aboutsummaryrefslogtreecommitdiffstats
path: root/contrib/restricted/abseil-cpp/absl/time/duration.cc
diff options
context:
space:
mode:
authoranastasy888 <anastasy888@yandex-team.ru>2022-02-10 16:45:54 +0300
committerDaniil Cherednik <dcherednik@yandex-team.ru>2022-02-10 16:45:54 +0300
commit49f765d71da452ea93138a25559dfa68dd76c7f3 (patch)
tree1016041feb637349e401dcc0fa85217dd2c2c639 /contrib/restricted/abseil-cpp/absl/time/duration.cc
parent7353a3fdea9c67c256980c00a2b3b67f09b23a27 (diff)
downloadydb-49f765d71da452ea93138a25559dfa68dd76c7f3.tar.gz
Restoring authorship annotation for <anastasy888@yandex-team.ru>. Commit 1 of 2.
Diffstat (limited to 'contrib/restricted/abseil-cpp/absl/time/duration.cc')
-rw-r--r--contrib/restricted/abseil-cpp/absl/time/duration.cc1718
1 files changed, 859 insertions, 859 deletions
diff --git a/contrib/restricted/abseil-cpp/absl/time/duration.cc b/contrib/restricted/abseil-cpp/absl/time/duration.cc
index 4443109a51..77ed7af93e 100644
--- a/contrib/restricted/abseil-cpp/absl/time/duration.cc
+++ b/contrib/restricted/abseil-cpp/absl/time/duration.cc
@@ -1,720 +1,720 @@
-// Copyright 2017 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.
-
-// The implementation of the absl::Duration class, which is declared in
-// //absl/time.h. This class behaves like a numeric type; it has no public
-// methods and is used only through the operators defined here.
-//
-// Implementation notes:
-//
-// An absl::Duration is represented as
-//
-// rep_hi_ : (int64_t) Whole seconds
-// rep_lo_ : (uint32_t) Fractions of a second
-//
-// The seconds value (rep_hi_) may be positive or negative as appropriate.
-// The fractional seconds (rep_lo_) is always a positive offset from rep_hi_.
-// The API for Duration guarantees at least nanosecond resolution, which
-// means rep_lo_ could have a max value of 1B - 1 if it stored nanoseconds.
-// However, to utilize more of the available 32 bits of space in rep_lo_,
-// we instead store quarters of a nanosecond in rep_lo_ resulting in a max
-// value of 4B - 1. This allows us to correctly handle calculations like
-// 0.5 nanos + 0.5 nanos = 1 nano. The following example shows the actual
-// Duration rep using quarters of a nanosecond.
-//
-// 2.5 sec = {rep_hi_=2, rep_lo_=2000000000} // lo = 4 * 500000000
-// -2.5 sec = {rep_hi_=-3, rep_lo_=2000000000}
-//
-// Infinite durations are represented as Durations with the rep_lo_ field set
-// to all 1s.
-//
-// +InfiniteDuration:
-// rep_hi_ : kint64max
-// rep_lo_ : ~0U
-//
-// -InfiniteDuration:
-// rep_hi_ : kint64min
-// rep_lo_ : ~0U
-//
-// Arithmetic overflows/underflows to +/- infinity and saturates.
-
-#if defined(_MSC_VER)
-#include <winsock2.h> // for timeval
-#endif
-
-#include <algorithm>
-#include <cassert>
-#include <cctype>
-#include <cerrno>
-#include <cmath>
-#include <cstdint>
-#include <cstdlib>
-#include <cstring>
-#include <ctime>
-#include <functional>
-#include <limits>
-#include <string>
-
-#include "absl/base/casts.h"
+// Copyright 2017 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.
+
+// The implementation of the absl::Duration class, which is declared in
+// //absl/time.h. This class behaves like a numeric type; it has no public
+// methods and is used only through the operators defined here.
+//
+// Implementation notes:
+//
+// An absl::Duration is represented as
+//
+// rep_hi_ : (int64_t) Whole seconds
+// rep_lo_ : (uint32_t) Fractions of a second
+//
+// The seconds value (rep_hi_) may be positive or negative as appropriate.
+// The fractional seconds (rep_lo_) is always a positive offset from rep_hi_.
+// The API for Duration guarantees at least nanosecond resolution, which
+// means rep_lo_ could have a max value of 1B - 1 if it stored nanoseconds.
+// However, to utilize more of the available 32 bits of space in rep_lo_,
+// we instead store quarters of a nanosecond in rep_lo_ resulting in a max
+// value of 4B - 1. This allows us to correctly handle calculations like
+// 0.5 nanos + 0.5 nanos = 1 nano. The following example shows the actual
+// Duration rep using quarters of a nanosecond.
+//
+// 2.5 sec = {rep_hi_=2, rep_lo_=2000000000} // lo = 4 * 500000000
+// -2.5 sec = {rep_hi_=-3, rep_lo_=2000000000}
+//
+// Infinite durations are represented as Durations with the rep_lo_ field set
+// to all 1s.
+//
+// +InfiniteDuration:
+// rep_hi_ : kint64max
+// rep_lo_ : ~0U
+//
+// -InfiniteDuration:
+// rep_hi_ : kint64min
+// rep_lo_ : ~0U
+//
+// Arithmetic overflows/underflows to +/- infinity and saturates.
+
+#if defined(_MSC_VER)
+#include <winsock2.h> // for timeval
+#endif
+
+#include <algorithm>
+#include <cassert>
+#include <cctype>
+#include <cerrno>
+#include <cmath>
+#include <cstdint>
+#include <cstdlib>
+#include <cstring>
+#include <ctime>
+#include <functional>
+#include <limits>
+#include <string>
+
+#include "absl/base/casts.h"
#include "absl/base/macros.h"
-#include "absl/numeric/int128.h"
+#include "absl/numeric/int128.h"
#include "absl/strings/string_view.h"
#include "absl/strings/strip.h"
-#include "absl/time/time.h"
-
-namespace absl {
+#include "absl/time/time.h"
+
+namespace absl {
ABSL_NAMESPACE_BEGIN
-
-namespace {
-
-using time_internal::kTicksPerNanosecond;
-using time_internal::kTicksPerSecond;
-
-constexpr int64_t kint64max = std::numeric_limits<int64_t>::max();
-constexpr int64_t kint64min = std::numeric_limits<int64_t>::min();
-
-// Can't use std::isinfinite() because it doesn't exist on windows.
-inline bool IsFinite(double d) {
- if (std::isnan(d)) return false;
- return d != std::numeric_limits<double>::infinity() &&
- d != -std::numeric_limits<double>::infinity();
-}
-
-inline bool IsValidDivisor(double d) {
- if (std::isnan(d)) return false;
- return d != 0.0;
-}
-
-// Can't use std::round() because it is only available in C++11.
-// Note that we ignore the possibility of floating-point over/underflow.
-template <typename Double>
-inline double Round(Double d) {
- return d < 0 ? std::ceil(d - 0.5) : std::floor(d + 0.5);
-}
-
-// *sec may be positive or negative. *ticks must be in the range
-// -kTicksPerSecond < *ticks < kTicksPerSecond. If *ticks is negative it
-// will be normalized to a positive value by adjusting *sec accordingly.
-inline void NormalizeTicks(int64_t* sec, int64_t* ticks) {
- if (*ticks < 0) {
- --*sec;
- *ticks += kTicksPerSecond;
- }
-}
-
-// Makes a uint128 from the absolute value of the given scalar.
-inline uint128 MakeU128(int64_t a) {
- uint128 u128 = 0;
- if (a < 0) {
- ++u128;
- ++a; // Makes it safe to negate 'a'
- a = -a;
- }
- u128 += static_cast<uint64_t>(a);
- return u128;
-}
-
-// Makes a uint128 count of ticks out of the absolute value of the Duration.
-inline uint128 MakeU128Ticks(Duration d) {
- int64_t rep_hi = time_internal::GetRepHi(d);
- uint32_t rep_lo = time_internal::GetRepLo(d);
- if (rep_hi < 0) {
- ++rep_hi;
- rep_hi = -rep_hi;
- rep_lo = kTicksPerSecond - rep_lo;
- }
- uint128 u128 = static_cast<uint64_t>(rep_hi);
- u128 *= static_cast<uint64_t>(kTicksPerSecond);
- u128 += rep_lo;
- return u128;
-}
-
-// Breaks a uint128 of ticks into a Duration.
-inline Duration MakeDurationFromU128(uint128 u128, bool is_neg) {
- int64_t rep_hi;
- uint32_t rep_lo;
- const uint64_t h64 = Uint128High64(u128);
- const uint64_t l64 = Uint128Low64(u128);
- if (h64 == 0) { // fastpath
- const uint64_t hi = l64 / kTicksPerSecond;
- rep_hi = static_cast<int64_t>(hi);
- rep_lo = static_cast<uint32_t>(l64 - hi * kTicksPerSecond);
- } else {
- // kMaxRepHi64 is the high 64 bits of (2^63 * kTicksPerSecond).
- // Any positive tick count whose high 64 bits are >= kMaxRepHi64
- // is not representable as a Duration. A negative tick count can
- // have its high 64 bits == kMaxRepHi64 but only when the low 64
- // bits are all zero, otherwise it is not representable either.
