mirror of
				https://github.com/LadybirdBrowser/ladybird.git
				synced 2025-10-31 13:20:59 +00:00 
			
		
		
		
	
		
			
				
	
	
		
			1255 lines
		
	
	
	
		
			43 KiB
		
	
	
	
		
			C++
		
	
	
	
	
	
			
		
		
	
	
			1255 lines
		
	
	
	
		
			43 KiB
		
	
	
	
		
			C++
		
	
	
	
	
	
| /*
 | ||
|  * Copyright (c) 2020-2023, Andreas Kling <andreas@ladybird.org>
 | ||
|  * Copyright (c) 2020-2023, Linus Groh <linusg@serenityos.org>
 | ||
|  * Copyright (c) 2021, Idan Horowitz <idan.horowitz@serenityos.org>
 | ||
|  * Copyright (c) 2023, Shannon Booth <shannon@serenityos.org>
 | ||
|  *
 | ||
|  * SPDX-License-Identifier: BSD-2-Clause
 | ||
|  */
 | ||
| 
 | ||
| #include <AK/BuiltinWrappers.h>
 | ||
| #include <AK/Function.h>
 | ||
| #include <AK/Random.h>
 | ||
| #include <LibJS/Runtime/AbstractOperations.h>
 | ||
| #include <LibJS/Runtime/GlobalObject.h>
 | ||
| #include <LibJS/Runtime/Iterator.h>
 | ||
| #include <LibJS/Runtime/MathObject.h>
 | ||
| #include <LibJS/Runtime/ValueInlines.h>
 | ||
| #include <math.h>
 | ||
| 
 | ||
| namespace JS {
 | ||
| 
 | ||
| GC_DEFINE_ALLOCATOR(MathObject);
 | ||
| 
 | ||
| MathObject::MathObject(Realm& realm)
 | ||
|     : Object(ConstructWithPrototypeTag::Tag, realm.intrinsics().object_prototype())
 | ||
| {
 | ||
| }
 | ||
| 
 | ||
| void MathObject::initialize(Realm& realm)
 | ||
| {
 | ||
|     auto& vm = this->vm();
 | ||
|     Base::initialize(realm);
 | ||
| 
 | ||
|     u8 attr = Attribute::Writable | Attribute::Configurable;
 | ||
|     define_native_function(realm, vm.names.abs, abs, 1, attr, Bytecode::Builtin::MathAbs);
 | ||
|     define_native_function(realm, vm.names.random, random, 0, attr, Bytecode::Builtin::MathRandom);
 | ||
|     define_native_function(realm, vm.names.sqrt, sqrt, 1, attr, Bytecode::Builtin::MathSqrt);
 | ||
|     define_native_function(realm, vm.names.floor, floor, 1, attr, Bytecode::Builtin::MathFloor);
 | ||
|     define_native_function(realm, vm.names.ceil, ceil, 1, attr, Bytecode::Builtin::MathCeil);
 | ||
|     define_native_function(realm, vm.names.round, round, 1, attr, Bytecode::Builtin::MathRound);
 | ||
|     define_native_function(realm, vm.names.max, max, 2, attr);
 | ||
|     define_native_function(realm, vm.names.min, min, 2, attr);
 | ||
|     define_native_function(realm, vm.names.trunc, trunc, 1, attr);
 | ||
|     define_native_function(realm, vm.names.sin, sin, 1, attr);
 | ||
|     define_native_function(realm, vm.names.cos, cos, 1, attr);
 | ||
|     define_native_function(realm, vm.names.tan, tan, 1, attr);
 | ||
|     define_native_function(realm, vm.names.pow, pow, 2, attr, Bytecode::Builtin::MathPow);
 | ||
|     define_native_function(realm, vm.names.exp, exp, 1, attr, Bytecode::Builtin::MathExp);
 | ||
|     define_native_function(realm, vm.names.expm1, expm1, 1, attr);
 | ||
|     define_native_function(realm, vm.names.sign, sign, 1, attr);
 | ||
