2014-07-30 13:54:45 +00:00
|
|
|
// Copyright 2014 the V8 project authors. All rights reserved.
|
|
|
|
// Use of this source code is governed by a BSD-style license that can be
|
|
|
|
// found in the LICENSE file.
|
|
|
|
|
|
|
|
#ifndef V8_CCTEST_COMPILER_VALUE_HELPER_H_
|
|
|
|
#define V8_CCTEST_COMPILER_VALUE_HELPER_H_
|
|
|
|
|
2015-12-16 15:34:24 +00:00
|
|
|
#include <stdint.h>
|
|
|
|
|
2017-10-18 10:13:34 +00:00
|
|
|
#include "src/base/template-utils.h"
|
2014-07-30 13:54:45 +00:00
|
|
|
#include "src/compiler/common-operator.h"
|
2017-10-18 13:55:18 +00:00
|
|
|
#include "src/compiler/node-matchers.h"
|
2017-10-18 10:13:34 +00:00
|
|
|
#include "src/compiler/node.h"
|
|
|
|
#include "src/heap/heap-inl.h"
|
2014-07-30 13:54:45 +00:00
|
|
|
#include "src/isolate.h"
|
|
|
|
#include "src/objects.h"
|
|
|
|
#include "test/cctest/cctest.h"
|
|
|
|
|
|
|
|
namespace v8 {
|
|
|
|
namespace internal {
|
|
|
|
namespace compiler {
|
|
|
|
|
|
|
|
// A collection of utilities related to numerical and heap values, including
|
|
|
|
// example input values of various types, including int32_t, uint32_t, double,
|
|
|
|
// etc.
|
|
|
|
class ValueHelper {
|
|
|
|
public:
|
|
|
|
Isolate* isolate_;
|
|
|
|
|
|
|
|
ValueHelper() : isolate_(CcTest::InitIsolateOnce()) {}
|
|
|
|
|
|
|
|
void CheckFloat64Constant(double expected, Node* node) {
|
|
|
|
CHECK_EQ(IrOpcode::kFloat64Constant, node->opcode());
|
2014-09-08 09:16:11 +00:00
|
|
|
CHECK_EQ(expected, OpParameter<double>(node));
|
2014-07-30 13:54:45 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
void CheckNumberConstant(double expected, Node* node) {
|
|
|
|
CHECK_EQ(IrOpcode::kNumberConstant, node->opcode());
|
2014-09-08 09:16:11 +00:00
|
|
|
CHECK_EQ(expected, OpParameter<double>(node));
|
2014-07-30 13:54:45 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
void CheckInt32Constant(int32_t expected, Node* node) {
|
|
|
|
CHECK_EQ(IrOpcode::kInt32Constant, node->opcode());
|
2014-09-08 09:16:11 +00:00
|
|
|
CHECK_EQ(expected, OpParameter<int32_t>(node));
|
2014-07-30 13:54:45 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
void CheckUint32Constant(int32_t expected, Node* node) {
|
|
|
|
CHECK_EQ(IrOpcode::kInt32Constant, node->opcode());
|
2015-01-30 09:29:25 +00:00
|
|
|
CHECK_EQ(expected, OpParameter<int32_t>(node));
|
2014-07-30 13:54:45 +00:00
|
|
|
}
|
|
|
|
|
2015-08-31 08:24:52 +00:00
|
|
|
void CheckHeapConstant(HeapObject* expected, Node* node) {
|
2014-07-30 13:54:45 +00:00
|
|
|
