e44777d69e7a949a3fbaaae37e3c7986c9847e49: Bug 1158920 - Yield to the event loop before reading the spelling suggestions
Ehsan Akhgari <ehsan@mozilla.com> - Mon, 27 Apr 2015 15:08:10 -0400 - rev 241241
Push 59064 by eakhgari@mozilla.com at Mon, 27 Apr 2015 19:09:35 +0000
Bug 1158920 - Yield to the event loop before reading the spelling suggestions
531631f340210c8e885b0f9a34ec00e1d474a15b: Bug 1058695 - Add member to nsIGlobalObject to detect it is going away. Make promises use it. r=bholley
Nikhil Marathe <nsm.nikhil@gmail.com> - Mon, 27 Apr 2015 12:00:41 -0700 - rev 241240
Push 59063 by nsm.nikhil@gmail.com at Mon, 27 Apr 2015 19:08:01 +0000
Bug 1058695 - Add member to nsIGlobalObject to detect it is going away. Make promises use it. r=bholley
b2dd209a43df1379f9f7e7272032ec4b94585b32: Bug 1157648 - Make nsScriptError::ToString use only the first 512 characters of mSourceName and mSourceLine. r=bholley
L. David Baron <dbaron@dbaron.org> - Mon, 27 Apr 2015 11:35:10 -0700 - rev 241239
Push 59062 by dbaron@mozilla.com at Mon, 27 Apr 2015 18:37:34 +0000
Bug 1157648 - Make nsScriptError::ToString use only the first 512 characters of mSourceName and mSourceLine. r=bholley This means that when mSourceName and mSourceLine are large, ToString is not excessively expensive. This is particularly important for CSS errors, where we don't make an attempt to truncate these prior to constructing the script error, but we do ensure that when we report multiple errors on the same line (which is common for minified CSS), we share from the same string buffer and avoid copying.
f284c64e55371a5ecbbb86d0790e9ce418767ef8: Bug 1158896 - prevent dead code from being generated in lower.py; r=bent
Nathan Froyd <froydnj@mozilla.com> - Mon, 27 Apr 2015 13:08:12 -0400 - rev 241238
Push 59061 by nfroyd@mozilla.com at Mon, 27 Apr 2015 18:29:45 +0000
Bug 1158896 - prevent dead code from being generated in lower.py; r=bent Coverity was complaining that we have things like: if ((!(actor))) { return false; } if ((!(actor))) { return false; } in the generated code, as the second return will clearly never be hit. To address this, let's remove a redundant call to dtorPrologue.
4f898b4ffd633ed3dc2c75cd5e38e8016b2e3ea6: Bug 1158772 - fix non-idiomatic memset call in nsDeque.cpp; r=erahm
Nathan Froyd <froydnj@mozilla.com> - Mon, 27 Apr 2015 07:55:32 -0400 - rev 241237
Push 59061 by nfroyd@mozilla.com at Mon, 27 Apr 2015 18:29:45 +0000
Bug 1158772 - fix non-idiomatic memset call in nsDeque.cpp; r=erahm Coverty complains that we're using sizeof(mData) here instead of sizeof(*mData). They're equivalent for all the architectures we care about, but go ahead and tidy up the syntax to silence the static analyzer.
d02374f93b37167b58a34a8f78ee785089027c3c: Backed out changeset c839f283bf28 (bug 1058695) for bustage.
