7a96e1cb0: Bug 1552964 - Remove the third argument passed to run_robocop_test(). r=ahal
Ting-Yu Lin <tlin@mozilla.com> - Tue, 21 May 2019 13:09:21 +0000 - rev 474741
Push 113171 by aiakab@mozilla.com at Tue, 21 May 2019 21:56:33 +0000
Bug 1552964 - Remove the third argument passed to run_robocop_test(). r=ahal Differential Revision: https://phabricator.services.mozilla.com/D31873
272a61f8978bd2ec910d6c4abc19d08d7d8cf5fd: bug 1553316: remote: update reference to sync communication; r=me a=doc
Andreas Tolfsen <ato@sny.no> - Tue, 21 May 2019 22:47:21 +0100 - rev 474740
Push 113170 by ato@sny.no at Tue, 21 May 2019 21:49:07 +0000
bug 1553316: remote: update reference to sync communication; r=me a=doc The remote protocol work has moved from IRC to Slack. DONTBUILD
17018058233cfc319552ca7d97f1dfae35c3c50b: Bug 1530190. Adjust assert expectations on toolkit/content/tests/chrome/test_bug437844.xul. r=aosmond
Timothy Nikkel <tnikkel@gmail.com> - Wed, 15 May 2019 00:54:31 -0500 - rev 474739
Push 113169 by tnikkel@gmail.com at Tue, 21 May 2019 21:38:13 +0000
Bug 1530190. Adjust assert expectations on toolkit/content/tests/chrome/test_bug437844.xul. r=aosmond We frequently get more asserts in this test after we increased the async-ness of AsyncOpen. Before the patches of this bug, in this test we would see 15 calls to nsIconChannel::AsyncOpen, and then a flurry of reflow asserts ("ASSERTION: bad inline size"). After the patches for this bug, sometimes we get all 15 AsyncOpen calls to complete (FinishAsyncOpen) before the asserts start. In that case we get an expected number of asserts. But frequently we get a case where some FinishAsyncOpen calls don't happen until after all of the asserts. In this case we get more asserts than we are expecting. So it seems the lack of a complete icon image is causing extra reflow churn (it's the same asserts, just more). Differential Revision: https://phabricator.services.mozilla.com/D31190
1d13cc5dbb8a842f602eaffecb109eed523fdb23: Bug 1530190. Cleanup GetStockIcon from first patch. r=aosmond
Timothy Nikkel <tnikkel@gmail.com> - Mon, 13 May 2019 21:34:01 -0500 - rev 474738
Push 113169 by tnikkel@gmail.com at Tue, 21 May 2019 21:38:13 +0000
Bug 1530190. Cleanup GetStockIcon from first patch. r=aosmond Looks like this is left over from a partial attempt at making GetStockIcon async like GetHIconFromFile. But I don't think it needs to be async, the only system call SHGetStockIconInfo doesn't have anything about background thread like SHGetFileInfo. Differential Revision: https://phabricator.services.mozilla.com/D31013
642316b68e6a267ce07e99d5e7506119820b437b: Bug 1530190. Allow nsIconChannel::Open to work on Windows. r=aosmond
Timothy Nikkel <tnikkel@gmail.com> - Thu, 02 May 2019 03:09:37 -0500 - rev 474737
Push 113169 by tnikkel@gmail.com at Tue, 21 May 2019 21:38:13 +0000
Bug 1530190. Allow nsIconChannel::Open to work on Windows. r=aosmond As stated in other patches in this bug, the system call has to happen off main thread, and it has to always be on the same thread. So we dispatch the task to that thread and synchronously wait on the main thread for it to finish. Differential Revision: https://phabricator.services.mozilla.com/D29626
2d9c02e85710b2fa095f7542b0b58996c7e8a77a: Bug 1530190. Use the image io thread to call SHGetFileInfo. r=aosmond
Timothy Nikkel <tnikkel@gmail.com> - Mon, 22 Apr 2019 17:30:41 -0500 - rev 474736
Push 113169 by tnikkel@gmail.com at Tue, 21 May 2019 21:38:13 +0000
Bug 1530190. Use the image io thread to call SHGetFileInfo. r=aosmond Windows doesn't like more than one thread (at the same time?) calling SHGetFileInfoW. So we use the same thread to avoid this. We also need to call CoInitialize on the thread we do this. Differential Revision: https://phabricator.services.mozilla.com/D28423
8010faead0fd05a943e260dc6b5a7d9b024eef88: Bug 1530190. De-unify image/DecodePool.cpp on Windows. r=aosmond
Timothy Nikkel <tnikkel@gmail.com> - Mon, 22 Apr 2019 17:30:40 -0500 - rev 474735
