Export of internal Abseil changes

--
e54b9c7bbb0c58475676c268e2e19c69f4bce48a by Jorg Brown <jorg@google.com>:

Tweak ABSL_PREDICT_TRUE slightly, for better code on some platforms and/or
optimization levels.  "false || (x)" is more verbose than "!!(x)", but
ultimately more efficient.

For example, given this code:

void InitIfNecessary() {
  if (ABSL_PREDICT_TRUE(NeedsInit())) {
    SlowInitIfNecessary();
  }
}

Clang with default optimization level will produce:

Before this CL              After this CL
InitIfNecessary:            InitIfNecessary:
  push rbp                    push rbp
  mov  rbp, rsp               mov  rbp, rsp
  call NeedsInit              call NeedsInit
  xor  al, -1
  xor  al, -1
  test al, 1                  test al, 1
  jne  .LBB2_1                jne  .LBB3_1
  jmp  .LBB2_2                jmp  .LBB3_2
.LBB2_1:                    .LBB3_1:
  call SlowInitIfNecessary    call SlowInitIfNecessary
.LBB2_2:                    .LBB3_2:
  pop  rbp                    pop  rbp
  ret                         ret
PiperOrigin-RevId: 276401386

--
0a3c4dfd8342bf2b1b11a87f1c662c883f73cab7 by Abseil Team <absl-team@google.com>:

Fix comment nit: sem_open => sem_init.

The code calls sem_init, not sem_open, to initialize an unnamed semaphore.
(sem_open creates or opens a named semaphore.)

PiperOrigin-RevId: 276344072

--
b36a664e9459057509a90e83d3482e1d3a4c44c7 by Abseil Team <absl-team@google.com>:

Fix typo in flat_hash_map.h: exchaged -> exchanged

PiperOrigin-RevId: 276295792

--
7bbd8d18276eb110c8335743e35fceb662ddf3d6 by Samuel Benzaquen <sbenza@google.com>:

Add assertions to verify use of iterators.

PiperOrigin-RevId: 276283300

--
677398a8ffcb1f59182cffe57a4fe7ff147a0404 by Laramie Leavitt <lar@google.com>:

Migrate distribution_impl.h/cc to generate_real.h/cc.

Combine the methods RandU64To<Float,Double> into a single method:
GenerateRealFromBits().

Remove rejection sampling from absl::uniform_real_distribution.

PiperOrigin-RevId: 276158675

--
c60c9d11d24b0c546329d998e78e15a84b3153f5 by Abseil Team <absl-team@google.com>:

Internal change

PiperOrigin-RevId: 276126962

--
4c840cab6a8d86efa29b397cafaf7520eece68cc by Andy Soffer <asoffer@google.com>:

Update CMakeLists.txt to address https://github.com/abseil/abseil-cpp/issues/365.
This does not cover every platform, but it does at least address the
first-order issue of assuming gcc implies x86.

PiperOrigin-RevId: 276116253

--
98da366e6b5d51afe5d7ac6722126aca23d85ee6 by Abseil Team <absl-team@google.com>:

Internal change

PiperOrigin-RevId: 276097452
GitOrigin-RevId: e54b9c7bbb0c58475676c268e2e19c69f4bce48a
Change-Id: I02d84454bb71ab21ad3d39650acf6cc6e36f58d7

1 files changed, 84 insertions, 0 deletions

diff --git a/absl/base/internal/exponential_biased.cc b/absl/base/internal/exponential_biased.cc
new file mode 100644
index 000000000000..d7ffd184e968
--- /dev/null
+++ b/absl/base/internal/exponential_biased.cc
@@ -0,0 +1,84 @@
+// Copyright 2019 The Abseil Authors.
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     https://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+#include "absl/base/internal/exponential_biased.h"
+
+#include <stdint.h>
+
+#include <atomic>
+#include <cmath>
+#include <limits>
+
+#include "absl/base/attributes.h"
+#include "absl/base/optimization.h"
+
+namespace absl {
+namespace base_internal {
+
+// The algorithm generates a random number between 0 and 1 and applies the
+// inverse cumulative distribution function for an exponential. Specifically:
+// Let m be the inverse of the sample period, then the probability
+// distribution function is m*exp(-mx) so the CDF is
+// p = 1 - exp(-mx), so
+// q = 1 - p = exp(-mx)
+// log_e(q) = -mx
+// -log_e(q)/m = x
+// log_2(q) * (-log_e(2) * 1/m) = x
+// In the code, q is actually in the range 1 to 2**26, hence the -26 below
+int64_t ExponentialBiased::Get(int64_t mean) {
+  if (ABSL_PREDICT_FALSE(!initialized_)) {
+    Initialize();
+  }
+
+  uint64_t rng = NextRandom(rng_);
+  rng_ = rng;
+
+  // Take the top 26 bits as the random number
+  // (This plus the 1<<58 sampling bound give a max possible step of
+  // 5194297183973780480 bytes.)
+  // The uint32_t cast is to prevent a (hard-to-reproduce) NAN
+  // under piii debug for some binaries.
+  double q = static_cast<uint32_t>(rng >> (kPrngNumBits - 26)) + 1.0;
+  // Put the computed p-value through the CDF of a geometric.
+  double interval = (std::log2(q) - 26) * (-std::log(2.0) * mean);
+  // Very large values of interval overflow int64_t. To avoid that, we will cheat
+  // and clamp any huge values to (int64_t max)/2. This is a potential source of
+  // bias, but the mean would need to be such a large value that it's not likely
+  // to come up. For example, with a mean of 1e18, the probability of hitting
+  // this condition is about 1/1000. For a mean of 1e17, standard calculators
+  // claim that this event won't happen.
+  if (interval > static_cast<double>(std::numeric_limits<int64_t>::max() / 2)) {
+    return std::numeric_limits<int64_t>::max() / 2;
+  }
+
+  return static_cast<int64_t>(interval);
+}
+
+void ExponentialBiased::Initialize() {
+  // We don't get well distributed numbers from `this` so we call NextRandom() a
+  // bunch to mush the bits around. We use a global_rand to handle the case
+  // where the same thread (by memory address) gets created and destroyed
+  // repeatedly.
+  ABSL_CONST_INIT static std::atomic<uint32_t> global_rand(0);
+  uint64_t r = reinterpret_cast<uint64_t>(this) +
+               global_rand.fetch_add(1, std::memory_order_relaxed);
+  for (int i = 0; i < 20; ++i) {
+    r = NextRandom(r);
+  }
+  rng_ = r;
+  initialized_ = true;
+}
+
+}  // namespace base_internal
+}  // namespace absl