Ryan VanderMeulen <ryanvm@gmail.com> - Mon, 27 Apr 2015 14:04:24 -0400 - rev 241236
Push 59060 by ryanvm@gmail.com at Mon, 27 Apr 2015 18:04:19 +0000
Backed out changeset c839f283bf28 (bug 1058695) for bustage. CLOSED TREE
600207b620204b60e41164b70e552cf764a80b7c: Bug 1158257 - Gecko should always provide SpiderMonkey with a debuggerMallocSizeOf; r=mccr8
Nick Fitzgerald <fitzgen@gmail.com> - Mon, 27 Apr 2015 10:58:24 -0700 - rev 241235
Push 59059 by nfitzgerald@mozilla.com at Mon, 27 Apr 2015 17:58:48 +0000
Bug 1158257 - Gecko should always provide SpiderMonkey with a debuggerMallocSizeOf; r=mccr8
ada25868ffa642d6a5b0e4c0db2e1f769656e09f: Bug 1134865 - Part 3: Add JS::ubi::Node::jsObjectConstructorName; r=djvj
Nick Fitzgerald <fitzgen@gmail.com> - Mon, 27 Apr 2015 10:57:06 -0700 - rev 241234
Push 59058 by nfitzgerald@mozilla.com at Mon, 27 Apr 2015 17:57:27 +0000
Bug 1134865 - Part 3: Add JS::ubi::Node::jsObjectConstructorName; r=djvj
626f0dd608b0a52e58e9fa9a655db2819f0c51a9: Bug 1134865 - Part 2: Add constructor name to the allocations log; r=djvj
Nick Fitzgerald <fitzgen@gmail.com> - Mon, 27 Apr 2015 10:57:06 -0700 - rev 241233
Push 59058 by nfitzgerald@mozilla.com at Mon, 27 Apr 2015 17:57:27 +0000
Bug 1134865 - Part 2: Add constructor name to the allocations log; r=djvj
50dedbefd46e856541d39560d1c68725157621f6: Bug 1134865 - Part 1: Add JSObject::constructorDisplayAtom; r=djvj
Nick Fitzgerald <fitzgen@gmail.com> - Mon, 27 Apr 2015 10:57:06 -0700 - rev 241232
Push 59058 by nfitzgerald@mozilla.com at Mon, 27 Apr 2015 17:57:27 +0000
Bug 1134865 - Part 1: Add JSObject::constructorDisplayAtom; r=djvj
ae91360cc160b403c82f2a7704a378707782ea7c: Bug 1128586 - Prefer hg.exe over hg; r=RyanVM
Gregory Szorc <gps@mozilla.com> - Mon, 27 Apr 2015 10:43:49 -0700 - rev 241231
Push 59057 by gszorc@mozilla.com at Mon, 27 Apr 2015 17:44:14 +0000
Bug 1128586 - Prefer hg.exe over hg; r=RyanVM This is needed for compatibility with an upcoming release of MozillaBuild, which distributes Mercurial as a Python package, not as a standalone Windows program. As a result, it introduces "hg" into $PATH, which "which" will happily prefer as the "hg" binary. This upsets subprocess. So, we explicitly prefer "hg.exe" over "hg". We could accomplish the same thing by calling which.whichall() and sorting results. But this is more code and IMO not worth the effort to implement.
d439522e45ae53b0c078979c77e89b202d2fe26c: Bug 1128586 - Properly look for Mercurial version; r=RyanVM
Gregory Szorc <gps@mozilla.com> - Mon, 27 Apr 2015 10:43:25 -0700 - rev 241230
Push 59057 by gszorc@mozilla.com at Mon, 27 Apr 2015 17:44:14 +0000
Bug 1128586 - Properly look for Mercurial version; r=RyanVM `mercurial version` doesn't always print the version number. `mercurial --version` does. Use the latter.