Push 113169 by tnikkel@gmail.com at Tue, 21 May 2019 21:38:13 +0000
Bug 1530190. De-unify image/DecodePool.cpp on Windows. r=aosmond We need to call CoInitialize, so we need to include the objbase.h header. That defines some generic boolean stuff, which conflicts with something that libjpeg defines. I tried working around it, but I gave up, this solution seems fine. Differential Revision: https://phabricator.services.mozilla.com/D28422
875d43d35a71e530b84388ad55d515015c6d3a68: Bug 1530190. Fix unified build bustage in image/Downscaler.cpp. r=aosmond
Timothy Nikkel <tnikkel@gmail.com> - Mon, 22 Apr 2019 17:30:40 -0500 - rev 474734
Push 113169 by tnikkel@gmail.com at Tue, 21 May 2019 21:38:13 +0000
Bug 1530190. Fix unified build bustage in image/Downscaler.cpp. r=aosmond The unifed tweak I make in the next patch exposes this. Differential Revision: https://phabricator.services.mozilla.com/D28421
bb340bd557d44a059953eb8b5f5e48ab3761a545: Bug 1530190. Change description of DecodePool::mMutex to be more accurate. r=aosmond
Timothy Nikkel <tnikkel@gmail.com> - Mon, 22 Apr 2019 17:30:40 -0500 - rev 474733
Push 113169 by tnikkel@gmail.com at Tue, 21 May 2019 21:38:13 +0000
Bug 1530190. Change description of DecodePool::mMutex to be more accurate. r=aosmond This mutex only protects the iothread, not the rest of the decodepool. There is a different monitor for that in DecodePoolImpl. Differential Revision: https://phabricator.services.mozilla.com/D28420
71bf2b9ca6ea2cbd97f4a07242e2ad7f20a1a2e5: Bug 1530190. Call SHGetFileInfo off the main thread in nsIconChannel on Windows. r=tnikkel
Andrew Osmond <aosmond@mozilla.com> - Mon, 22 Apr 2019 17:30:40 -0500 - rev 474732
Push 113169 by tnikkel@gmail.com at Tue, 21 May 2019 21:38:13 +0000
Bug 1530190. Call SHGetFileInfo off the main thread in nsIconChannel on Windows. r=tnikkel There are two reasons for this. One is performance, it can be slow. The second is that is can spin the event loop which can re-enter into things like layout. This patch uses the image decoder thread pool for that. But this is unsuitable (reason in next patch), the next patch changes it to use the img io thread used for network to pass data to imglib off main thread. Differential Revision: https://phabricator.services.mozilla.com/D28419
3c0f78074b727fbae112b6eda111d4c4d30cc3ec: Merge mozilla-inbound to mozilla-central a=merge
Razvan Maries <rmaries@mozilla.com> - Tue, 21 May 2019 19:29:01 +0300 - rev 474731
Push 113168 by rmaries@mozilla.com at Tue, 21 May 2019 16:39:23 +0000
Merge mozilla-inbound to mozilla-central a=merge
78571bb1f20e643cdab9791f61b0aa7b6a8bdf90: Bug 1534932 - Ensure filtering works with warningGroups. r=Honza.
Nicolas Chevobbe <nchevobbe@mozilla.com> - Tue, 21 May 2019 12:43:04 +0000 - rev 474730
Push 113168 by rmaries@mozilla.com at Tue, 21 May 2019 16:39:23 +0000
Bug 1534932 - Ensure filtering works with warningGroups. r=Honza. We want the warningGroup to be displayed if at least one of its children will be visible. We also want all the children if the warningGroup should be visible. This requires a few changes in the message reducers, mainly in the getVisibility function. But we also modify maybeSortVisibleMessages to place the messages in warningGroups at the expected positions. A complete mochitest is added to ensure the output always has the expected messages in the expected order. Differential Revision: https://phabricator.services.mozilla.com/D31220
be01ec66f386d44d59fc56ab904a84c299521177: Bug 1419358 - Never filter-out navigation markers message. r=Honza.
Nicolas Chevobbe <nchevobbe@mozilla.com> - Tue, 21 May 2019 12:42:34 +0000 - rev 474729
Push 113168 by rmaries@mozilla.com at Tue, 21 May 2019 16:39:23 +0000
Bug 1419358 - Never filter-out navigation markers message. r=Honza. Differential Revision: https://phabricator.services.mozilla.com/D31219
4039c6761d62239b4e55691f381cf2953969dcca: Bug 1533277 - Use SearchBox component in console filter bar. r=ntim.
