403e4c993b48a832d50b4f44738c5b5c6d5ce085: Bug 1257446 part.1 ContentCache should store previous character of selection r=m_kato draft
Masayuki Nakano <masayuki@d-toybox.com> - Fri, 22 Jul 2016 20:47:51 +0900 - rev 399826
Push 26006 by masayuki@d-toybox.com at Fri, 12 Aug 2016 06:28:16 +0000
Bug 1257446 part.1 ContentCache should store previous character of selection r=m_kato This patch makes ContentCache store previous character's rect of selection anchor and selection focus because if caret is at end of a line, IME may query the last character of the line. MozReview-Commit-ID: 5X1K8KtrYfl
77d541f21b78b4270c109f2a4e9db59c19129e03: Bug 1257446 part.0 ContentCacheInParent::HandleQueryContentEvent() should log the cause of failure when it makes the event's input offset absolute r=m_kato draft
Masayuki Nakano <masayuki@d-toybox.com> - Tue, 12 Jul 2016 17:44:44 +0900 - rev 399825
Push 26006 by masayuki@d-toybox.com at Fri, 12 Aug 2016 06:28:16 +0000
Bug 1257446 part.0 ContentCacheInParent::HandleQueryContentEvent() should log the cause of failure when it makes the event's input offset absolute r=m_kato MozReview-Commit-ID: JVGHPQ5gC3G
9507ccd2dfdf4fe039435d04c1f79d490be785fd: Bug 1286157 TSFTextStore should use insertion point relative offset query when cached contents for TSF and actual content (or content cache) are different r=m_kato draft
Masayuki Nakano <masayuki@d-toybox.com> - Tue, 12 Jul 2016 16:52:48 +0900 - rev 399824
Push 26005 by masayuki@d-toybox.com at Fri, 12 Aug 2016 06:25:43 +0000
Bug 1286157 TSFTextStore should use insertion point relative offset query when cached contents for TSF and actual content (or content cache) are different r=m_kato MozReview-Commit-ID: 3Q9T3XVvyCj
a27248675e81c6f207b6433fe0a67a1ba83b5283: Bug 1294585 - Move --with-*-keyfile options to python configure. r?chmanchester draft
Mike Hommey <mh+mozilla@glandium.org> - Tue, 09 Aug 2016 18:15:53 +0900 - rev 399823
Push 26004 by bmo:mh+mozilla@glandium.org at Fri, 12 Aug 2016 06:25:34 +0000
Bug 1294585 - Move --with-*-keyfile options to python configure. r?chmanchester
f0b21956f737f9520e8e253f20b7972b6c08c219: Bug 1294585 - Remove the --with-google-oauth-api-keyfile configure flag. r=mikedeboer draft
Mike Hommey <mh+mozilla@glandium.org> - Tue, 09 Aug 2016 19:43:40 +0900 - rev 399822
Push 26004 by bmo:mh+mozilla@glandium.org at Fri, 12 Aug 2016 06:25:34 +0000
Bug 1294585 - Remove the --with-google-oauth-api-keyfile configure flag. r=mikedeboer
177391a25a6b2a3ddc86bea56dbb18da0c9491f8: Bug 1294585 - Remove --with-google-oauth-api-keyfile from mozconfigs. r=mikedeboer draft
Mike Hommey <mh+mozilla@glandium.org> - Tue, 09 Aug 2016 19:40:59 +0900 - rev 399821
Push 26004 by bmo:mh+mozilla@glandium.org at Fri, 12 Aug 2016 06:25:34 +0000
Bug 1294585 - Remove --with-google-oauth-api-keyfile from mozconfigs. r=mikedeboer The flag has not had an effect since bug 1213984, and the file given as argument is not even on the build slaves anymore.
5e422fe55ef64121b5277aa43a7b14c2e07f0269: Bug 1293840 - Determine HAVE_64BIT_BUILD from --target. r?chmanchester draft
Mike Hommey <mh+mozilla@glandium.org> - Wed, 10 Aug 2016 11:39:16 +0900 - rev 399820
Push 26004 by bmo:mh+mozilla@glandium.org at Fri, 12 Aug 2016 06:25:34 +0000
Bug 1293840 - Determine HAVE_64BIT_BUILD from --target. r?chmanchester Until now, HAVE_64BIT_BUILD was entirely determined by a compiler check. But we didn't run the check on e.g. artifact builds, while relying on its result for some non-compilation related things, leading to subtle discrepancies. This changes the configure check to derive HAVE_64BIT_BUILD from bitness determined by the target CPU, and double checked with a compiler check.