|     define_native_function(realm, vm.names.clz32, clz32, 1, attr);
 | ||
|     define_native_function(realm, vm.names.acos, acos, 1, attr);
 | ||
|     define_native_function(realm, vm.names.acosh, acosh, 1, attr);
 | ||
|     define_native_function(realm, vm.names.asin, asin, 1, attr);
 | ||
|     define_native_function(realm, vm.names.asinh, asinh, 1, attr);
 | ||
|     define_native_function(realm, vm.names.atan, atan, 1, attr);
 | ||
|     define_native_function(realm, vm.names.atanh, atanh, 1, attr);
 | ||
|     define_native_function(realm, vm.names.log1p, log1p, 1, attr);
 | ||
|     define_native_function(realm, vm.names.cbrt, cbrt, 1, attr);
 | ||
|     define_native_function(realm, vm.names.atan2, atan2, 2, attr);
 | ||
|     define_native_function(realm, vm.names.fround, fround, 1, attr);
 | ||
|     define_native_function(realm, vm.names.f16round, f16round, 1, attr);
 | ||
|     define_native_function(realm, vm.names.hypot, hypot, 2, attr);
 | ||
|     define_native_function(realm, vm.names.imul, imul, 2, attr, Bytecode::Builtin::MathImul);
 | ||
|     define_native_function(realm, vm.names.log, log, 1, attr, Bytecode::Builtin::MathLog);
 | ||
|     define_native_function(realm, vm.names.log2, log2, 1, attr);
 | ||
|     define_native_function(realm, vm.names.log10, log10, 1, attr);
 | ||
|     define_native_function(realm, vm.names.sinh, sinh, 1, attr);
 | ||
|     define_native_function(realm, vm.names.cosh, cosh, 1, attr);
 | ||
|     define_native_function(realm, vm.names.tanh, tanh, 1, attr);
 | ||
|     define_native_function(realm, vm.names.sumPrecise, sumPrecise, 1, attr);
 | ||
| 
 | ||
|     // 21.3.1 Value Properties of the Math Object, https://tc39.es/ecma262/#sec-value-properties-of-the-math-object
 | ||
|     define_direct_property(vm.names.E, Value(M_E), 0);
 | ||
|     define_direct_property(vm.names.LN2, Value(M_LN2), 0);
 | ||
|     define_direct_property(vm.names.LN10, Value(M_LN10), 0);
 | ||
|     define_direct_property(vm.names.LOG2E, Value(::log2(M_E)), 0);
 | ||
|     define_direct_property(vm.names.LOG10E, Value(::log10(M_E)), 0);
 | ||
|     define_direct_property(vm.names.PI, Value(M_PI), 0);
 | ||
|     define_direct_property(vm.names.SQRT1_2, Value(M_SQRT1_2), 0);
 | ||
|     define_direct_property(vm.names.SQRT2, Value(M_SQRT2), 0);
 | ||
| 
 | ||
|     // 21.3.1.9 Math [ @@toStringTag ], https://tc39.es/ecma262/#sec-math-@@tostringtag
 | ||
|     define_direct_property(vm.well_known_symbol_to_string_tag(), PrimitiveString::create(vm, vm.names.Math.as_string()), Attribute::Configurable);
 | ||
| }
 | ||
| 
 | ||
| // 21.3.2.1 Math.abs ( x ), https://tc39.es/ecma262/#sec-math.abs
 | ||
| ThrowCompletionOr<Value> MathObject::abs_impl(VM& vm, Value x)
 | ||
| {
 | ||
|     // OPTIMIZATION: Fast path for Int32 values.
 | ||
|     if (x.is_int32())
 | ||
|         return Value(AK::abs(x.as_i32()));
 | ||
| 
 | ||