CHECK_EQ(IrOpcode::kHeapConstant, node->opcode());
|
2015-08-31 08:24:52 +00:00
|
|
|
CHECK_EQ(expected, *OpParameter<Handle<HeapObject>>(node));
|
2014-07-30 13:54:45 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
void CheckTrue(Node* node) {
|
|
|
|
CheckHeapConstant(isolate_->heap()->true_value(), node);
|
|
|
|
}
|
|
|
|
|
|
|
|
void CheckFalse(Node* node) {
|
|
|
|
CheckHeapConstant(isolate_->heap()->false_value(), node);
|
|
|
|
}
|
|
|
|
|
2017-10-18 10:13:34 +00:00
|
|
|
static constexpr float float32_array[] = {
|
|
|
|
-std::numeric_limits<float>::infinity(),
|
|
|
|
-2.70497e+38f,
|
|
|
|
-1.4698e+37f,
|
|
|
|
-1.22813e+35f,
|
|
|
|
-1.20555e+35f,
|
|
|
|
-1.34584e+34f,
|
|
|
|
-1.0079e+32f,
|
|
|
|
-6.49364e+26f,
|
|
|
|
-3.06077e+25f,
|
|
|
|
-1.46821e+25f,
|
|
|
|
-1.17658e+23f,
|
|
|
|
-1.9617e+22f,
|
|
|
|
-2.7357e+20f,
|
|
|
|
-9223372036854775808.0f, // INT64_MIN
|
|
|
|
-1.48708e+13f,
|
|
|
|
-1.89633e+12f,
|
|
|
|
-4.66622e+11f,
|
|
|
|
-2.22581e+11f,
|
|
|
|
-1.45381e+10f,
|
|
|
|
-2147483904.0f, // First float32 after INT32_MIN
|
|
|
|
-2147483648.0f, // INT32_MIN
|
|
|
|
-2147483520.0f, // Last float32 before INT32_MIN
|
|
|
|
-1.3956e+09f,
|
|
|
|
-1.32951e+09f,
|
|
|
|
-1.30721e+09f,
|
|
|
|
-1.19756e+09f,
|
|
|
|
-9.26822e+08f,
|
|
|
|
-6.35647e+08f,
|
|
|
|
-4.00037e+08f,
|
|
|
|
-1.81227e+08f,
|
|
|
|
-5.09256e+07f,
|
|
|
|
-964300.0f,
|
|
|
|
-192446.0f,
|
|
|
|
-28455.0f,
|
|
|
|
-27194.0f,
|
|
|
|
-26401.0f,
|
|
|
|
-20575.0f,
|
|
|
|
-17069.0f,
|
|
|
|
-9167.0f,
|
|
|
|
-960.178f,
|
|
|
|
-113.0f,
|
|
|
|
-62.0f,
|
|
|
|
-15.0f,
|
|
|
|
-7.0f,
|
|
|
|
-1.0f,
|
|
|
|
-0.0256635f,
|
|
|
|
-4.60374e-07f,
|
|
|
|
-3.63759e-10f,
|
|
|
|
-4.30175e-14f,
|
|
|
|
-5.27385e-15f,
|
|
|
|
-1.5707963267948966f,
|
|
|
|
-1.48084e-15f,
|
|
|
|
-2.220446049250313e-16f,
|
|
|
|
-1.05755e-19f,
|
|
|
|
-3.2995e-21f,
|
|
|
|
-1.67354e-23f,
|
|
|
|
-1.11885e-23f,
|
|
|
|
-1.78506e-30f,
|
|
|
|
-5.07594e-31f,
|
|
|
|
-3.65799e-31f,
|
|
|
|
-1.43718e-34f,
|
|
|
|
-1.27126e-38f,
|
|
|
|
-0.0f,
|
|
|
|
0.0f,
|
|
|
|
1.17549e-38f,
|
|
|
|
1.56657e-37f,
|
|
|
|
4.08512e-29f,
|
|
|
|
3.31357e-28f,
|
|
|
|
6.25073e-22f,
|
|
|
|
4.1723e-13f,
|
|
|
|
1.44343e-09f,
|
|
|
|
1.5707963267948966f,
|
|
|
|
5.27004e-08f,
|
|
|
|
9.48298e-08f,
|
|
|
|
5.57888e-07f,
|
|
|
|
4.89988e-05f,
|
|
|
|
0.244326f,
|
|
|
|
1.0f,
|
|
|
|
12.4895f,
|
|
|
|
19.0f,
|
|
|
|
47.0f,
|
|
|
|
106.0f,
|
|
|
|
538.324f,
|
|
|
|
564.536f,
|
|
|
|
819.124f,
|
|
|
|
7048.0f,
|
|
|
|
12611.0f,
|
|
|
|
19878.0f,
|
|
|
|
20309.0f,