- const uint64_t kMaxRepHi64 = 0x77359400UL;
- if (h64 >= kMaxRepHi64) {
- if (is_neg && h64 == kMaxRepHi64 && l64 == 0) {
- // Avoid trying to represent -kint64min below.
- return time_internal::MakeDuration(kint64min);
- }
- return is_neg ? -InfiniteDuration() : InfiniteDuration();
- }
- const uint128 kTicksPerSecond128 = static_cast<uint64_t>(kTicksPerSecond);
- const uint128 hi = u128 / kTicksPerSecond128;
- rep_hi = static_cast<int64_t>(Uint128Low64(hi));
- rep_lo =
- static_cast<uint32_t>(Uint128Low64(u128 - hi * kTicksPerSecond128));
- }
- if (is_neg) {
- rep_hi = -rep_hi;
- if (rep_lo != 0) {
- --rep_hi;
- rep_lo = kTicksPerSecond - rep_lo;
- }
- }
- return time_internal::MakeDuration(rep_hi, rep_lo);
-}
-
-// Convert between int64_t and uint64_t, preserving representation. This
-// allows us to do arithmetic in the unsigned domain, where overflow has
-// well-defined behavior. See operator+=() and operator-=().
-//
-// C99 7.20.1.1.1, as referenced by C++11 18.4.1.2, says, "The typedef
-// name intN_t designates a signed integer type with width N, no padding
-// bits, and a two's complement representation." So, we can convert to
-// and from the corresponding uint64_t value using a bit cast.
-inline uint64_t EncodeTwosComp(int64_t v) {
- return absl::bit_cast<uint64_t>(v);
-}
-inline int64_t DecodeTwosComp(uint64_t v) { return absl::bit_cast<int64_t>(v); }
-
-// Note: The overflow detection in this function is done using greater/less *or
-// equal* because kint64max/min is too large to be represented exactly in a
-// double (which only has 53 bits of precision). In order to avoid assigning to
-// rep->hi a double value that is too large for an int64_t (and therefore is
-// undefined), we must consider computations that equal kint64max/min as a
-// double as overflow cases.
-inline bool SafeAddRepHi(double a_hi, double b_hi, Duration* d) {
- double c = a_hi + b_hi;
- if (c >= static_cast<double>(kint64max)) {
- *d = InfiniteDuration();
- return false;
- }
- if (c <= static_cast<double>(kint64min)) {
- *d = -InfiniteDuration();
- return false;
- }
- *d = time_internal::MakeDuration(c, time_internal::GetRepLo(*d));
- return true;
-}
-
-// A functor that's similar to std::multiplies<T>, except this returns the max
-// T value instead of overflowing. This is only defined for uint128.
-template <typename Ignored>
-struct SafeMultiply {
- uint128 operator()(uint128 a, uint128 b) const {
- // b hi is always zero because it originated as an int64_t.
- assert(Uint128High64(b) == 0);
- // Fastpath to avoid the expensive overflow check with division.
- if (Uint128High64(a) == 0) {
- return (((Uint128Low64(a) | Uint128Low64(b)) >> 32) == 0)
- ? static_cast<uint128>(Uint128Low64(a) * Uint128Low64(b))
- : a * b;
- }
- return b == 0 ? b : (a > kuint128max / b) ? kuint128max : a * b;
- }
-};
-
-// Scales (i.e., multiplies or divides, depending on the Operation template)
-// the Duration d by the int64_t r.
-template <template <typename> class Operation>
-inline Duration ScaleFixed(Duration d, int64_t r) {
- const uint128 a = MakeU128Ticks(d);
- const uint128 b = MakeU128(r);
- const uint128 q = Operation<uint128>()(a, b);
- const bool is_neg = (time_internal::GetRepHi(d) < 0) != (r < 0);
- return MakeDurationFromU128(q, is_neg);
-}
-
-// Scales (i.e., multiplies or divides, depending on the Operation template)
-// the Duration d by the double r.
-template <template <typename> class Operation>
-inline Duration ScaleDouble(Duration d, double r) {
- Operation<double> op;
- double hi_doub = op(time_internal::GetRepHi(d), r);
- double lo_doub = op(time_internal::GetRepLo(d), r);
-
- double hi_int = 0;
- double hi_frac = std::modf(hi_doub, &hi_int);
-
- // Moves hi's fractional bits to lo.
- lo_doub /= kTicksPerSecond;
- lo_doub += hi_frac;
-
- double lo_int = 0;
- double lo_frac = std::modf(lo_doub, &lo_int);
-
- // Rolls lo into hi if necessary.
- int64_t lo64 = Round(lo_frac * kTicksPerSecond);
-
- Duration ans;
- if (!SafeAddRepHi(hi_int, lo_int, &ans)) return ans;
- int64_t hi64 = time_internal::GetRepHi(ans);
- if (!SafeAddRepHi(hi64, lo64 / kTicksPerSecond, &ans)) return ans;
- hi64 = time_internal::GetRepHi(ans);
- lo64 %= kTicksPerSecond;
- NormalizeTicks(&hi64, &lo64);
- return time_internal::MakeDuration(hi64, lo64);
-}
-
-// Tries to divide num by den as fast as possible by looking for common, easy
-// cases. If the division was done, the quotient is in *q and the remainder is
-// in *rem and true will be returned.
-inline bool IDivFastPath(const Duration num, const Duration den, int64_t* q,
- Duration* rem) {
- // Bail if num or den is an infinity.
- if (time_internal::IsInfiniteDuration(num) ||
- time_internal::IsInfiniteDuration(den))
- return false;
-
- int64_t num_hi = time_internal::GetRepHi(num);
- uint32_t num_lo = time_internal::GetRepLo(num);
- int64_t den_hi = time_internal::GetRepHi(den);
- uint32_t den_lo = time_internal::GetRepLo(den);
-
- if (den_hi == 0 && den_lo == kTicksPerNanosecond) {
- // Dividing by 1ns
- if (num_hi >= 0 && num_hi < (kint64max - kTicksPerSecond) / 1000000000) {
- *q = num_hi * 1000000000 + num_lo / kTicksPerNanosecond;
- *rem = time_internal::MakeDuration(0, num_lo % den_lo);
- return true;
- }
- } else if (den_hi == 0 && den_lo == 100 * kTicksPerNanosecond) {
- // Dividing by 100ns (common when converting to Universal time)
- if (num_hi >= 0 && num_hi < (kint64max - kTicksPerSecond) / 10000000) {
- *q = num_hi * 10000000 + num_lo / (100 * kTicksPerNanosecond);
- *rem = time_internal::MakeDuration(0, num_lo % den_lo);
- return true;
- }
- } else if (den_hi == 0 && den_lo == 1000 * kTicksPerNanosecond) {
- // Dividing by 1us
- if (num_hi >= 0 && num_hi < (kint64max - kTicksPerSecond) / 1000000) {
- *q = num_hi * 1000000 + num_lo / (1000 * kTicksPerNanosecond);
- *rem = time_internal::MakeDuration(0, num_lo % den_lo);
- return true;
- }
- } else if (den_hi == 0 && den_lo == 1000000 * kTicksPerNanosecond) {
- // Dividing by 1ms
- if (num_hi >= 0 && num_hi < (kint64max - kTicksPerSecond) / 1000) {
- *q = num_hi * 1000 + num_lo / (1000000 * kTicksPerNanosecond);
- *rem = time_internal::MakeDuration(0, num_lo % den_lo);
- return true;
- }
- } else if (den_hi > 0 && den_lo == 0) {
- // Dividing by positive multiple of 1s
- if (num_hi >= 0) {
- if (den_hi == 1) {
- *q = num_hi;
- *rem = time_internal::MakeDuration(0, num_lo);
- return true;
- }
- *q = num_hi / den_hi;
- *rem = time_internal::MakeDuration(num_hi % den_hi, num_lo);
- return true;
- }
- if (num_lo != 0) {
- num_hi += 1;
- }
- int64_t quotient = num_hi / den_hi;
- int64_t rem_sec = num_hi % den_hi;
- if (rem_sec > 0) {
- rem_sec -= den_hi;
- quotient += 1;
- }
- if (num_lo != 0) {
- rem_sec -= 1;
- }
- *q = quotient;
- *rem = time_internal::MakeDuration(rem_sec, num_lo);
- return true;
- }
-
- return false;
-}
-
-} // namespace
-
-namespace time_internal {
-
-// The 'satq' argument indicates whether the quotient should saturate at the
-// bounds of int64_t. If it does saturate, the difference will spill over to
-// the remainder. If it does not saturate, the remainder remain accurate,
-// but the returned quotient will over/underflow int64_t and should not be used.
-int64_t IDivDuration(bool satq, const Duration num, const Duration den,
+
+namespace {
+
+using time_internal::kTicksPerNanosecond;
+using time_internal::kTicksPerSecond;
+
+constexpr int64_t kint64max = std::numeric_limits<int64_t>::max();
+constexpr int64_t kint64min = std::numeric_limits<int64_t>::min();
+
+// Can't use std::isinfinite() because it doesn't exist on windows.