|     // Let n be ? ToNumber(x).
 | ||
|     auto number = TRY(x.to_number(vm));
 | ||
| 
 | ||
|     // 2. If n is NaN, return NaN.
 | ||
|     if (number.is_nan())
 | ||
|         return js_nan();
 | ||
| 
 | ||
|     // 3. If n is -0𝔽, return +0𝔽.
 | ||
|     if (number.is_negative_zero())
 | ||
|         return Value(0);
 | ||
| 
 | ||
|     // 4. If n is -∞𝔽, return +∞𝔽.
 | ||
|     if (number.is_negative_infinity())
 | ||
|         return js_infinity();
 | ||
| 
 | ||
|     // 5. If n < -0𝔽, return -n.
 | ||
|     // 6. Return n.
 | ||
|     return Value(number.as_double() < 0 ? -number.as_double() : number.as_double());
 | ||
| }
 | ||
| 
 | ||
| // 21.3.2.1 Math.abs ( x ), https://tc39.es/ecma262/#sec-math.abs
 | ||
| JS_DEFINE_NATIVE_FUNCTION(MathObject::abs)
 | ||
| {
 | ||
|     return abs_impl(vm, vm.argument(0));
 | ||
| }
 | ||
| 
 | ||
| // 21.3.2.2 Math.acos ( x ), https://tc39.es/ecma262/#sec-math.acos
 | ||
| JS_DEFINE_NATIVE_FUNCTION(MathObject::acos)
 | ||
| {
 | ||
|     // 1. Let n be ? ToNumber(x).
 | ||
|     auto number = TRY(vm.argument(0).to_number(vm));
 | ||
| 
 | ||
|     // 2. If n is NaN, n > 1𝔽, or n < -1𝔽, return NaN.
 | ||
|     if (number.is_nan() || number.as_double() > 1 || number.as_double() < -1)
 | ||
|         return js_nan();
 | ||
| 
 | ||
|     // 3. If n is 1𝔽, return +0𝔽.
 | ||
|     if (number.as_double() == 1)
 | ||
|         return Value(0);
 | ||
| 
 | ||
|     // 4. Return an implementation-approximated Number value representing the result of the inverse cosine of ℝ(n).
 | ||
|     return Value(::acos(number.as_double()));
 | ||
| }
 | ||
| 
 | ||
| // 21.3.2.3 Math.acosh ( x ), https://tc39.es/ecma262/#sec-math.acosh
 | ||
| JS_DEFINE_NATIVE_FUNCTION(MathObject::acosh)
 | ||
| {
 | ||
|     // 1. Let n be ? ToNumber(x).
 | ||
|     auto number = TRY(vm.argument(0).to_number(vm));
 | ||
| 
 | ||
|     // 2. If n is NaN or n is +∞𝔽, return n.
 | ||
|     if (number.is_nan() || number.is_positive_infinity())
 | ||
|         return number;
 | ||
| 
 | ||
|     // 3. If n is 1𝔽, return +0𝔽.
 | ||
|     if (number.as_double() == 1.0)
 | ||
|         return Value(0.0);
 | ||
| 
 | ||
|     // 4. If n < 1𝔽, return NaN.
 | ||
|     if (number.as_double() < 1)
 | ||
|         return js_nan();
 | ||
| 
 | ||
|     // 5. Return an implementation-approximated Number value representing the result of the inverse hyperbolic cosine of ℝ(n).
 | ||
|     return Value(::acosh(number.as_double()));
 | ||
| }
 | ||
| 
 | ||
| // 21.3.2.4 Math.asin ( x ), https://tc39.es/ecma262/#sec-math.asin
 | ||
| JS_DEFINE_NATIVE_FUNCTION(MathObject::asin)
 | ||
| {
 | ||
|     // 1. Let n be ? ToNumber(x).
 | ||
|     auto number = TRY(vm.argument(0).to_number(vm));
 | ||
| 
 | ||
|     // 2. If n is NaN, n is +0𝔽, or n is -0𝔽, return n.
 | ||
|     if (number.is_nan() || number.is_positive_zero() || number.is_negative_zero())
 | ||
|         return number;