06558d5af091c5f6da56070b4c8baa764254a228: Bug 1293837 - Move HAVE_64BIT_BUILD test to python configure. r?chmanchester draft
Mike Hommey <mh+mozilla@glandium.org> - Wed, 10 Aug 2016 09:00:23 +0900 - rev 399819
Push 26004 by bmo:mh+mozilla@glandium.org at Fri, 12 Aug 2016 06:25:34 +0000
Bug 1293837 - Move HAVE_64BIT_BUILD test to python configure. r?chmanchester Also, now that we're using modern C++11 compilers, we can just rely on static_assert, instead of the pile of macros used in the autoconf test.
6c16daa1ef8dcb77891f2fd20ae67b9ab1579811: Bug 1293837 - Change try_compile() to return True when the test passes. r?chmanchester draft
Mike Hommey <mh+mozilla@glandium.org> - Wed, 10 Aug 2016 10:00:43 +0900 - rev 399818
Push 26004 by bmo:mh+mozilla@glandium.org at Fri, 12 Aug 2016 06:25:34 +0000
Bug 1293837 - Change try_compile() to return True when the test passes. r?chmanchester Currently, it returns either None or the contents of the compiler's stdout, which is always expected to be an empty string, and is not very useful. So instead, return True in the latter case.
3fe6b6caeb0919cd695a9e231f25552c3f0794bd: Bug 1293579 - Use the new `dependable` function in a few places. r?chmanchester draft
Mike Hommey <mh+mozilla@glandium.org> - Tue, 09 Aug 2016 15:55:10 +0900 - rev 399817
Push 26004 by bmo:mh+mozilla@glandium.org at Fri, 12 Aug 2016 06:25:34 +0000
Bug 1293579 - Use the new `dependable` function in a few places. r?chmanchester
8c968870824ef74420982502852c6c26489a6c7c: Bug 1293579 - Add a `dependable` function/decorator to avoid some of the depends('--help') weirdness in python configure. r?chmanchester draft
Mike Hommey <mh+mozilla@glandium.org> - Tue, 09 Aug 2016 15:54:20 +0900 - rev 399816
Push 26004 by bmo:mh+mozilla@glandium.org at Fri, 12 Aug 2016 06:25:34 +0000
Bug 1293579 - Add a `dependable` function/decorator to avoid some of the depends('--help') weirdness in python configure. r?chmanchester
97128589cc388a6c293cf89ebdde83a929bcccc8: Bug 1290334 - Use a LD variable instead of LINK. r?wtc draft
Mike Hommey <mh@glandium.org> - Fri, 29 Jul 2016 21:13:39 +0900 - rev 399815
Push 26004 by bmo:mh+mozilla@glandium.org at Fri, 12 Aug 2016 06:25:34 +0000
Bug 1290334 - Use a LD variable instead of LINK. r?wtc MSVC's link.exe allows to pass linker flags through the LINK environment variable[1]. When running under a shell that has the GNU coreutils's link.exe installed (a tool to create hard links), it can become necessary to override LINK from the command line. In that case, MAKE passes down LINK as an environment variable to submake and subprocesses, which then means it's passed down to link.exe itself, which then tries to link itself to the rest of the passed arguments, and fails, presumably because it can't open the executable while it's being executed (and even if it could, that would still be an invalid thing to link). So, change the variable name. 1. https://msdn.microsoft.com/en-us/library/6y6t9esh.aspx
0d745c5a062241a927a2406ca0a710a9186bb1be: Bug 1290334 - Use single quotes for core_abspath. r=wtc draft
Mike Hommey <mh@glandium.org> - Fri, 29 Jul 2016 20:54:48 +0900 - rev 399814
Push 26004 by bmo:mh+mozilla@glandium.org at Fri, 12 Aug 2016 06:25:34 +0000
Bug 1290334 - Use single quotes for core_abspath. r=wtc Double quotes allow all sorts of escaping, which are not necessarily wanted. Moreover, it forces make to run a subshell instead of executing the command directly on its own (because of the possibility of there being something escaped between the double quotes), and that causes problems when running under MINGW make with a MSYS shell. When using single quotes, those problems go away, because make runs the subprocess without an intermediate shell, and avoids a painful MINGW->MSYS->MINGW transition.
dbf40c85a61d7103c899a5d934afaed238d04103: Bug 1290334 - Use $(CURDIR) for PWD when it's a Windows path. r=wtc draft
Mike Hommey <mh@glandium.org> - Fri, 29 Jul 2016 20:50:58 +0900 - rev 399813
Push 26004 by bmo:mh+mozilla@glandium.org at Fri, 12 Aug 2016 06:25:34 +0000
Bug 1290334 - Use $(CURDIR) for PWD when it's a Windows path. r=wtc
e548a933d0cb4d1f85846fd1c76f4426f8dda228: Bug 1290334 - Only build SSE2 chacha20 implementation with GCC/Clang. r?wtc draft
Mike Hommey <mh@glandium.org> - Fri, 29 Jul 2016 20:37:08 +0900 - rev 399812
Push 26004 by bmo:mh+mozilla@glandium.org at Fri, 12 Aug 2016 06:25:34 +0000
Bug 1290334 - Only build SSE2 chacha20 implementation with GCC/Clang. r?wtc The SSE2 code in chacha20_vec.c relies on syntax that only builds with GCC/Clang and not e.g. MSVC. When building with MSVC and CPU_ARCH is x86_64, the generic non-SSE2 chacha20.c needs to be used. Only build chacha20_vec.c when the compiler is GCC or Clang.