|
|
|
|
797056.0f,
|
|
|
|
1.77219e+09f,
|
|
|
|
2147483648.0f, // INT32_MAX + 1
|
|
|
|
4294967296.0f, // UINT32_MAX + 1
|
|
|
|
1.51116e+11f,
|
|
|
|
4.18193e+13f,
|
|
|
|
3.59167e+16f,
|
|
|
|
9223372036854775808.0f, // INT64_MAX + 1
|
|
|
|
18446744073709551616.0f, // UINT64_MAX + 1
|
|
|
|
3.38211e+19f,
|
|
|
|
2.67488e+20f,
|
|
|
|
1.78831e+21f,
|
|
|
|
9.20914e+21f,
|
|
|
|
8.35654e+23f,
|
|
|
|
1.4495e+24f,
|
|
|
|
5.94015e+25f,
|
|
|
|
4.43608e+30f,
|
|
|
|
2.44502e+33f,
|
|
|
|
2.61152e+33f,
|
|
|
|
1.38178e+37f,
|
|
|
|
1.71306e+37f,
|
|
|
|
3.31899e+38f,
|
|
|
|
3.40282e+38f,
|
|
|
|
std::numeric_limits<float>::infinity(),
|
|
|
|
std::numeric_limits<float>::quiet_NaN(),
|
|
|
|
-std::numeric_limits<float>::quiet_NaN()};
|
2014-07-30 13:54:45 +00:00
|
|
|
|
2017-10-18 10:13:34 +00:00
|
|
|
static constexpr Vector<const float> float32_vector() {
|
|
|
|
return ArrayVector(float32_array);
|
2014-07-30 13:54:45 +00:00
|
|
|
}
|
|
|
|
|
2017-10-18 10:13:34 +00:00
|
|
|
static constexpr double float64_array[] = {
|
|
|
|
-2e66,
|
|
|
|
-2.220446049250313e-16,
|
|
|
|
-9223373136366403584.0,
|
|
|
|
-9223372036854775808.0, // INT64_MIN
|
|
|
|
-2147483649.5,
|
|
|
|
-2147483648.25,
|
|
|
|
-2147483648.0,
|
|
|
|
-2147483647.875,
|
|
|
|
-2147483647.125,
|
|
|
|
-2147483647.0,
|
|
|
|
-999.75,
|
|
|
|
-2e66,
|
|
|
|
-1.75,
|
|
|
|
-1.5707963267948966,
|
|
|
|
-1.0,
|
|
|
|
-0.5,
|
|
|
|
-0.0,
|
|
|
|
0.0,
|
|
|
|
3e-88,
|
|
|
|
0.125,
|
|
|
|
0.25,
|
|
|
|
0.375,
|
|
|
|
0.5,
|
|
|
|
1.0,
|
|
|
|
1.17549e-38,
|
|
|
|
1.56657e-37,
|
|
|
|
1.0000001,
|
|
|
|
1.25,
|
|
|
|
1.5707963267948966,
|
|
|
|
2,
|
|
|
|
3.1e7,
|
|
|
|
5.125,
|
|
|
|
6.25,
|
|
|
|
888,
|
|
|
|
982983.25,
|
|
|
|
2147483647.0,
|
|
|
|
2147483647.375,
|
|
|
|
2147483647.75,
|
|
|
|
2147483648.0,
|
|
|
|
2147483648.25,
|
|
|
|
2147483649.25,
|
|
|
|
9223372036854775808.0, // INT64_MAX + 1
|
|
|
|
9223373136366403584.0,
|
|
|
|
18446744073709551616.0, // UINT64_MAX + 1
|
|
|
|
2e66,
|
|
|
|
V8_INFINITY,
|
|
|
|
-V8_INFINITY,
|
|
|
|
std::numeric_limits<double>::quiet_NaN(),
|
|
|
|
-std::numeric_limits<double>::quiet_NaN()};
|
|
|
|
|
|
|
|
static constexpr Vector<const double> float64_vector() {
|
|
|
|
return ArrayVector(float64_array);
|
2014-07-30 13:54:45 +00:00
|
|
|
}
|
|
|
|
|
2017-10-18 10:13:34 +00:00
|
|
|
static constexpr uint32_t uint32_array[] = {
|
|
|
|
0x00000000, 0x00000001, 0xffffffff, 0x1b09788b, 0x04c5fce8, 0xcc0de5bf,
|
|
|
|
// This row is useful for testing lea optimizations on intel.