 | ||
| 
 | ||
|     // 3. If n > 1𝔽 or n < -1𝔽, return NaN.
 | ||
|     if (number.as_double() > 1 || number.as_double() < -1)
 | ||
|         return js_nan();
 | ||
| 
 | ||
|     // 4. Return an implementation-approximated Number value representing the result of the inverse sine of ℝ(n).
 | ||
|     return Value(::asin(number.as_double()));
 | ||
| }
 | ||
| 
 | ||
| // 21.3.2.5 Math.asinh ( x ), https://tc39.es/ecma262/#sec-math.asinh
 | ||
| JS_DEFINE_NATIVE_FUNCTION(MathObject::asinh)
 | ||
| {
 | ||
|     // 1. Let n be ? ToNumber(x).
 | ||
|     auto number = TRY(vm.argument(0).to_number(vm));
 | ||
| 
 | ||
|     // 2. If n is not finite or n is either +0𝔽 or -0𝔽, return n.
 | ||
|     if (!number.is_finite_number() || number.is_positive_zero() || number.is_negative_zero())
 | ||
|         return number;
 | ||
| 
 | ||
|     // 3. Return an implementation-approximated Number value representing the result of the inverse hyperbolic sine of ℝ(n).
 | ||
|     return Value(::asinh(number.as_double()));
 | ||
| }
 | ||
| 
 | ||
| // 21.3.2.6 Math.atan ( x ), https://tc39.es/ecma262/#sec-math.atan
 | ||
| JS_DEFINE_NATIVE_FUNCTION(MathObject::atan)
 | ||
| {
 | ||
|     // Let n be ? ToNumber(x).
 | ||
|     auto number = TRY(vm.argument(0).to_number(vm));
 | ||
| 
 | ||
|     // 2. If n is one of NaN, +0𝔽, or -0𝔽, return n.
 | ||
|     if (number.is_nan() || number.as_double() == 0)
 | ||
|         return number;
 | ||
| 
 | ||
|     // 3. If n is +∞𝔽, return an implementation-approximated Number value representing π / 2.
 | ||
|     if (number.is_positive_infinity())
 | ||
|         return Value(M_PI_2);
 | ||
| 
 | ||
|     // 4. If n is -∞𝔽, return an implementation-approximated Number value representing -π / 2.
 | ||
|     if (number.is_negative_infinity())
 | ||
|         return Value(-M_PI_2);
 | ||
| 
 | ||
|     // 5. Return an implementation-approximated Number value representing the result of the inverse tangent of ℝ(n).
 | ||
|     return Value(::atan(number.as_double()));
 | ||
| }
 | ||
| 
 | ||
| // 21.3.2.7 Math.atanh ( x ), https://tc39.es/ecma262/#sec-math.atanh
 | ||
| JS_DEFINE_NATIVE_FUNCTION(MathObject::atanh)
 | ||
| {
 | ||
|     // 1. Let n be ? ToNumber(x).
 | ||
|     auto number = TRY(vm.argument(0).to_number(vm));
 | ||
| 
 | ||
|     // 2. If n is NaN, n is +0𝔽, or n is -0𝔽, return n.
 | ||
|     if (number.is_nan() || number.is_positive_zero() || number.is_negative_zero())
 | ||
|         return number;
 | ||
| 
 | ||
|     // 3. If n > 1𝔽 or n < -1𝔽, return NaN.
 | ||
|     if (number.as_double() > 1. || number.as_double() < -1.)
 | ||
|         return js_nan();
 | ||
| 
 | ||
|     // 4. If n is 1𝔽, return +∞𝔽.
 | ||
|     if (number.as_double() == 1.)
 | ||
|         return js_infinity();
 | ||
| 
 | ||
|     // 5. If n is -1𝔽, return -∞𝔽.
 | ||
|     if (number.as_double() == -1.)
 | ||
|         return js_negative_infinity();