Aaditya Arora <b17071@students.iitmandi.ac.in> - Tue, 21 May 2019 12:20:50 +0000 - rev 474728
Push 113168 by rmaries@mozilla.com at Tue, 21 May 2019 16:39:23 +0000
Bug 1533277 - Use SearchBox component in console filter bar. r=ntim. The code is adapted to work with the SearchBox, which means we can get rid of an event listener in webconsole-ui.js. Also, we take this as an opportunity to remove the element refs as they were only used in tests, where we can more easily use query selectors to get those elements. Some changes were needed in SearchBox's `onChange` function since `setState` is asynchronous. Differential Revision: https://phabricator.services.mozilla.com/D31762
086a7a5fa9eb35bd1e341a91aa8cb934241b5b3f: Backed out changeset 754426d70d28 (bug 1551729) for causing wpt failures. CLOSED TREE
Mihai Alexandru Michis <malexandru@mozilla.com> - Tue, 21 May 2019 15:17:37 +0300 - rev 474727
Push 113168 by rmaries@mozilla.com at Tue, 21 May 2019 16:39:23 +0000
Backed out changeset 754426d70d28 (bug 1551729) for causing wpt failures. CLOSED TREE
03d3b6d0eeca1bcf833bef4d519848220d5c78bd: Bug 1383650 - followup: Regenerate property database again. r=bustage
Emilio Cobos Álvarez <emilio@crisal.io> - Tue, 21 May 2019 13:59:18 +0200 - rev 474726
Push 113168 by rmaries@mozilla.com at Tue, 21 May 2019 16:39:23 +0000
Bug 1383650 - followup: Regenerate property database again. r=bustage CLOSED TREE
28504d015b50e07394582a0b7a166e9a98d5e6a5: Bug 1520191 - Polish style for error and warning messages r=jdescottes,ladybenko,flod
Ola Gasidlo <ogasidlo@mozilla.com> - Tue, 21 May 2019 11:14:47 +0000 - rev 474725
Push 113168 by rmaries@mozilla.com at Tue, 21 May 2019 16:39:23 +0000
Bug 1520191 - Polish style for error and warning messages r=jdescottes,ladybenko,flod Wrapped error log in <details> Differential Revision: https://phabricator.services.mozilla.com/D31046
979750dd65da33eebd25999f286b98914639a735: Bug 1552789 - Check NS_FRAME_FORCE_DISPLAY_LIST_DESCEND_INTO flag for every line r=mattwoodrow
Miko Mynttinen <mikokm@gmail.com> - Mon, 20 May 2019 22:54:18 +0000 - rev 474724
Push 113168 by rmaries@mozilla.com at Tue, 21 May 2019 16:39:23 +0000
Bug 1552789 - Check NS_FRAME_FORCE_DISPLAY_LIST_DESCEND_INTO flag for every line r=mattwoodrow Differential Revision: https://phabricator.services.mozilla.com/D31865
07ce2dd7db9d0ff8f68e5ab8d913aa75fe463399: Bug 1498187 - Let the user sort by column in about:performance. r=florian
fluks <fluks.github@gmail.com> - Tue, 21 May 2019 10:46:35 +0000 - rev 474723
Push 113168 by rmaries@mozilla.com at Tue, 21 May 2019 16:39:23 +0000
Bug 1498187 - Let the user sort by column in about:performance. r=florian Differential Revision: https://phabricator.services.mozilla.com/D31612
21b3ac9c491755466306d68e1e198a6891c0e561: Bug 1383650 - Run mach devtools-css-db to sync css db r=emilio
violet <violet.bugreport@gmail.com> - Thu, 16 May 2019 13:21:22 +0000 - rev 474722
Push 113168 by rmaries@mozilla.com at Tue, 21 May 2019 16:39:23 +0000
Bug 1383650 - Run mach devtools-css-db to sync css db r=emilio This is the last part of this seris of patches to implement geometry property. This particular patch just run `./mach devtools-css-db` to update db per instruction at the beginning of devtools/shared/tests/unit/test_css-properties-db.js, and also a manual addition to the animation property db. After this patch, the SVG geometry propery is implemented for <rect>, <circle>, <ellipse> and <foreignObject>. We already implemented outer <svg>. Thus the remainings are inner <svg> and <image>, which are kind of different to the others, so they will be handled in some follow-ups. Note that these patches won't impact old behavior of inner <svg> and <image>. Differential Revision: https://phabricator.services.mozilla.com/D30808
(0) -300000 -100000 -30000 -10000 -3000 -1000 -300 -100 -50 -20 +20 +50 +100 +300 +1000 +3000 +10000 tip