+inline bool IsFinite(double d) {
+ if (std::isnan(d)) return false;
+ return d != std::numeric_limits<double>::infinity() &&
+ d != -std::numeric_limits<double>::infinity();
+}
+
+inline bool IsValidDivisor(double d) {
+ if (std::isnan(d)) return false;
+ return d != 0.0;
+}
+
+// Can't use std::round() because it is only available in C++11.
+// Note that we ignore the possibility of floating-point over/underflow.
+template <typename Double>
+inline double Round(Double d) {
+ return d < 0 ? std::ceil(d - 0.5) : std::floor(d + 0.5);
+}
+
+// *sec may be positive or negative. *ticks must be in the range
+// -kTicksPerSecond < *ticks < kTicksPerSecond. If *ticks is negative it
+// will be normalized to a positive value by adjusting *sec accordingly.
+inline void NormalizeTicks(int64_t* sec, int64_t* ticks) {
+ if (*ticks < 0) {
+ --*sec;
+ *ticks += kTicksPerSecond;
+ }
+}
+
+// Makes a uint128 from the absolute value of the given scalar.
+inline uint128 MakeU128(int64_t a) {
+ uint128 u128 = 0;
+ if (a < 0) {
+ ++u128;
+ ++a; // Makes it safe to negate 'a'
+ a = -a;
+ }
+ u128 += static_cast<uint64_t>(a);
+ return u128;
+}
+
+// Makes a uint128 count of ticks out of the absolute value of the Duration.
+inline uint128 MakeU128Ticks(Duration d) {
+ int64_t rep_hi = time_internal::GetRepHi(d);
+ uint32_t rep_lo = time_internal::GetRepLo(d);
+ if (rep_hi < 0) {
+ ++rep_hi;
+ rep_hi = -rep_hi;
+ rep_lo = kTicksPerSecond - rep_lo;
+ }
+ uint128 u128 = static_cast<uint64_t>(rep_hi);
+ u128 *= static_cast<uint64_t>(kTicksPerSecond);
+ u128 += rep_lo;
+ return u128;
+}
+
+// Breaks a uint128 of ticks into a Duration.
+inline Duration MakeDurationFromU128(uint128 u128, bool is_neg) {
+ int64_t rep_hi;
+ uint32_t rep_lo;
+ const uint64_t h64 = Uint128High64(u128);
+ const uint64_t l64 = Uint128Low64(u128);
+ if (h64 == 0) { // fastpath
+ const uint64_t hi = l64 / kTicksPerSecond;
+ rep_hi = static_cast<int64_t>(hi);
+ rep_lo = static_cast<uint32_t>(l64 - hi * kTicksPerSecond);
+ } else {
+ // kMaxRepHi64 is the high 64 bits of (2^63 * kTicksPerSecond).
+ // Any positive tick count whose high 64 bits are >= kMaxRepHi64
+ // is not representable as a Duration. A negative tick count can
+ // have its high 64 bits == kMaxRepHi64 but only when the low 64
+ // bits are all zero, otherwise it is not representable either.
+ const uint64_t kMaxRepHi64 = 0x77359400UL;
+ if (h64 >= kMaxRepHi64) {
+ if (is_neg && h64 == kMaxRepHi64 && l64 == 0) {
+ // Avoid trying to represent -kint64min below.
+ return time_internal::MakeDuration(kint64min);
+ }
+ return is_neg ? -InfiniteDuration() : InfiniteDuration();
+ }
+ const uint128 kTicksPerSecond128 = static_cast<uint64_t>(kTicksPerSecond);
+ const uint128 hi = u128 / kTicksPerSecond128;
+ rep_hi = static_cast<int64_t>(Uint128Low64(hi));
+ rep_lo =
+ static_cast<uint32_t>(Uint128Low64(u128 - hi * kTicksPerSecond128));
+ }
+ if (is_neg) {
+ rep_hi = -rep_hi;
+ if (rep_lo != 0) {
+ --rep_hi;
+ rep_lo = kTicksPerSecond - rep_lo;
+ }
+ }
+ return time_internal::MakeDuration(rep_hi, rep_lo);
+}
+
+// Convert between int64_t and uint64_t, preserving representation. This
+// allows us to do arithmetic in the unsigned domain, where overflow has
+// well-defined behavior. See operator+=() and operator-=().
+//
+// C99 7.20.1.1.1, as referenced by C++11 18.4.1.2, says, "The typedef
+// name intN_t designates a signed integer type with width N, no padding
+// bits, and a two's complement representation." So, we can convert to
+// and from the corresponding uint64_t value using a bit cast.
+inline uint64_t EncodeTwosComp(int64_t v) {
+ return absl::bit_cast<uint64_t>(v);
+}
+inline int64_t DecodeTwosComp(uint64_t v) { return absl::bit_cast<int64_t>(v); }
+
+// Note: The overflow detection in this function is done using greater/less *or
+// equal* because kint64max/min is too large to be represented exactly in a
+// double (which only has 53 bits of precision). In order to avoid assigning to
+// rep->hi a double value that is too large for an int64_t (and therefore is
+// undefined), we must consider computations that equal kint64max/min as a
+// double as overflow cases.
+inline bool SafeAddRepHi(double a_hi, double b_hi, Duration* d) {
+ double c = a_hi + b_hi;
+ if (c >= static_cast<double>(kint64max)) {
+ *d = InfiniteDuration();
+ return false;
+ }
+ if (c <= static_cast<double>(kint64min)) {
+ *d = -InfiniteDuration();
+ return false;
+ }
+ *d = time_internal::MakeDuration(c, time_internal::GetRepLo(*d));
+ return true;
+}
+
+// A functor that's similar to std::multiplies<T>, except this returns the max
+// T value instead of overflowing. This is only defined for uint128.
+template <typename Ignored>
+struct SafeMultiply {
+ uint128 operator()(uint128 a, uint128 b) const {
+ // b hi is always zero because it originated as an int64_t.
+ assert(Uint128High64(b) == 0);
+ // Fastpath to avoid the expensive overflow check with division.
+ if (Uint128High64(a) == 0) {
+ return (((Uint128Low64(a) | Uint128Low64(b)) >> 32) == 0)
+ ? static_cast<uint128>(Uint128Low64(a) * Uint128Low64(b))
+ : a * b;
+ }
+ return b == 0 ? b : (a > kuint128max / b) ? kuint128max : a * b;
+ }
+};
+
+// Scales (i.e., multiplies or divides, depending on the Operation template)
+// the Duration d by the int64_t r.
+template <template <typename> class Operation>
+inline Duration ScaleFixed(Duration d, int64_t r) {
+ const uint128 a = MakeU128Ticks(d);
+ const uint128 b = MakeU128(r);
+ const uint128 q = Operation<uint128>()(a, b);
+ const bool is_neg = (time_internal::GetRepHi(d) < 0) != (r < 0);
+ return MakeDurationFromU128(q, is_neg);
+}
+
+// Scales (i.e., multiplies or divides, depending on the Operation template)
+// the Duration d by the double r.
+template <template <typename> class Operation>
+inline Duration ScaleDouble(Duration d, double r) {
+ Operation<double> op;
+ double hi_doub = op(time_internal::GetRepHi(d), r);
+ double lo_doub = op(time_internal::GetRepLo(d), r);
+
+ double hi_int = 0;
+ double hi_frac = std::modf(hi_doub, &hi_int);
+
+ // Moves hi's fractional bits to lo.
+ lo_doub /= kTicksPerSecond;
+ lo_doub += hi_frac;
+
+ double lo_int = 0;
+ double lo_frac = std::modf(lo_doub, &lo_int);
+
+ // Rolls lo into hi if necessary.
+ int64_t lo64 = Round(lo_frac * kTicksPerSecond);
+
+ Duration ans;
+ if (!SafeAddRepHi(hi_int, lo_int, &ans)) return ans;
+ int64_t hi64 = time_internal::GetRepHi(ans);
+ if (!SafeAddRepHi(hi64, lo64 / kTicksPerSecond, &ans)) return ans;
+ hi64 = time_internal::GetRepHi(ans);
+ lo64 %= kTicksPerSecond;
+ NormalizeTicks(&hi64, &lo64);
+ return time_internal::MakeDuration(hi64, lo64);
+}
+
+// Tries to divide num by den as fast as possible by looking for common, easy
+// cases. If the division was done, the quotient is in *q and the remainder is
+// in *rem and true will be returned.
+inline bool IDivFastPath(const Duration num, const Duration den, int64_t* q,
+ Duration* rem) {
+ // Bail if num or den is an infinity.