|
|
|
|
0x00000002, 0x00000003, 0x00000004, 0x00000005, 0x00000008, 0x00000009,
|
|
|
|
0x273a798e, 0x187937a3, 0xece3af83, 0x5495a16b, 0x0b668ecc, 0x11223344,
|
|
|
|
0x0000009e, 0x00000043, 0x0000af73, 0x0000116b, 0x00658ecc, 0x002b3b4c,
|
|
|
|
0x88776655, 0x70000000, 0x07200000, 0x7fffffff, 0x56123761, 0x7fffff00,
|
|
|
|
0x761c4761, 0x80000000, 0x88888888, 0xa0000000, 0xdddddddd, 0xe0000000,
|
|
|
|
0xeeeeeeee, 0xfffffffd, 0xf0000000, 0x007fffff, 0x003fffff, 0x001fffff,
|
|
|
|
0x000fffff, 0x0007ffff, 0x0003ffff, 0x0001ffff, 0x0000ffff, 0x00007fff,
|
2017-10-20 09:12:46 +00:00
|
|
|
0x00003fff, 0x00001fff, 0x00000fff, 0x000007ff, 0x000003ff, 0x000001ff,
|
|
|
|
// Bit pattern of a quiet NaN and signaling NaN, with or without
|
|
|
|
// additional payload.
|
|
|
|
0x7fc00000, 0x7f800000, 0x7fffffff, 0x7f876543};
|
2017-10-18 10:13:34 +00:00
|
|
|
|
|
|
|
static constexpr Vector<const uint32_t> uint32_vector() {
|
|
|
|
return ArrayVector(uint32_array);
|
2015-09-23 11:04:44 +00:00
|
|
|
}
|
|
|
|
|
2017-10-18 10:13:34 +00:00
|
|
|
static constexpr Vector<const int32_t> int32_vector() {
|
|
|
|
return Vector<const int32_t>::cast(uint32_vector());
|
2015-09-23 11:04:44 +00:00
|
|
|
}
|
|
|
|
|
2017-10-18 10:13:34 +00:00
|
|
|
static constexpr uint64_t uint64_array[] = {
|
2017-10-20 09:12:46 +00:00
|
|
|
0x00000000, 0x00000001, 0xffffffff, 0x1b09788b, 0x04c5fce8, 0xcc0de5bf,
|
|
|
|
0x00000002, 0x00000003, 0x00000004, 0x00000005, 0x00000008, 0x00000009,
|
2017-10-18 10:13:34 +00:00
|
|
|
0xffffffffffffffff, 0xfffffffffffffffe, 0xfffffffffffffffd,
|
|
|
|
0x0000000000000000, 0x0000000100000000, 0xffffffff00000000,
|
|
|
|
0x1b09788b00000000, 0x04c5fce800000000, 0xcc0de5bf00000000,
|
|
|
|
0x0000000200000000, 0x0000000300000000, 0x0000000400000000,
|
|
|
|
0x0000000500000000, 0x0000000800000000, 0x0000000900000000,
|
2017-10-20 09:12:46 +00:00
|
|
|
0x273a798e187937a3, 0xece3af835495a16b, 0x0b668ecc11223344, 0x0000009e,
|
|
|
|
0x00000043, 0x0000af73, 0x0000116b, 0x00658ecc, 0x002b3b4c, 0x88776655,
|
|
|
|
0x70000000, 0x07200000, 0x7fffffff, 0x56123761, 0x7fffff00,
|
2017-10-18 10:13:34 +00:00
|
|
|
0x761c4761eeeeeeee, 0x80000000eeeeeeee, 0x88888888dddddddd,
|
|
|
|
0xa0000000dddddddd, 0xddddddddaaaaaaaa, 0xe0000000aaaaaaaa,
|
|
|
|
0xeeeeeeeeeeeeeeee, 0xfffffffdeeeeeeee, 0xf0000000dddddddd,
|
2017-10-20 09:12:46 +00:00
|
|
|
0x007fffffdddddddd, 0x003fffffaaaaaaaa, 0x001fffffaaaaaaaa, 0x000fffff,
|
|
|
|
0x0007ffff, 0x0003ffff, 0x0001ffff, 0x0000ffff, 0x00007fff, 0x00003fff,
|
|
|
|
0x00001fff, 0x00000fff, 0x000007ff, 0x000003ff, 0x000001ff,
|
2017-10-18 10:13:34 +00:00
|
|
|
0x00003fffffffffff, 0x00001fffffffffff, 0x00000fffffffffff,
|
|
|
|
0x000007ffffffffff, 0x000003ffffffffff, 0x000001ffffffffff,
|
|
|
|
0x8000008000000000, 0x8000008000000001, 0x8000000000000400,
|
2017-10-20 09:12:46 +00:00
|
|
|
0x8000000000000401, 0x0000000000000020,
|
|
|
|
// Bit pattern of a quiet NaN and signaling NaN, with or without
|
|
|
|
// additional payload.