 | ||
| 
 | ||
|     // 6. Return an implementation-approximated Number value representing the result of the inverse hyperbolic tangent of ℝ(n).
 | ||
|     return Value(::atanh(number.as_double()));
 | ||
| }
 | ||
| 
 | ||
| // 21.3.2.8 Math.atan2 ( y, x ), https://tc39.es/ecma262/#sec-math.atan2
 | ||
| JS_DEFINE_NATIVE_FUNCTION(MathObject::atan2)
 | ||
| {
 | ||
|     auto constexpr three_quarters_pi = M_PI_4 + M_PI_2;
 | ||
| 
 | ||
|     // 1. Let ny be ? ToNumber(y).
 | ||
|     auto y = TRY(vm.argument(0).to_number(vm));
 | ||
| 
 | ||
|     // 2. Let nx be ? ToNumber(x).
 | ||
|     auto x = TRY(vm.argument(1).to_number(vm));
 | ||
| 
 | ||
|     // 3. If ny is NaN or nx is NaN, return NaN.
 | ||
|     if (y.is_nan() || x.is_nan())
 | ||
|         return js_nan();
 | ||
| 
 | ||
|     // 4. If ny is +∞𝔽, then
 | ||
|     if (y.is_positive_infinity()) {
 | ||
|         // a. If nx is +∞𝔽, return an implementation-approximated Number value representing π / 4.
 | ||
|         if (x.is_positive_infinity())
 | ||
|             return Value(M_PI_4);
 | ||
| 
 | ||
|         // b. If nx is -∞𝔽, return an implementation-approximated Number value representing 3π / 4.
 | ||
|         if (x.is_negative_infinity())
 | ||
|             return Value(three_quarters_pi);
 | ||
| 
 | ||
|         // c. Return an implementation-approximated Number value representing π / 2.
 | ||
|         return Value(M_PI_2);
 | ||
|     }
 | ||
| 
 | ||
|     // 5. If ny is -∞𝔽, then
 | ||
|     if (y.is_negative_infinity()) {
 | ||
|         // a. If nx is +∞𝔽, return an implementation-approximated Number value representing -π / 4.
 | ||
|         if (x.is_positive_infinity())
 | ||
|             return Value(-M_PI_4);
 | ||
| 
 | ||
|         // b. If nx is -∞𝔽, return an implementation-approximated Number value representing -3π / 4.
 | ||
|         if (x.is_negative_infinity())
 | ||
|             return Value(-three_quarters_pi);
 | ||
| 
 | ||
|         // c. Return an implementation-approximated Number value representing -π / 2.
 | ||
|         return Value(-M_PI_2);
 | ||
|     }
 | ||
| 
 | ||
|     // 6. If ny is +0𝔽, then
 | ||
|     if (y.is_positive_zero()) {
 | ||
|         // a. If nx > +0𝔽 or nx is +0𝔽, return +0𝔽.
 | ||
|         if (x.as_double() > 0 || x.is_positive_zero())
 | ||
|             return Value(0.0);
 | ||
| 
 | ||
|         // b. Return an implementation-approximated Number value representing π.
 | ||
|         return Value(M_PI);
 | ||
|     }
 | ||
| 
 | ||
|     // 7. If ny is -0𝔽, then
 | ||
|     if (y.is_negative_zero()) {
 | ||
|         // a. If nx > +0𝔽 or nx is +0𝔽, return -0𝔽
 | ||
|         if (x.as_double() > 0 || x.is_positive_zero())
 | ||
|             return Value(-0.0);
 | ||
| 
 | ||
|         // b. Return an implementation-approximated Number value representing -π.
 | ||
|         return Value(-M_PI);
 | ||
|     }
 | ||