+ if (time_internal::IsInfiniteDuration(num) ||
+ time_internal::IsInfiniteDuration(den))
+ return false;
+
+ int64_t num_hi = time_internal::GetRepHi(num);
+ uint32_t num_lo = time_internal::GetRepLo(num);
+ int64_t den_hi = time_internal::GetRepHi(den);
+ uint32_t den_lo = time_internal::GetRepLo(den);
+
+ if (den_hi == 0 && den_lo == kTicksPerNanosecond) {
+ // Dividing by 1ns
+ if (num_hi >= 0 && num_hi < (kint64max - kTicksPerSecond) / 1000000000) {
+ *q = num_hi * 1000000000 + num_lo / kTicksPerNanosecond;
+ *rem = time_internal::MakeDuration(0, num_lo % den_lo);
+ return true;
+ }
+ } else if (den_hi == 0 && den_lo == 100 * kTicksPerNanosecond) {
+ // Dividing by 100ns (common when converting to Universal time)
+ if (num_hi >= 0 && num_hi < (kint64max - kTicksPerSecond) / 10000000) {
+ *q = num_hi * 10000000 + num_lo / (100 * kTicksPerNanosecond);
+ *rem = time_internal::MakeDuration(0, num_lo % den_lo);
+ return true;
+ }
+ } else if (den_hi == 0 && den_lo == 1000 * kTicksPerNanosecond) {
+ // Dividing by 1us
+ if (num_hi >= 0 && num_hi < (kint64max - kTicksPerSecond) / 1000000) {
+ *q = num_hi * 1000000 + num_lo / (1000 * kTicksPerNanosecond);
+ *rem = time_internal::MakeDuration(0, num_lo % den_lo);
+ return true;
+ }
+ } else if (den_hi == 0 && den_lo == 1000000 * kTicksPerNanosecond) {
+ // Dividing by 1ms
+ if (num_hi >= 0 && num_hi < (kint64max - kTicksPerSecond) / 1000) {
+ *q = num_hi * 1000 + num_lo / (1000000 * kTicksPerNanosecond);
+ *rem = time_internal::MakeDuration(0, num_lo % den_lo);
+ return true;
+ }
+ } else if (den_hi > 0 && den_lo == 0) {
+ // Dividing by positive multiple of 1s
+ if (num_hi >= 0) {
+ if (den_hi == 1) {
+ *q = num_hi;
+ *rem = time_internal::MakeDuration(0, num_lo);
+ return true;
+ }
+ *q = num_hi / den_hi;
+ *rem = time_internal::MakeDuration(num_hi % den_hi, num_lo);
+ return true;
+ }
+ if (num_lo != 0) {
+ num_hi += 1;
+ }
+ int64_t quotient = num_hi / den_hi;
+ int64_t rem_sec = num_hi % den_hi;
+ if (rem_sec > 0) {
+ rem_sec -= den_hi;
+ quotient += 1;
+ }
+ if (num_lo != 0) {
+ rem_sec -= 1;
+ }
+ *q = quotient;
+ *rem = time_internal::MakeDuration(rem_sec, num_lo);
+ return true;
+ }
+
+ return false;
+}
+
+} // namespace
+
+namespace time_internal {
+
+// The 'satq' argument indicates whether the quotient should saturate at the
+// bounds of int64_t. If it does saturate, the difference will spill over to
+// the remainder. If it does not saturate, the remainder remain accurate,
+// but the returned quotient will over/underflow int64_t and should not be used.
+int64_t IDivDuration(bool satq, const Duration num, const Duration den,
Duration* rem) {
- int64_t q = 0;
- if (IDivFastPath(num, den, &q, rem)) {
- return q;
- }
-
- const bool num_neg = num < ZeroDuration();
- const bool den_neg = den < ZeroDuration();
- const bool quotient_neg = num_neg != den_neg;
-
- if (time_internal::IsInfiniteDuration(num) || den == ZeroDuration()) {
- *rem = num_neg ? -InfiniteDuration() : InfiniteDuration();
- return quotient_neg ? kint64min : kint64max;
- }
- if (time_internal::IsInfiniteDuration(den)) {
- *rem = num;
- return 0;
- }
-
- const uint128 a = MakeU128Ticks(num);
- const uint128 b = MakeU128Ticks(den);
- uint128 quotient128 = a / b;
-
- if (satq) {
- // Limits the quotient to the range of int64_t.
- if (quotient128 > uint128(static_cast<uint64_t>(kint64max))) {
- quotient128 = quotient_neg ? uint128(static_cast<uint64_t>(kint64min))
- : uint128(static_cast<uint64_t>(kint64max));
- }
- }
-
- const uint128 remainder128 = a - quotient128 * b;
- *rem = MakeDurationFromU128(remainder128, num_neg);
-
- if (!quotient_neg || quotient128 == 0) {
- return Uint128Low64(quotient128) & kint64max;
- }
- // The quotient needs to be negated, but we need to carefully handle
- // quotient128s with the top bit on.
- return -static_cast<int64_t>(Uint128Low64(quotient128 - 1) & kint64max) - 1;
-}
-
-} // namespace time_internal
-
-//
-// Additive operators.
-//
-
-Duration& Duration::operator+=(Duration rhs) {
- if (time_internal::IsInfiniteDuration(*this)) return *this;
- if (time_internal::IsInfiniteDuration(rhs)) return *this = rhs;
- const int64_t orig_rep_hi = rep_hi_;
- rep_hi_ =
- DecodeTwosComp(EncodeTwosComp(rep_hi_) + EncodeTwosComp(rhs.rep_hi_));
- if (rep_lo_ >= kTicksPerSecond - rhs.rep_lo_) {
- rep_hi_ = DecodeTwosComp(EncodeTwosComp(rep_hi_) + 1);
- rep_lo_ -= kTicksPerSecond;
- }
- rep_lo_ += rhs.rep_lo_;
- if (rhs.rep_hi_ < 0 ? rep_hi_ > orig_rep_hi : rep_hi_ < orig_rep_hi) {
- return *this = rhs.rep_hi_ < 0 ? -InfiniteDuration() : InfiniteDuration();
- }
- return *this;
-}
-
-Duration& Duration::operator-=(Duration rhs) {
- if (time_internal::IsInfiniteDuration(*this)) return *this;
- if (time_internal::IsInfiniteDuration(rhs)) {
- return *this = rhs.rep_hi_ >= 0 ? -InfiniteDuration() : InfiniteDuration();
- }
- const int64_t orig_rep_hi = rep_hi_;
- rep_hi_ =
- DecodeTwosComp(EncodeTwosComp(rep_hi_) - EncodeTwosComp(rhs.rep_hi_));
- if (rep_lo_ < rhs.rep_lo_) {
- rep_hi_ = DecodeTwosComp(EncodeTwosComp(rep_hi_) - 1);
- rep_lo_ += kTicksPerSecond;
- }
- rep_lo_ -= rhs.rep_lo_;
- if (rhs.rep_hi_ < 0 ? rep_hi_ < orig_rep_hi : rep_hi_ > orig_rep_hi) {
- return *this = rhs.rep_hi_ >= 0 ? -InfiniteDuration() : InfiniteDuration();
- }
- return *this;
-}
-
-//
-// Multiplicative operators.
-//
-
-Duration& Duration::operator*=(int64_t r) {
- if (time_internal::IsInfiniteDuration(*this)) {
- const bool is_neg = (r < 0) != (rep_hi_ < 0);
- return *this = is_neg ? -InfiniteDuration() : InfiniteDuration();
- }
- return *this = ScaleFixed<SafeMultiply>(*this, r);
-}
-
-Duration& Duration::operator*=(double r) {
- if (time_internal::IsInfiniteDuration(*this) || !IsFinite(r)) {
- const bool is_neg = (std::signbit(r) != 0) != (rep_hi_ < 0);
- return *this = is_neg ? -InfiniteDuration() : InfiniteDuration();
- }
- return *this = ScaleDouble<std::multiplies>(*this, r);
-}
-
-Duration& Duration::operator/=(int64_t r) {
- if (time_internal::IsInfiniteDuration(*this) || r == 0) {
- const bool is_neg = (r < 0) != (rep_hi_ < 0);
- return *this = is_neg ? -InfiniteDuration() : InfiniteDuration();
- }
- return *this = ScaleFixed<std::divides>(*this, r);
-}
-
-Duration& Duration::operator/=(double r) {
- if (time_internal::IsInfiniteDuration(*this) || !IsValidDivisor(r)) {
- const bool is_neg = (std::signbit(r) != 0) != (rep_hi_ < 0);
- return *this = is_neg ? -InfiniteDuration() : InfiniteDuration();
- }
- return *this = ScaleDouble<std::divides>(*this, r);
-}
-
-Duration& Duration::operator%=(Duration rhs) {
- time_internal::IDivDuration(false, *this, rhs, this);
- return *this;
-}
-
-double FDivDuration(Duration num, Duration den) {
- // Arithmetic with infinity is sticky.
- if (time_internal::IsInfiniteDuration(num) || den == ZeroDuration()) {
- return (num < ZeroDuration()) == (den < ZeroDuration())
- ? std::numeric_limits<double>::infinity()
- : -std::numeric_limits<double>::infinity();
- }
- if (time_internal::IsInfiniteDuration(den)) return 0.0;
-
- double a =
- static_cast<double>(time_internal::GetRepHi(num)) * kTicksPerSecond +
- time_internal::GetRepLo(num);
- double b =
- static_cast<double>(time_internal::GetRepHi(den)) * kTicksPerSecond +
- time_internal::GetRepLo(den);
- return a / b;
-}
-
-//
-// Trunc/Floor/Ceil.
