Bug 903389 - Make Make NativeGet[Getter]PureInline handle dense/typed array shapes. r=arai

Bug 903389 - Fix uses of ClassMethodIsNative. r=Waldo

Bug 1323303 - require building with Skia. r=glandium MozReview-Commit-ID: 56oawitOHEP

Bug 1029245 - part 2 - upgrade builds to use GCC 4.9.4; r=glandium

Bug 1029245 - part 1 - modify build-gcc.sh to build GCC 4.9.4; r=glandium PR 64905 apparently never got backported to 4.9.x, so we still need the patch for that.

Bug 1029245 - part 0 - tweak Skia's SkOnce.h header to work around issues with std::atomic::compare_exchange_strong; r=lsalzman Building Skia inside of mozilla-central with GCC 4.9.4 causes problems: [...]c++/4.9.4/bits/atomic_base.h:581:70: error: failure memory model cannot be stronger than success memory model for '__atomic_compare_exchange' The error stack accompanying this message points at SkEventTracer::GetInstance: SkEventTracer* SkEventTracer::GetInstance() { if (SkEventTracer* tracer = sk_atomic_load(&gUserTracer, sk_memory_order_acquire)) { return tracer; } static SkOnce once; static SkDefaultEventTracer* defaultTracer; once([] { defaultTracer = new SkDefaultEventTracer; }); return defaultTracer; } The only place that compare_exchange_strong could be called here is from SkOnce::operator(): template <typename Fn, typename... Args> void operator()(Fn&& fn, Args&&... args) { auto state = fState.load(std::memory_order_acquire); if (state == Done) { return; } // If it looks like no one has started calling fn(), try to claim that job. if (state == NotStarted && fState.compare_exchange_strong(state, Claimed, std::memory_order_relaxed)) { // Great! We'll run fn() then notify the other threads by releasing Done into fState. fn(std::forward<Args>(args)...); return fState.store(Done, std::memory_order_release); } [...code elided...] where |fState| is an atomic<uint8_t>. The three-argument form of atomic<uint8_t>::compare_exchange_strong is defined as: _GLIBCXX_ALWAYS_INLINE bool compare_exchange_strong(__int_type& __i1, __int_type __i2, memory_order __m = memory_order_seq_cst) noexcept { return compare_exchange_strong(__i1, __i2, __m, __cmpexch_failure_order(__m)); } __cmpexch_failure_order relaxes the given memory_order: // Drop release ordering as per [atomics.types.operations.req]/21 constexpr memory_order __cmpexch_failure_order2(memory_order __m) noexcept { return __m == memory_order_acq_rel ? memory_order_acquire : __m == memory_order_release ? memory_order_relaxed : __m; } constexpr memory_order __cmpexch_failure_order(memory_order __m) noexcept { return memory_order(__cmpexch_failure_order2(__m & __memory_order_mask) | (__m & __memory_order_modifier_mask)); } which then gets us to the four-argument version of compare_exchange_strong: _GLIBCXX_ALWAYS_INLINE bool compare_exchange_strong(__int_type& __i1, __int_type __i2, memory_order __m1, memory_order __m2) noexcept { memory_order __b2 = __m2 & __memory_order_mask; memory_order __b1 = __m1 & __memory_order_mask; __glibcxx_assert(__b2 != memory_order_release); __glibcxx_assert(__b2 != memory_order_acq_rel); __glibcxx_assert(__b2 <= __b1); return __atomic_compare_exchange_n(&_M_i, &__i1, __i2, 0, __m1, __m2); } Despite the constexpr annotation on __cmpexch_failure_order and friends, which ought to imply that they get constant-folded, I think what is happening is that GCC doesn't see |memory_order_relaxed| when it examines __m2. Instead, it seems some internal tree representation for the call to __cmpexch_failure_order. Since this is not an integer constant, GCC treats __m2 as being equivalent to memory_order_seq_cst (see gcc/builtins.c:get_memmodel). And since memory_order_seq_cst is stronger than memory_order_relaxed, we get the above error. In any event, the easiest fix is to simply use the four-argument form of compare_exchange_strong directly, explicitly specifying the failure memory order.