| 
 | ||
|     // 8. Assert: ny is finite and is neither +0𝔽 nor -0𝔽.
 | ||
|     VERIFY(y.is_finite_number() && !y.is_positive_zero() && !y.is_negative_zero());
 | ||
| 
 | ||
|     // 9. If ny > +0𝔽, then
 | ||
|     if (y.as_double() > 0) {
 | ||
|         // a. If nx is +∞𝔽, return +0𝔽.
 | ||
|         if (x.is_positive_infinity())
 | ||
|             return Value(0);
 | ||
| 
 | ||
|         // b. If nx is -∞𝔽, return an implementation-approximated Number value representing π.
 | ||
|         if (x.is_negative_infinity())
 | ||
|             return Value(M_PI);
 | ||
| 
 | ||
|         // c. If nx is either +0𝔽 or -0𝔽, return an implementation-approximated Number value representing π / 2.
 | ||
|         if (x.is_positive_zero() || x.is_negative_zero())
 | ||
|             return Value(M_PI_2);
 | ||
|     }
 | ||
| 
 | ||
|     // 10. If ny < -0𝔽, then
 | ||
|     if (y.as_double() < -0) {
 | ||
|         // a. If nx is +∞𝔽, return -0𝔽.
 | ||
|         if (x.is_positive_infinity())
 | ||
|             return Value(-0.0);
 | ||
| 
 | ||
|         // b. If nx is -∞𝔽, return an implementation-approximated Number value representing -π.
 | ||
|         if (x.is_negative_infinity())
 | ||
|             return Value(-M_PI);
 | ||
| 
 | ||
|         // c. If nx is either +0𝔽 or -0𝔽, return an implementation-approximated Number value representing -π / 2.
 | ||
|         if (x.is_positive_zero() || x.is_negative_zero())
 | ||
|             return Value(-M_PI_2);
 | ||
|     }
 | ||
| 
 | ||
|     // 11. Assert: nx is finite and is neither +0𝔽 nor -0𝔽.
 | ||
|     VERIFY(x.is_finite_number() && !x.is_positive_zero() && !x.is_negative_zero());
 | ||
| 
 | ||
|     // 12. Return an implementation-approximated Number value representing the result of the inverse tangent of the quotient ℝ(ny) / ℝ(nx).
 | ||
|     return Value(::atan2(y.as_double(), x.as_double()));
 | ||
| }
 | ||
| 
 | ||
| // 21.3.2.9 Math.cbrt ( x ), https://tc39.es/ecma262/#sec-math.cbrt
 | ||
| JS_DEFINE_NATIVE_FUNCTION(MathObject::cbrt)
 | ||
| {
 | ||
|     // 1. Let n be ? ToNumber(x).
 | ||
|     auto number = TRY(vm.argument(0).to_number(vm));
 | ||
| 
 | ||
|     // 2. If n is not finite or n is either +0𝔽 or -0𝔽, return n.
 | ||
|     if (!number.is_finite_number() || number.as_double() == 0)
 | ||
|         return number;
 | ||
| 
 | ||
|     // 3. Return an implementation-approximated Number value representing the result of the cube root of ℝ(n).
 | ||
|     return Value(::cbrt(number.as_double()));
 | ||
| }
 | ||
| 
 | ||
| // 21.3.2.10 Math.ceil ( x ), https://tc39.es/ecma262/#sec-math.ceil
 | ||
| ThrowCompletionOr<Value> MathObject::ceil_impl(VM& vm, Value x)
 | ||
| {
 | ||
|     // 1. Let n be ? ToNumber(x).
 | ||
|     auto number = TRY(x.to_number(vm));
 | ||
| 
 | ||
|     // 2. If n is not finite or n is either +0𝔽 or -0𝔽, return n.
 | ||
|     if (!number.is_finite_number() || number.as_double() == 0)