-//
-
-Duration Trunc(Duration d, Duration unit) {
- return d - (d % unit);
-}
-
-Duration Floor(const Duration d, const Duration unit) {
- const absl::Duration td = Trunc(d, unit);
- return td <= d ? td : td - AbsDuration(unit);
-}
-
-Duration Ceil(const Duration d, const Duration unit) {
- const absl::Duration td = Trunc(d, unit);
- return td >= d ? td : td + AbsDuration(unit);
-}
-
-//
-// Factory functions.
-//
-
-Duration DurationFromTimespec(timespec ts) {
- if (static_cast<uint64_t>(ts.tv_nsec) < 1000 * 1000 * 1000) {
- int64_t ticks = ts.tv_nsec * kTicksPerNanosecond;
- return time_internal::MakeDuration(ts.tv_sec, ticks);
- }
- return Seconds(ts.tv_sec) + Nanoseconds(ts.tv_nsec);
-}
-
-Duration DurationFromTimeval(timeval tv) {
- if (static_cast<uint64_t>(tv.tv_usec) < 1000 * 1000) {
- int64_t ticks = tv.tv_usec * 1000 * kTicksPerNanosecond;
- return time_internal::MakeDuration(tv.tv_sec, ticks);
- }
- return Seconds(tv.tv_sec) + Microseconds(tv.tv_usec);
-}
-
-//
-// Conversion to other duration types.
-//
-
-int64_t ToInt64Nanoseconds(Duration d) {
- if (time_internal::GetRepHi(d) >= 0 &&
- time_internal::GetRepHi(d) >> 33 == 0) {
- return (time_internal::GetRepHi(d) * 1000 * 1000 * 1000) +
- (time_internal::GetRepLo(d) / kTicksPerNanosecond);
- }
- return d / Nanoseconds(1);
-}
-int64_t ToInt64Microseconds(Duration d) {
- if (time_internal::GetRepHi(d) >= 0 &&
- time_internal::GetRepHi(d) >> 43 == 0) {
- return (time_internal::GetRepHi(d) * 1000 * 1000) +
- (time_internal::GetRepLo(d) / (kTicksPerNanosecond * 1000));
- }
- return d / Microseconds(1);
-}
-int64_t ToInt64Milliseconds(Duration d) {
- if (time_internal::GetRepHi(d) >= 0 &&
- time_internal::GetRepHi(d) >> 53 == 0) {
- return (time_internal::GetRepHi(d) * 1000) +
- (time_internal::GetRepLo(d) / (kTicksPerNanosecond * 1000 * 1000));
- }
- return d / Milliseconds(1);
-}
-int64_t ToInt64Seconds(Duration d) {
- int64_t hi = time_internal::GetRepHi(d);
- if (time_internal::IsInfiniteDuration(d)) return hi;
- if (hi < 0 && time_internal::GetRepLo(d) != 0) ++hi;
- return hi;
-}
-int64_t ToInt64Minutes(Duration d) {
- int64_t hi = time_internal::GetRepHi(d);
- if (time_internal::IsInfiniteDuration(d)) return hi;
- if (hi < 0 && time_internal::GetRepLo(d) != 0) ++hi;
- return hi / 60;
-}
-int64_t ToInt64Hours(Duration d) {
- int64_t hi = time_internal::GetRepHi(d);
- if (time_internal::IsInfiniteDuration(d)) return hi;
- if (hi < 0 && time_internal::GetRepLo(d) != 0) ++hi;
- return hi / (60 * 60);
-}
-
-double ToDoubleNanoseconds(Duration d) {
- return FDivDuration(d, Nanoseconds(1));
-}
-double ToDoubleMicroseconds(Duration d) {
- return FDivDuration(d, Microseconds(1));
-}
-double ToDoubleMilliseconds(Duration d) {
- return FDivDuration(d, Milliseconds(1));
-}
-double ToDoubleSeconds(Duration d) {
- return FDivDuration(d, Seconds(1));
-}
-double ToDoubleMinutes(Duration d) {
- return FDivDuration(d, Minutes(1));
-}
-double ToDoubleHours(Duration d) {
- return FDivDuration(d, Hours(1));
-}
-
-timespec ToTimespec(Duration d) {
- timespec ts;
- if (!time_internal::IsInfiniteDuration(d)) {
- int64_t rep_hi = time_internal::GetRepHi(d);
- uint32_t rep_lo = time_internal::GetRepLo(d);
- if (rep_hi < 0) {
- // Tweak the fields so that unsigned division of rep_lo
- // maps to truncation (towards zero) for the timespec.
- rep_lo += kTicksPerNanosecond - 1;
- if (rep_lo >= kTicksPerSecond) {
- rep_hi += 1;
- rep_lo -= kTicksPerSecond;
- }
- }
- ts.tv_sec = rep_hi;
- if (ts.tv_sec == rep_hi) { // no time_t narrowing
- ts.tv_nsec = rep_lo / kTicksPerNanosecond;
- return ts;
- }
- }
- if (d >= ZeroDuration()) {
- ts.tv_sec = std::numeric_limits<time_t>::max();
- ts.tv_nsec = 1000 * 1000 * 1000 - 1;
- } else {
- ts.tv_sec = std::numeric_limits<time_t>::min();
- ts.tv_nsec = 0;
- }
- return ts;
-}
-
-timeval ToTimeval(Duration d) {
- timeval tv;
- timespec ts = ToTimespec(d);
- if (ts.tv_sec < 0) {
- // Tweak the fields so that positive division of tv_nsec
- // maps to truncation (towards zero) for the timeval.
- ts.tv_nsec += 1000 - 1;
- if (ts.tv_nsec >= 1000 * 1000 * 1000) {
- ts.tv_sec += 1;
- ts.tv_nsec -= 1000 * 1000 * 1000;
- }
- }
- tv.tv_sec = ts.tv_sec;
- if (tv.tv_sec != ts.tv_sec) { // narrowing
- if (ts.tv_sec < 0) {
- tv.tv_sec = std::numeric_limits<decltype(tv.tv_sec)>::min();
- tv.tv_usec = 0;
- } else {
- tv.tv_sec = std::numeric_limits<decltype(tv.tv_sec)>::max();
- tv.tv_usec = 1000 * 1000 - 1;
- }
- return tv;
- }
- tv.tv_usec = static_cast<int>(ts.tv_nsec / 1000); // suseconds_t
- return tv;
-}
-
-std::chrono::nanoseconds ToChronoNanoseconds(Duration d) {
- return time_internal::ToChronoDuration<std::chrono::nanoseconds>(d);
-}
-std::chrono::microseconds ToChronoMicroseconds(Duration d) {
- return time_internal::ToChronoDuration<std::chrono::microseconds>(d);
-}
-std::chrono::milliseconds ToChronoMilliseconds(Duration d) {
- return time_internal::ToChronoDuration<std::chrono::milliseconds>(d);
-}
-std::chrono::seconds ToChronoSeconds(Duration d) {
- return time_internal::ToChronoDuration<std::chrono::seconds>(d);
-}
-std::chrono::minutes ToChronoMinutes(Duration d) {
- return time_internal::ToChronoDuration<std::chrono::minutes>(d);
-}
-std::chrono::hours ToChronoHours(Duration d) {
- return time_internal::ToChronoDuration<std::chrono::hours>(d);
-}
-
-//
-// To/From string formatting.
-//
-
-namespace {
-
-// Formats a positive 64-bit integer in the given field width. Note that
-// it is up to the caller of Format64() to ensure that there is sufficient
-// space before ep to hold the conversion.
-char* Format64(char* ep, int width, int64_t v) {
- do {
- --width;
- *--ep = '0' + (v % 10); // contiguous digits
- } while (v /= 10);
- while (--width >= 0) *--ep = '0'; // zero pad
- return ep;
-}
-
-// Helpers for FormatDuration() that format 'n' and append it to 'out'
-// followed by the given 'unit'. If 'n' formats to "0", nothing is
-// appended (not even the unit).
-
-// A type that encapsulates how to display a value of a particular unit. For
-// values that are displayed with fractional parts, the precision indicates
-// where to round the value. The precision varies with the display unit because
-// a Duration can hold only quarters of a nanosecond, so displaying information
-// beyond that is just noise.
-//
-// For example, a microsecond value of 42.00025xxxxx should not display beyond 5
-// fractional digits, because it is in the noise of what a Duration can
-// represent.
-struct DisplayUnit {
+ int64_t q = 0;
+ if (IDivFastPath(num, den, &q, rem)) {
+ return q;
+ }
+
+ const bool num_neg = num < ZeroDuration();
+ const bool den_neg = den < ZeroDuration();
+ const bool quotient_neg = num_neg != den_neg;
+
+ if (time_internal::IsInfiniteDuration(num) || den == ZeroDuration()) {
+ *rem = num_neg ? -InfiniteDuration() : InfiniteDuration();
+ return quotient_neg ? kint64min : kint64max;
+ }
+ if (time_internal::IsInfiniteDuration(den)) {
+ *rem = num;
+ return 0;
+ }
+
+ const uint128 a = MakeU128Ticks(num);
+ const uint128 b = MakeU128Ticks(den);
+ uint128 quotient128 = a / b;
+
+ if (satq) {
+ // Limits the quotient to the range of int64_t.