|
|
|
|
0x7ff8000000000000, 0x7ff0000000000000, 0x7ff8123456789abc,
|
|
|
|
0x7ff7654321fedcba};
|
2017-10-18 10:13:34 +00:00
|
|
|
|
|
|
|
static constexpr Vector<const uint64_t> uint64_vector() {
|
|
|
|
return ArrayVector(uint64_array);
|
2014-07-30 13:54:45 +00:00
|
|
|
}
|
2014-07-31 07:44:29 +00:00
|
|
|
|
2017-10-18 10:13:34 +00:00
|
|
|
static constexpr Vector<const int64_t> int64_vector() {
|
|
|
|
return Vector<const int64_t>::cast(uint64_vector());
|
2017-01-26 02:18:00 +00:00
|
|
|
}
|
|
|
|
|
2017-10-18 10:13:34 +00:00
|
|
|
static constexpr int16_t int16_array[] = {
|
|
|
|
0, 1, 2, INT16_MAX - 1, INT16_MAX, INT16_MIN, INT16_MIN + 1, -2, -1};
|
|
|
|
|
|
|
|
static constexpr Vector<const int16_t> int16_vector() {
|
|
|
|
return ArrayVector(int16_array);
|
2017-08-02 21:40:45 +00:00
|
|
|
}
|
|
|
|
|
2017-10-18 10:13:34 +00:00
|
|
|
static constexpr Vector<const uint16_t> uint16_vector() {
|
|
|
|
return Vector<const uint16_t>::cast(int16_vector());
|
2017-01-26 02:18:00 +00:00
|
|
|
}
|
|
|
|
|
2017-10-18 10:13:34 +00:00
|
|
|
static constexpr int8_t int8_array[] = {
|
|
|
|
0, 1, 2, INT8_MAX - 1, INT8_MAX, INT8_MIN, INT8_MIN + 1, -2, -1};
|
|
|
|
|
|
|
|
static constexpr Vector<const int8_t> int8_vector() {
|
|
|
|
return ArrayVector(int8_array);
|
2017-08-02 21:40:45 +00:00
|
|
|
}
|
|
|
|
|
2017-10-18 10:13:34 +00:00
|
|
|
static constexpr Vector<const uint8_t> uint8_vector() {
|
|
|
|
return Vector<const uint8_t>::cast(ArrayVector(int8_array));
|
2017-10-18 13:55:18 +00:00
|
|
|
}
|
2017-10-18 10:13:34 +00:00
|
|
|
|
2017-10-18 10:13:34 +00:00
|
|
|
static constexpr uint32_t ror_array[31] = {
|
|
|
|
1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16,
|
|
|
|
17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31};
|
|
|
|
|
|
|
|
static constexpr Vector<const uint32_t> ror_vector() {
|
|
|
|
return ArrayVector(ror_array);
|
2014-07-31 07:44:29 +00:00
|
|
|
}
|
2014-07-30 13:54:45 +00:00
|
|
|
};
|
|
|
|
|
|
|
|
// Helper macros that can be used in FOR_INT32_INPUTS(i) { ... *i ... }
|
|
|
|
// Watch out, these macros aren't hygenic; they pollute your scope. Thanks STL.