c839f283bf289e73bbfb66995b5e9f31ba48d204: Bug 1058695 - Add member to nsIGlobalObject to detect it is going away. Make promises use it. r=bholley
Nikhil Marathe <nsm.nikhil@gmail.com> - Wed, 22 Apr 2015 16:34:21 -0700 - rev 241229
Push 59056 by nsm.nikhil@gmail.com at Mon, 27 Apr 2015 17:41:03 +0000
Bug 1058695 - Add member to nsIGlobalObject to detect it is going away. Make promises use it. r=bholley
c2c74fe15d1e19cd75e2e9685634e54c9c92ceb5: Bug 1155984. Improve the performance of the "self" getter in both window and workers. r=peterv,jorendorff
Boris Zbarsky <bzbarsky@mit.edu> - Mon, 27 Apr 2015 13:38:12 -0400 - rev 241228
Push 59055 by bzbarsky@mozilla.com at Mon, 27 Apr 2015 17:39:04 +0000
Bug 1155984. Improve the performance of the "self" getter in both window and workers. r=peterv,jorendorff
1bfcd43acd3c9454ff8ed1d1a40907793d5c8e73: Bug 892973 - Add support for the YouCompleteMe vim plugin; r=gps
Ehsan Akhgari <ehsan@mozilla.com> - Fri, 24 Apr 2015 15:12:50 -0400 - rev 241227
Push 59054 by eakhgari@mozilla.com at Mon, 27 Apr 2015 17:08:52 +0000
Bug 892973 - Add support for the YouCompleteMe vim plugin; r=gps
fe8d97018a57e96f51a0ad3ab1e423f9e34e7f56: Bug 1158442 - Remove the "Performance Entry buffer size maximum reached" warning; r=baku
Ehsan Akhgari <ehsan@mozilla.com> - Fri, 24 Apr 2015 23:59:42 -0400 - rev 241226
Push 59054 by eakhgari@mozilla.com at Mon, 27 Apr 2015 17:08:52 +0000
Bug 1158442 - Remove the "Performance Entry buffer size maximum reached" warning; r=baku This happens far too often in our tests. For example, in our mochitest-3 debug mac test, we hit this warning 4,472 times. At this point, this warning is only adding unneeded noise, slowing down the tests and increasing the size of the logs unnecessarily. Let's remove it.
5496344c05e9947d961d8ae939ed49ddbfc93b04: Bug 1136780 - Re-enable test_fetch_event.html now that bug 1156771 has been fixed; r=nsm
Ehsan Akhgari <ehsan@mozilla.com> - Tue, 21 Apr 2015 21:30:24 -0400 - rev 241225
Push 59053 by eakhgari@mozilla.com at Mon, 27 Apr 2015 17:04:38 +0000
Bug 1136780 - Re-enable test_fetch_event.html now that bug 1156771 has been fixed; r=nsm
b0480e99fa3838dedb90eb67fe38551177528b0e: Bug 1156771 - Part 2: Assert that the cache entry for the intercepted doesn't exist in the cache storage before we try to open it for the first time; r=michal
Ehsan Akhgari <ehsan@mozilla.com> - Tue, 21 Apr 2015 21:25:30 -0400 - rev 241224
Push 59053 by eakhgari@mozilla.com at Mon, 27 Apr 2015 17:04:38 +0000
Bug 1156771 - Part 2: Assert that the cache entry for the intercepted doesn't exist in the cache storage before we try to open it for the first time; r=michal
e0f7cbf55fdf504001b81475b9d1d38f2e9ce433: Bug 1156771 - Part 1: Ensure that each channel uses a unique ID to compute its cache entry extension; r=michal
Ehsan Akhgari <ehsan@mozilla.com> - Tue, 21 Apr 2015 21:13:30 -0400 - rev 241223
Push 59053 by eakhgari@mozilla.com at Mon, 27 Apr 2015 17:04:38 +0000
Bug 1156771 - Part 1: Ensure that each channel uses a unique ID to compute its cache entry extension; r=michal This makes sure that concurrent synthesized HTTP channels for the same URL do not try to use the same cache entry accidentally. We have so far observed this issue as an intermittent test failure (see bug 1136780), and it's hard to test this standalone, so enabling that test will serve as an automated test for this patch as well.
333aaffb8ad7eb94ff1a1cdc9f9e4cd76e16b61c: Bug 1156533 - Simplify how we trace Shapes for marking; r=sfink
Terrence Cole <terrence@mozilla.com> - Thu, 16 Apr 2015 13:45:24 -0700 - rev 241222
Push 59052 by tcole@mozilla.com at Mon, 27 Apr 2015 16:27:49 +0000
Bug 1156533 - Simplify how we trace Shapes for marking; r=sfink
(0) -100000 -30000 -10000 -3000 -1000 -300 -100 -50 -20 +20 +50 +100 +300 +1000 +3000 +10000 +30000 +100000 tip