+ if (quotient128 > uint128(static_cast<uint64_t>(kint64max))) {
+ quotient128 = quotient_neg ? uint128(static_cast<uint64_t>(kint64min))
+ : uint128(static_cast<uint64_t>(kint64max));
+ }
+ }
+
+ const uint128 remainder128 = a - quotient128 * b;
+ *rem = MakeDurationFromU128(remainder128, num_neg);
+
+ if (!quotient_neg || quotient128 == 0) {
+ return Uint128Low64(quotient128) & kint64max;
+ }
+ // The quotient needs to be negated, but we need to carefully handle
+ // quotient128s with the top bit on.
+ return -static_cast<int64_t>(Uint128Low64(quotient128 - 1) & kint64max) - 1;
+}
+
+} // namespace time_internal
+
+//
+// Additive operators.
+//
+
+Duration& Duration::operator+=(Duration rhs) {
+ if (time_internal::IsInfiniteDuration(*this)) return *this;
+ if (time_internal::IsInfiniteDuration(rhs)) return *this = rhs;
+ const int64_t orig_rep_hi = rep_hi_;
+ rep_hi_ =
+ DecodeTwosComp(EncodeTwosComp(rep_hi_) + EncodeTwosComp(rhs.rep_hi_));
+ if (rep_lo_ >= kTicksPerSecond - rhs.rep_lo_) {
+ rep_hi_ = DecodeTwosComp(EncodeTwosComp(rep_hi_) + 1);
+ rep_lo_ -= kTicksPerSecond;
+ }
+ rep_lo_ += rhs.rep_lo_;
+ if (rhs.rep_hi_ < 0 ? rep_hi_ > orig_rep_hi : rep_hi_ < orig_rep_hi) {
+ return *this = rhs.rep_hi_ < 0 ? -InfiniteDuration() : InfiniteDuration();
+ }
+ return *this;
+}
+
+Duration& Duration::operator-=(Duration rhs) {
+ if (time_internal::IsInfiniteDuration(*this)) return *this;
+ if (time_internal::IsInfiniteDuration(rhs)) {
+ return *this = rhs.rep_hi_ >= 0 ? -InfiniteDuration() : InfiniteDuration();
+ }
+ const int64_t orig_rep_hi = rep_hi_;
+ rep_hi_ =
+ DecodeTwosComp(EncodeTwosComp(rep_hi_) - EncodeTwosComp(rhs.rep_hi_));
+ if (rep_lo_ < rhs.rep_lo_) {
+ rep_hi_ = DecodeTwosComp(EncodeTwosComp(rep_hi_) - 1);
+ rep_lo_ += kTicksPerSecond;
+ }
+ rep_lo_ -= rhs.rep_lo_;
+ if (rhs.rep_hi_ < 0 ? rep_hi_ < orig_rep_hi : rep_hi_ > orig_rep_hi) {
+ return *this = rhs.rep_hi_ >= 0 ? -InfiniteDuration() : InfiniteDuration();
+ }
+ return *this;
+}
+
+//
+// Multiplicative operators.
+//
+
+Duration& Duration::operator*=(int64_t r) {
+ if (time_internal::IsInfiniteDuration(*this)) {
+ const bool is_neg = (r < 0) != (rep_hi_ < 0);
+ return *this = is_neg ? -InfiniteDuration() : InfiniteDuration();
+ }
+ return *this = ScaleFixed<SafeMultiply>(*this, r);
+}
+
+Duration& Duration::operator*=(double r) {
+ if (time_internal::IsInfiniteDuration(*this) || !IsFinite(r)) {
+ const bool is_neg = (std::signbit(r) != 0) != (rep_hi_ < 0);
+ return *this = is_neg ? -InfiniteDuration() : InfiniteDuration();
+ }
+ return *this = ScaleDouble<std::multiplies>(*this, r);
+}
+
+Duration& Duration::operator/=(int64_t r) {
+ if (time_internal::IsInfiniteDuration(*this) || r == 0) {
+ const bool is_neg = (r < 0) != (rep_hi_ < 0);
+ return *this = is_neg ? -InfiniteDuration() : InfiniteDuration();
+ }
+ return *this = ScaleFixed<std::divides>(*this, r);
+}
+
+Duration& Duration::operator/=(double r) {
+ if (time_internal::IsInfiniteDuration(*this) || !IsValidDivisor(r)) {
+ const bool is_neg = (std::signbit(r) != 0) != (rep_hi_ < 0);
+ return *this = is_neg ? -InfiniteDuration() : InfiniteDuration();
+ }
+ return *this = ScaleDouble<std::divides>(*this, r);
+}
+
+Duration& Duration::operator%=(Duration rhs) {
+ time_internal::IDivDuration(false, *this, rhs, this);
+ return *this;
+}
+
+double FDivDuration(Duration num, Duration den) {
+ // Arithmetic with infinity is sticky.
+ if (time_internal::IsInfiniteDuration(num) || den == ZeroDuration()) {
+ return (num < ZeroDuration()) == (den < ZeroDuration())
+ ? std::numeric_limits<double>::infinity()
+ : -std::numeric_limits<double>::infinity();
+ }
+ if (time_internal::IsInfiniteDuration(den)) return 0.0;
+
+ double a =
+ static_cast<double>(time_internal::GetRepHi(num)) * kTicksPerSecond +
+ time_internal::GetRepLo(num);
+ double b =
+ static_cast<double>(time_internal::GetRepHi(den)) * kTicksPerSecond +
+ time_internal::GetRepLo(den);
+ return a / b;
+}
+
+//
+// Trunc/Floor/Ceil.
+//
+
+Duration Trunc(Duration d, Duration unit) {
+ return d - (d % unit);
+}
+
+Duration Floor(const Duration d, const Duration unit) {
+ const absl::Duration td = Trunc(d, unit);
+ return td <= d ? td : td - AbsDuration(unit);
+}
+
+Duration Ceil(const Duration d, const Duration unit) {
+ const absl::Duration td = Trunc(d, unit);
+ return td >= d ? td : td + AbsDuration(unit);
+}
+
+//
+// Factory functions.
+//
+
+Duration DurationFromTimespec(timespec ts) {
+ if (static_cast<uint64_t>(ts.tv_nsec) < 1000 * 1000 * 1000) {
+ int64_t ticks = ts.tv_nsec * kTicksPerNanosecond;
+ return time_internal::MakeDuration(ts.tv_sec, ticks);
+ }
+ return Seconds(ts.tv_sec) + Nanoseconds(ts.tv_nsec);
+}
+
+Duration DurationFromTimeval(timeval tv) {
+ if (static_cast<uint64_t>(tv.tv_usec) < 1000 * 1000) {
+ int64_t ticks = tv.tv_usec * 1000 * kTicksPerNanosecond;
+ return time_internal::MakeDuration(tv.tv_sec, ticks);
+ }
+ return Seconds(tv.tv_sec) + Microseconds(tv.tv_usec);
+}
+
+//
+// Conversion to other duration types.
+//
+
+int64_t ToInt64Nanoseconds(Duration d) {
+ if (time_internal::GetRepHi(d) >= 0 &&
+ time_internal::GetRepHi(d) >> 33 == 0) {
+ return (time_internal::GetRepHi(d) * 1000 * 1000 * 1000) +
+ (time_internal::GetRepLo(d) / kTicksPerNanosecond);
+ }
+ return d / Nanoseconds(1);
+}
+int64_t ToInt64Microseconds(Duration d) {
+ if (time_internal::GetRepHi(d) >= 0 &&
+ time_internal::GetRepHi(d) >> 43 == 0) {
+ return (time_internal::GetRepHi(d) * 1000 * 1000) +
+ (time_internal::GetRepLo(d) / (kTicksPerNanosecond * 1000));
+ }
+ return d / Microseconds(1);
+}
+int64_t ToInt64Milliseconds(Duration d) {
+ if (time_internal::GetRepHi(d) >= 0 &&
+ time_internal::GetRepHi(d) >> 53 == 0) {
+ return (time_internal::GetRepHi(d) * 1000) +
+ (time_internal::GetRepLo(d) / (kTicksPerNanosecond * 1000 * 1000));
+ }
+ return d / Milliseconds(1);
+}
+int64_t ToInt64Seconds(Duration d) {
+ int64_t hi = time_internal::GetRepHi(d);
+ if (time_internal::IsInfiniteDuration(d)) return hi;
+ if (hi < 0 && time_internal::GetRepLo(d) != 0) ++hi;
+ return hi;
+}
+int64_t ToInt64Minutes(Duration d) {
+ int64_t hi = time_internal::GetRepHi(d);
+ if (time_internal::IsInfiniteDuration(d)) return hi;
+ if (hi < 0 && time_internal::GetRepLo(d) != 0) ++hi;
+ return hi / 60;
+}
+int64_t ToInt64Hours(Duration d) {
+ int64_t hi = time_internal::GetRepHi(d);
+ if (time_internal::IsInfiniteDuration(d)) return hi;
+ if (hi < 0 && time_internal::GetRepLo(d) != 0) ++hi;
+ return hi / (60 * 60);
+}
+
+double ToDoubleNanoseconds(Duration d) {
+ return FDivDuration(d, Nanoseconds(1));
+}
+double ToDoubleMicroseconds(Duration d) {
+ return FDivDuration(d, Microseconds(1));
+}
+double ToDoubleMilliseconds(Duration d) {
+ return FDivDuration(d, Milliseconds(1));
+}
+double ToDoubleSeconds(Duration d) {
+ return FDivDuration(d, Seconds(1));
+}
+double ToDoubleMinutes(Duration d) {
+ return FDivDuration(d, Minutes(1));
+}
+double ToDoubleHours(Duration d) {
+ return FDivDuration(d, Hours(1));
+}
+
+timespec ToTimespec(Duration d) {
+ timespec ts;
+ if (!time_internal::IsInfiniteDuration(d)) {
+ int64_t rep_hi = time_internal::GetRepHi(d);
+ uint32_t rep_lo = time_internal::GetRepLo(d);
+ if (rep_hi < 0) {
+ // Tweak the fields so that unsigned division of rep_lo
+ // maps to truncation (towards zero) for the timespec.