|
2017-10-18 10:13:34 +00:00
|
|
|
#define FOR_INPUTS(ctype, itype, var) \
|
|
|
|
Vector<const ctype> var##_vec = \
|
|
|
|
::v8::internal::compiler::ValueHelper::itype##_vector(); \
|
|
|
|
for (Vector<const ctype>::iterator var = var##_vec.begin(), \
|
|
|
|
var##_end = var##_vec.end(); \
|
|
|
|
var != var##_end; ++var)
|
2014-07-30 13:54:45 +00:00
|
|
|
|
|
|
|
#define FOR_INT32_INPUTS(var) FOR_INPUTS(int32_t, int32, var)
|
|
|
|
#define FOR_UINT32_INPUTS(var) FOR_INPUTS(uint32_t, uint32, var)
|
2017-01-26 02:18:00 +00:00
|
|
|
#define FOR_INT16_INPUTS(var) FOR_INPUTS(int16_t, int16, var)
|
2017-08-02 21:40:45 +00:00
|
|
|
#define FOR_UINT16_INPUTS(var) FOR_INPUTS(uint16_t, uint16, var)
|
2017-01-26 02:18:00 +00:00
|
|
|
#define FOR_INT8_INPUTS(var) FOR_INPUTS(int8_t, int8, var)
|
2017-08-02 21:40:45 +00:00
|
|
|
#define FOR_UINT8_INPUTS(var) FOR_INPUTS(uint8_t, uint8, var)
|
2015-09-23 11:04:44 +00:00
|
|
|
#define FOR_INT64_INPUTS(var) FOR_INPUTS(int64_t, int64, var)
|
|
|
|
#define FOR_UINT64_INPUTS(var) FOR_INPUTS(uint64_t, uint64, var)
|
2014-09-24 11:08:35 +00:00
|
|
|
#define FOR_FLOAT32_INPUTS(var) FOR_INPUTS(float, float32, var)
|
2014-07-30 13:54:45 +00:00
|
|
|
#define FOR_FLOAT64_INPUTS(var) FOR_INPUTS(double, float64, var)
|
2014-07-31 07:44:29 +00:00
|
|
|
|
2014-08-13 10:30:00 +00:00
|
|
|
#define FOR_INT32_SHIFTS(var) for (int32_t var = 0; var < 32; var++)
|
|
|
|
|
|
|
|
#define FOR_UINT32_SHIFTS(var) for (uint32_t var = 0; var < 32; var++)
|
|
|
|
|
2016-02-18 09:55:25 +00:00
|
|
|
// TODO(bmeurer): Drop this crap once we switch to GTest/Gmock.
|
|
|
|
static inline void CheckFloatEq(volatile float x, volatile float y) {
|
|
|
|
if (std::isnan(x)) {
|
|
|
|
CHECK(std::isnan(y));
|
|
|
|
} else {
|
|
|
|
CHECK_EQ(x, y);
|
2016-08-22 13:50:23 +00:00
|
|
|
CHECK_EQ(std::signbit(x), std::signbit(y));
|
2016-02-18 09:55:25 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2017-09-01 12:57:34 +00:00
|
|
|
#define CHECK_FLOAT_EQ(lhs, rhs) \
|
|
|
|
do { \
|
|
|
|
volatile float tmp = lhs; \
|
|
|
|
::v8::internal::compiler::CheckFloatEq(tmp, rhs); \
|
2016-03-07 12:29:06 +00:00
|
|
|
} while (0)
|
|
|
|
|
2016-02-18 09:55:25 +00:00
|
|
|
static inline void CheckDoubleEq(volatile double x, volatile double y) {
|
|
|
|
if (std::isnan(x)) {
|
|
|
|
CHECK(std::isnan(y));
|
|
|
|
} else {
|
|
|
|
CHECK_EQ(x, y);
|
2016-08-22 13:50:23 +00:00
|
|
|
CHECK_EQ(std::signbit(x), std::signbit(y));
|
2016-02-18 09:55:25 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2017-09-01 12:57:34 +00:00
|
|
|
#define CHECK_DOUBLE_EQ(lhs, rhs) \
|
|
|
|
do { \
|
|
|
|
volatile double tmp = lhs; \
|
|
|
|
::v8::internal::compiler::CheckDoubleEq(tmp, rhs); \
|
2016-03-07 12:29:06 +00:00
|
|
|
} while (0)
|
|
|
|
|
2014-07-31 07:44:29 +00:00
|
|
|
} // namespace compiler
|
|
|
|
} // namespace internal
|
|
|
|
} // namespace v8
|
2014-07-30 13:54:45 +00:00
|
|
|
|
|
|
|
#endif // V8_CCTEST_COMPILER_VALUE_HELPER_H_
|