Backed out changeset 28c47fbedac4 (bug 1323303) for bustage

Bug 1315906 - Add test coverage. r=gfritzsche MozReview-Commit-ID: FFskMOrwvLO

Bug 1315906 - Change JS histogram add functions so that it doesn't throw. r=gfritzsche

Bug 1323303 - require building with Skia. r=glandium MozReview-Commit-ID: 56oawitOHEP

Bug 1318965 - Improves the logging functions provided in ClearKeyUtils r?cpearce MozReview-Commit-ID: LRimaj7hMh7

Bug 1318965 - Fixes a bug in the WideVineAdapter wherein session load failures were not adapted correctly r?cpearce MozReview-Commit-ID: 2ze4d8EuHr9

Bug 1318965 - Adds code for deferring the initialization of sessions until persistent sessions have been loaded r?cpearce MozReview-Commit-ID: F46wvAmlbep

Bug 1318965 - Converts the way keys are persisted to WideVine's system. r?cpearce MozReview-Commit-ID: KyWprtSjfah

Bug 1322041 - DO NOT LAND - changes applied only to date branch MozReview-Commit-ID: 7SL5RZv4Y3X

Bug 1322041 - Sign nightly l10n jobs. r=dustin Involved work which landed on the date project branch with: * Bug 1312000, by Callek@gmail.com (https://hg.mozilla.org/projects/date/rev/79a2f66ff5c0) * Bug 1312500, by Callek@gmail.com (https://hg.mozilla.org/projects/date/rev/6ae07fa4b011) * Bug 1312585, by Callek@gmail.com (https://hg.mozilla.org/projects/date/rev/7cf267b61752) * Bug 1313662, by Callek@gmail.com (https://hg.mozilla.org/projects/date/rev/92f531d1cd77) * Bug 1314008, by Callek@gmail.com (https://hg.mozilla.org/projects/date/rev/bbc7608106c0) * Bug 1314847, by Callek@gmail.com (https://hg.mozilla.org/projects/date/rev/4a1231655fbb) * Bug 1316214, by Callek@gmail.com (https://hg.mozilla.org/projects/date/rev/edae37481cab, https://hg.mozilla.org/projects/date/rev/62bd3371e954, https://hg.mozilla.org/projects/date/rev/2cc558343951) * Bug 1319189, by Callek@gmail.com (https://hg.mozilla.org/projects/date/rev/4c33f8ccecf5) MozReview-Commit-ID: 3hadIi757jn

Bug 1322041 - Support SIMPLE_NAMES in a forced-way (via mozharness) for l10n. r=jlund Involved work which landed on the date project branch with: * Bug 1313662, by Callek@gmail.com (https://hg.mozilla.org/projects/date/rev/b8b3192cf5d9, https://hg.mozilla.org/projects/date/rev/92f531d1cd77) MozReview-Commit-ID: ItcVPN6KRd

Bug 1322041 - Add chain of trust to l10n jobs. r=dustin Involved work which landed on the date project branch with: * No Bug, by asasaki@mozilla.com (https://hg.mozilla.org/projects/date/rev/b82459a453a6) MozReview-Commit-ID: Fmg2nouDedc

Bug 1322041 - Nightly l10n support, and docs. r=dustin Involved work which landed on the date project branch with: * No Bug, by Callek@gmail.com (https://hg.mozilla.org/projects/date/rev/e7af20515b1c, https://hg.mozilla.org/projects/date/rev/d7f0f1c6d564) * Bug 1171738, by Callek@gmail.com (https://hg.mozilla.org/projects/date/rev/a906f6997731, https://hg.mozilla.org/projects/date/rev/ada87315d4c0, https://hg.mozilla.org/projects/date/rev/fbee52f145e9, https://hg.mozilla.org/projects/date/rev/2939b68e59fc) * Bug 1312000, by Callek@gmail.com (https://hg.mozilla.org/projects/date/rev/79a2f66ff5c0) * Bug 1312396, by Callek@gmail.com (https://hg.mozilla.org/projects/date/rev/dc06c58a5663) * Bug 1312585, by Callek@gmail.com (https://hg.mozilla.org/projects/date/rev/7cf267b61752) * Bug 1316527, by Callek@gmail.com (https://hg.mozilla.org/projects/date/rev/cc2d9365d9b9) * Bug 1318068, by Callek@gmail.com (https://hg.mozilla.org/projects/date/rev/7cb11cb76a09) MozReview-Commit-ID: Fhom5OcIWru

Bug 1322041 - Add Docs for L10n and add locales_file parsing. r=dustin Involved work which landed on the date project branch with: * Bug 1312585, by Callek@gmail.com (https://hg.mozilla.org/projects/date/rev/dc1926812404, https://hg.mozilla.org/projects/date/rev/7cf267b61752) * Bug 1171738, by Callek@gmail.com (https://hg.mozilla.org/projects/date/rev/a906f6997731) * Bug 1314008, by Callek@gmail.com (https://hg.mozilla.org/projects/date/rev/bbc7608106c0, https://hg.mozilla.org/projects/date/rev/a80373508881) MozReview-Commit-ID: 3MVTRgzLYc0