+ rep_lo += kTicksPerNanosecond - 1;
+ if (rep_lo >= kTicksPerSecond) {
+ rep_hi += 1;
+ rep_lo -= kTicksPerSecond;
+ }
+ }
+ ts.tv_sec = rep_hi;
+ if (ts.tv_sec == rep_hi) { // no time_t narrowing
+ ts.tv_nsec = rep_lo / kTicksPerNanosecond;
+ return ts;
+ }
+ }
+ if (d >= ZeroDuration()) {
+ ts.tv_sec = std::numeric_limits<time_t>::max();
+ ts.tv_nsec = 1000 * 1000 * 1000 - 1;
+ } else {
+ ts.tv_sec = std::numeric_limits<time_t>::min();
+ ts.tv_nsec = 0;
+ }
+ return ts;
+}
+
+timeval ToTimeval(Duration d) {
+ timeval tv;
+ timespec ts = ToTimespec(d);
+ if (ts.tv_sec < 0) {
+ // Tweak the fields so that positive division of tv_nsec
+ // maps to truncation (towards zero) for the timeval.
+ ts.tv_nsec += 1000 - 1;
+ if (ts.tv_nsec >= 1000 * 1000 * 1000) {
+ ts.tv_sec += 1;
+ ts.tv_nsec -= 1000 * 1000 * 1000;
+ }
+ }
+ tv.tv_sec = ts.tv_sec;
+ if (tv.tv_sec != ts.tv_sec) { // narrowing
+ if (ts.tv_sec < 0) {
+ tv.tv_sec = std::numeric_limits<decltype(tv.tv_sec)>::min();
+ tv.tv_usec = 0;
+ } else {
+ tv.tv_sec = std::numeric_limits<decltype(tv.tv_sec)>::max();
+ tv.tv_usec = 1000 * 1000 - 1;
+ }
+ return tv;
+ }
+ tv.tv_usec = static_cast<int>(ts.tv_nsec / 1000); // suseconds_t
+ return tv;
+}
+
+std::chrono::nanoseconds ToChronoNanoseconds(Duration d) {
+ return time_internal::ToChronoDuration<std::chrono::nanoseconds>(d);
+}
+std::chrono::microseconds ToChronoMicroseconds(Duration d) {
+ return time_internal::ToChronoDuration<std::chrono::microseconds>(d);
+}
+std::chrono::milliseconds ToChronoMilliseconds(Duration d) {
+ return time_internal::ToChronoDuration<std::chrono::milliseconds>(d);
+}
+std::chrono::seconds ToChronoSeconds(Duration d) {
+ return time_internal::ToChronoDuration<std::chrono::seconds>(d);
+}
+std::chrono::minutes ToChronoMinutes(Duration d) {
+ return time_internal::ToChronoDuration<std::chrono::minutes>(d);
+}
+std::chrono::hours ToChronoHours(Duration d) {
+ return time_internal::ToChronoDuration<std::chrono::hours>(d);
+}
+
+//
+// To/From string formatting.
+//
+
+namespace {
+
+// Formats a positive 64-bit integer in the given field width. Note that
+// it is up to the caller of Format64() to ensure that there is sufficient
+// space before ep to hold the conversion.
+char* Format64(char* ep, int width, int64_t v) {
+ do {
+ --width;
+ *--ep = '0' + (v % 10); // contiguous digits
+ } while (v /= 10);
+ while (--width >= 0) *--ep = '0'; // zero pad
+ return ep;
+}
+
+// Helpers for FormatDuration() that format 'n' and append it to 'out'
+// followed by the given 'unit'. If 'n' formats to "0", nothing is
+// appended (not even the unit).
+
+// A type that encapsulates how to display a value of a particular unit. For
+// values that are displayed with fractional parts, the precision indicates
+// where to round the value. The precision varies with the display unit because
+// a Duration can hold only quarters of a nanosecond, so displaying information
+// beyond that is just noise.
+//
+// For example, a microsecond value of 42.00025xxxxx should not display beyond 5
+// fractional digits, because it is in the noise of what a Duration can
+// represent.
+struct DisplayUnit {
absl::string_view abbr;
- int prec;
- double pow10;
-};
+ int prec;
+ double pow10;
+};
ABSL_CONST_INIT const DisplayUnit kDisplayNano = {"ns", 2, 1e2};
ABSL_CONST_INIT const DisplayUnit kDisplayMicro = {"us", 5, 1e5};
ABSL_CONST_INIT const DisplayUnit kDisplayMilli = {"ms", 8, 1e8};
@@ -722,123 +722,123 @@ ABSL_CONST_INIT const DisplayUnit kDisplaySec = {"s", 11, 1e11};
ABSL_CONST_INIT const DisplayUnit kDisplayMin = {"m", -1, 0.0}; // prec ignored
ABSL_CONST_INIT const DisplayUnit kDisplayHour = {"h", -1,
0.0}; // prec ignored
-
-void AppendNumberUnit(std::string* out, int64_t n, DisplayUnit unit) {
- char buf[sizeof("2562047788015216")]; // hours in max duration
- char* const ep = buf + sizeof(buf);
- char* bp = Format64(ep, 0, n);
- if (*bp != '0' || bp + 1 != ep) {
- out->append(bp, ep - bp);
+
+void AppendNumberUnit(std::string* out, int64_t n, DisplayUnit unit) {
+ char buf[sizeof("2562047788015216")]; // hours in max duration
+ char* const ep = buf + sizeof(buf);
+ char* bp = Format64(ep, 0, n);
+ if (*bp != '0' || bp + 1 != ep) {
+ out->append(bp, ep - bp);
out->append(unit.abbr.data(), unit.abbr.size());
- }
-}
-
-// Note: unit.prec is limited to double's digits10 value (typically 15) so it
-// always fits in buf[].
-void AppendNumberUnit(std::string* out, double n, DisplayUnit unit) {
+ }
+}
+
+// Note: unit.prec is limited to double's digits10 value (typically 15) so it
+// always fits in buf[].
+void AppendNumberUnit(std::string* out, double n, DisplayUnit unit) {
constexpr int kBufferSize = std::numeric_limits<double>::digits10;
const int prec = std::min(kBufferSize, unit.prec);
char buf[kBufferSize]; // also large enough to hold integer part
- char* ep = buf + sizeof(buf);
- double d = 0;
- int64_t frac_part = Round(std::modf(n, &d) * unit.pow10);
- int64_t int_part = d;
- if (int_part != 0 || frac_part != 0) {
- char* bp = Format64(ep, 0, int_part); // always < 1000
- out->append(bp, ep - bp);
- if (frac_part != 0) {
- out->push_back('.');
- bp = Format64(ep, prec, frac_part);
- while (ep[-1] == '0') --ep;
- out->append(bp, ep - bp);
- }
+ char* ep = buf + sizeof(buf);
+ double d = 0;
+ int64_t frac_part = Round(std::modf(n, &d) * unit.pow10);
+ int64_t int_part = d;
+ if (int_part != 0 || frac_part != 0) {
+ char* bp = Format64(ep, 0, int_part); // always < 1000
+ out->append(bp, ep - bp);
+ if (frac_part != 0) {
+ out->push_back('.');
+ bp = Format64(ep, prec, frac_part);
+ while (ep[-1] == '0') --ep;
+ out->append(bp, ep - bp);
+ }
out->append(unit.abbr.data(), unit.abbr.size());
- }
-}
-
-} // namespace
-
-// From Go's doc at https://golang.org/pkg/time/#Duration.String
-// [FormatDuration] returns a string representing the duration in the
-// form "72h3m0.5s". Leading zero units are omitted. As a special
-// case, durations less than one second format use a smaller unit
-// (milli-, micro-, or nanoseconds) to ensure that the leading digit
+ }
+}
+
+} // namespace
+
+// From Go's doc at https://golang.org/pkg/time/#Duration.String
+// [FormatDuration] returns a string representing the duration in the
+// form "72h3m0.5s". Leading zero units are omitted. As a special
+// case, durations less than one second format use a smaller unit
+// (milli-, micro-, or nanoseconds) to ensure that the leading digit
// is non-zero.
// Unlike Go, we format the zero duration as 0, with no unit.
-std::string FormatDuration(Duration d) {
- const Duration min_duration = Seconds(kint64min);
- if (d == min_duration) {
- // Avoid needing to negate kint64min by directly returning what the
- // following code should produce in that case.