88c97731bde5bcaaee930375225357fadcb4a0cf: Bug 1290334 - Force-set CPU_ARCH to x386 or x86_64 depending on USE_64. r=wtc draft
Mike Hommey <mh@glandium.org> - Fri, 29 Jul 2016 20:30:56 +0900 - rev 399811
Push 26004 by bmo:mh+mozilla@glandium.org at Fri, 12 Aug 2016 06:25:34 +0000
Bug 1290334 - Force-set CPU_ARCH to x386 or x86_64 depending on USE_64. r=wtc Let's put things as they are: the NSS build system is a mess. It is not consistent across platforms, and sometimes relies on user input, sometimes not. On Windows x86/x86_64, the USE_64 variable essentially chooses between actual x86 and x86_64. Except there are CPU_ARCH based tests in cross-platform Makefiles, and CPU_ARCH doesn't necessarily match what USE_64 says is wanted (not even mentioning what the compiler actually will generate). For instance, building a 32-bits NSS under MINGW64 will say CPU_ARCH is x86_64. This change makes CPU_ARCH match the reality imposed by USE_64 on Windows.
d557cd9b7fa70be38765d366ed84d9228a0a059f: Bug 1290334 - Support building under MINGW64. r?wtc draft
Mike Hommey <mh@glandium.org> - Fri, 29 Jul 2016 20:25:21 +0900 - rev 399810
Push 26004 by bmo:mh+mozilla@glandium.org at Fri, 12 Aug 2016 06:25:34 +0000
Bug 1290334 - Support building under MINGW64. r?wtc
02141d0351fed13066010a4e1a63835946964725: Bug 1294384 - make video-only attribute orthogonal to the seek type; r?jwwang draft
Kaku Kuo <tkuo@mozilla.com> - Fri, 12 Aug 2016 11:45:02 +0800 - rev 399809
Push 26003 by tkuo@mozilla.com at Fri, 12 Aug 2016 06:23:15 +0000
Bug 1294384 - make video-only attribute orthogonal to the seek type; r?jwwang MozReview-Commit-ID: 2oBjuiN7pLn
3e55b17c30c451717ad931e79875516fc8b01c35: Bug 1294615. Part 2 - Refactor MDSM::SetDormant() to remove unnecessary checks for mQueuedSeek.Exists() and mCurrentSeek.Exists(). draft
JW Wang <jwwang@mozilla.com> - Thu, 11 Aug 2016 17:33:07 +0800 - rev 399808
Push 26002 by jwwang@mozilla.com at Fri, 12 Aug 2016 06:22:54 +0000
Bug 1294615. Part 2 - Refactor MDSM::SetDormant() to remove unnecessary checks for mQueuedSeek.Exists() and mCurrentSeek.Exists(). We've proven that when seek is in action, mQueuedSeek.Exists() is always false and mCurrentSeek.Exists() is always true. MozReview-Commit-ID: CxNU45NXG88
dc006ed31315c00f3a4096d49fa3323155e3be48: Bug 1294615. Part 1 - Assert mQueuedSeek.Exists() is false in InitiateSeek(). draft
JW Wang <jwwang@mozilla.com> - Thu, 11 Aug 2016 16:07:45 +0800 - rev 399807
Push 26002 by jwwang@mozilla.com at Fri, 12 Aug 2016 06:22:54 +0000
Bug 1294615. Part 1 - Assert mQueuedSeek.Exists() is false in InitiateSeek(). InitiateSeek() is called from several places where we can prove mQueuedSeek.Exists() is false: 1. MaybeFinishDecodeFirstFrame(): mQueuedSeek is moved when calling InitiateSeek(). 2. Seek() rejects mQueuedSeek before calling InitiateSeek(). 3. StartDecoding(): mQueuedSeek is moved when calling InitiateSeek(). 4. VisibilityChanged(): it doesn't call InitiateSeek() when mQueuedSeek.Exists() is true. MozReview-Commit-ID: BriPw0VID5O
(0) -300000 -100000 -30000 -10000 -3000 -1000 -300 -100 -50 -20 +20 +50 +100 +300 +1000 +3000 +10000 +30000 +100000 +300000 tip