- return "-2562047788015215h30m8s";
- }
- std::string s;
- if (d < ZeroDuration()) {
- s.append("-");
- d = -d;
- }
- if (d == InfiniteDuration()) {
- s.append("inf");
- } else if (d < Seconds(1)) {
- // Special case for durations with a magnitude < 1 second. The duration
- // is printed as a fraction of a single unit, e.g., "1.2ms".
- if (d < Microseconds(1)) {
- AppendNumberUnit(&s, FDivDuration(d, Nanoseconds(1)), kDisplayNano);
- } else if (d < Milliseconds(1)) {
- AppendNumberUnit(&s, FDivDuration(d, Microseconds(1)), kDisplayMicro);
- } else {
- AppendNumberUnit(&s, FDivDuration(d, Milliseconds(1)), kDisplayMilli);
- }
- } else {
- AppendNumberUnit(&s, IDivDuration(d, Hours(1), &d), kDisplayHour);
- AppendNumberUnit(&s, IDivDuration(d, Minutes(1), &d), kDisplayMin);
- AppendNumberUnit(&s, FDivDuration(d, Seconds(1)), kDisplaySec);
- }
- if (s.empty() || s == "-") {
- s = "0";
- }
- return s;
-}
-
-namespace {
-
-// A helper for ParseDuration() that parses a leading number from the given
-// string and stores the result in *int_part/*frac_part/*frac_scale. The
-// given string pointer is modified to point to the first unconsumed char.
+std::string FormatDuration(Duration d) {
+ const Duration min_duration = Seconds(kint64min);
+ if (d == min_duration) {
+ // Avoid needing to negate kint64min by directly returning what the
+ // following code should produce in that case.
+ return "-2562047788015215h30m8s";
+ }
+ std::string s;
+ if (d < ZeroDuration()) {
+ s.append("-");
+ d = -d;
+ }
+ if (d == InfiniteDuration()) {
+ s.append("inf");
+ } else if (d < Seconds(1)) {
+ // Special case for durations with a magnitude < 1 second. The duration
+ // is printed as a fraction of a single unit, e.g., "1.2ms".
+ if (d < Microseconds(1)) {
+ AppendNumberUnit(&s, FDivDuration(d, Nanoseconds(1)), kDisplayNano);
+ } else if (d < Milliseconds(1)) {
+ AppendNumberUnit(&s, FDivDuration(d, Microseconds(1)), kDisplayMicro);
+ } else {
+ AppendNumberUnit(&s, FDivDuration(d, Milliseconds(1)), kDisplayMilli);
+ }
+ } else {
+ AppendNumberUnit(&s, IDivDuration(d, Hours(1), &d), kDisplayHour);
+ AppendNumberUnit(&s, IDivDuration(d, Minutes(1), &d), kDisplayMin);
+ AppendNumberUnit(&s, FDivDuration(d, Seconds(1)), kDisplaySec);
+ }
+ if (s.empty() || s == "-") {
+ s = "0";
+ }
+ return s;
+}
+
+namespace {
+
+// A helper for ParseDuration() that parses a leading number from the given
+// string and stores the result in *int_part/*frac_part/*frac_scale. The
+// given string pointer is modified to point to the first unconsumed char.
bool ConsumeDurationNumber(const char** dpp, const char* ep, int64_t* int_part,
- int64_t* frac_part, int64_t* frac_scale) {
- *int_part = 0;
- *frac_part = 0;
- *frac_scale = 1; // invariant: *frac_part < *frac_scale
- const char* start = *dpp;
+ int64_t* frac_part, int64_t* frac_scale) {
+ *int_part = 0;
+ *frac_part = 0;
+ *frac_scale = 1; // invariant: *frac_part < *frac_scale
+ const char* start = *dpp;
for (; *dpp != ep; *dpp += 1) {
- const int d = **dpp - '0'; // contiguous digits
+ const int d = **dpp - '0'; // contiguous digits
if (d < 0 || 10 <= d) break;
- if (*int_part > kint64max / 10) return false;
- *int_part *= 10;
- if (*int_part > kint64max - d) return false;
- *int_part += d;
- }
- const bool int_part_empty = (*dpp == start);
+ if (*int_part > kint64max / 10) return false;
+ *int_part *= 10;
+ if (*int_part > kint64max - d) return false;
+ *int_part += d;
+ }
+ const bool int_part_empty = (*dpp == start);
if (*dpp == ep || **dpp != '.') return !int_part_empty;
for (*dpp += 1; *dpp != ep; *dpp += 1) {
- const int d = **dpp - '0'; // contiguous digits
+ const int d = **dpp - '0'; // contiguous digits
if (d < 0 || 10 <= d) break;
- if (*frac_scale <= kint64max / 10) {
- *frac_part *= 10;
- *frac_part += d;
- *frac_scale *= 10;
- }
- }
- return !int_part_empty || *frac_scale != 1;
-}
-
-// A helper for ParseDuration() that parses a leading unit designator (e.g.,
-// ns, us, ms, s, m, h) from the given string and stores the resulting unit
-// in "*unit". The given string pointer is modified to point to the first
-// unconsumed char.
+ if (*frac_scale <= kint64max / 10) {
+ *frac_part *= 10;
+ *frac_part += d;
+ *frac_scale *= 10;
+ }
+ }
+ return !int_part_empty || *frac_scale != 1;
+}
+
+// A helper for ParseDuration() that parses a leading unit designator (e.g.,
+// ns, us, ms, s, m, h) from the given string and stores the resulting unit
+// in "*unit". The given string pointer is modified to point to the first
+// unconsumed char.
bool ConsumeDurationUnit(const char** start, const char* end, Duration* unit) {
size_t size = end - *start;
switch (size) {
@@ -888,67 +888,67 @@ bool ConsumeDurationUnit(const char** start, const char* end, Duration* unit) {
default:
return false;
}
- }
-}
-
-} // namespace
-
-// From Go's doc at https://golang.org/pkg/time/#ParseDuration
-// [ParseDuration] parses a duration string. A duration string is
-// a possibly signed sequence of decimal numbers, each with optional
-// fraction and a unit suffix, such as "300ms", "-1.5h" or "2h45m".
-// Valid time units are "ns", "us" "ms", "s", "m", "h".
+ }
+}
+
+} // namespace
+
+// From Go's doc at https://golang.org/pkg/time/#ParseDuration
+// [ParseDuration] parses a duration string. A duration string is
+// a possibly signed sequence of decimal numbers, each with optional
+// fraction and a unit suffix, such as "300ms", "-1.5h" or "2h45m".
+// Valid time units are "ns", "us" "ms", "s", "m", "h".
bool ParseDuration(absl::string_view dur_sv, Duration* d) {
- int sign = 1;
+ int sign = 1;
if (absl::ConsumePrefix(&dur_sv, "-")) {
sign = -1;
} else {
absl::ConsumePrefix(&dur_sv, "+");
- }
+ }
if (dur_sv.empty()) return false;
-
+
// Special case for a string of "0".
if (dur_sv == "0") {
- *d = ZeroDuration();
- return true;
- }
-
+ *d = ZeroDuration();
+ return true;
+ }
+
if (dur_sv == "inf") {
- *d = sign * InfiniteDuration();
- return true;
- }
-
+ *d = sign * InfiniteDuration();
+ return true;
+ }
+
const char* start = dur_sv.data();
const char* end = start + dur_sv.size();
- Duration dur;
+ Duration dur;
while (start != end) {
- int64_t int_part;
- int64_t frac_part;
- int64_t frac_scale;
- Duration unit;
+ int64_t int_part;
+ int64_t frac_part;
+ int64_t frac_scale;
+ Duration unit;
if (!ConsumeDurationNumber(&start, end, &int_part, &frac_part,
&frac_scale) ||
!ConsumeDurationUnit(&start, end, &unit)) {
- return false;
- }
- if (int_part != 0) dur += sign * int_part * unit;
- if (frac_part != 0) dur += sign * frac_part * unit / frac_scale;
- }
- *d = dur;
- return true;
-}
-
-bool AbslParseFlag(absl::string_view text, Duration* dst, std::string*) {
- return ParseDuration(text, dst);
-}
-
-std::string AbslUnparseFlag(Duration d) { return FormatDuration(d); }
-bool ParseFlag(const std::string& text, Duration* dst, std::string* ) {
+ return false;
+ }
+ if (int_part != 0) dur += sign * int_part * unit;
+ if (frac_part != 0) dur += sign * frac_part * unit / frac_scale;
+ }
+ *d = dur;
+ return true;
+}
+
+bool AbslParseFlag(absl::string_view text, Duration* dst, std::string*) {
return ParseDuration(text, dst);
-}
-
-std::string UnparseFlag(Duration d) { return FormatDuration(d); }
-
+}
+
+std::string AbslUnparseFlag(Duration d) { return FormatDuration(d); }
+bool ParseFlag(const std::string& text, Duration* dst, std::string* ) {
+ return ParseDuration(text, dst);
+}
+
+std::string UnparseFlag(Duration d) { return FormatDuration(d); }
+
ABSL_NAMESPACE_END
-} // namespace absl
+} // namespace absl