 | ||
|         return number;
 | ||
| 
 | ||
|     // 3. If n < -0𝔽 and n > -1𝔽, return -0𝔽.
 | ||
|     if (number.as_double() < 0 && number.as_double() > -1)
 | ||
|         return Value(-0.f);
 | ||
| 
 | ||
|     // 4. If n is an integral Number, return n.
 | ||
|     // 5. Return the smallest (closest to -∞) integral Number value that is not less than n.
 | ||
|     return Value(::ceil(number.as_double()));
 | ||
| }
 | ||
| 
 | ||
| // 21.3.2.10 Math.ceil ( x ), https://tc39.es/ecma262/#sec-math.ceil
 | ||
| JS_DEFINE_NATIVE_FUNCTION(MathObject::ceil)
 | ||
| {
 | ||
|     return ceil_impl(vm, vm.argument(0));
 | ||
| }
 | ||
| 
 | ||
| // 21.3.2.11 Math.clz32 ( x ), https://tc39.es/ecma262/#sec-math.clz32
 | ||
| JS_DEFINE_NATIVE_FUNCTION(MathObject::clz32)
 | ||
| {
 | ||
|     // 1. Let n be ? ToUint32(x).
 | ||
|     auto number = TRY(vm.argument(0).to_u32(vm));
 | ||
| 
 | ||
|     // 2. Let p be the number of leading zero bits in the unsigned 32-bit binary representation of n.
 | ||
|     // 3. Return 𝔽(p).
 | ||
|     return Value(count_leading_zeroes_safe(number));
 | ||
| }
 | ||
| 
 | ||
| // 21.3.2.12 Math.cos ( x ), https://tc39.es/ecma262/#sec-math.cos
 | ||
| JS_DEFINE_NATIVE_FUNCTION(MathObject::cos)
 | ||
| {
 | ||
|     // 1. Let n be ? ToNumber(x).
 | ||
|     auto number = TRY(vm.argument(0).to_number(vm));
 | ||
| 
 | ||
|     // 2. If n is NaN, n is +∞𝔽, or n is -∞𝔽, return NaN.
 | ||
|     if (number.is_nan() || number.is_infinity())
 | ||
|         return js_nan();
 | ||
| 
 | ||
|     // 3. If n is +0𝔽 or n is -0𝔽, return 1𝔽.
 | ||
|     if (number.is_positive_zero() || number.is_negative_zero())
 | ||
|         return Value(1);
 | ||
| 
 | ||
|     // 4. Return an implementation-approximated Number value representing the result of the cosine of ℝ(n).
 | ||
|     return Value(::cos(number.as_double()));
 | ||
| }
 | ||
| 
 | ||
| // 21.3.2.13 Math.cosh ( x ), https://tc39.es/ecma262/#sec-math.cosh
 | ||
| JS_DEFINE_NATIVE_FUNCTION(MathObject::cosh)
 | ||
| {
 | ||
|     // 1. Let n be ? ToNumber(x).
 | ||
|     auto number = TRY(vm.argument(0).to_number(vm));
 | ||
| 
 | ||
|     // 2. If n is NaN, return NaN.
 | ||
|     if (number.is_nan())
 | ||
|         return js_nan();
 | ||
| 
 | ||
|     // 3. If n is +∞𝔽 or n is -∞𝔽, return +∞𝔽.
 | ||
|     if (number.is_positive_infinity() || number.is_negative_infinity())
 | ||
|         return js_infinity();
 | ||
| 
 | ||
|     // 4. If n is +0𝔽 or n is -0𝔽, return 1𝔽.
 | ||
|     if (number.is_positive_zero() || number.is_negative_zero())
 | ||
|         return Value(1);
 | ||
| 
 | ||
|     // 5. Return an implementation-approximated Number value representing the result of the hyperbolic cosine of ℝ(n).
 | ||
|     return Value(::cosh(number.as_double()));
 | ||
| }
 | ||
| 
 | ||
| // 21.3.2.14 Math.exp ( x ), https://tc39.es/ecma262/#sec-math.exp
 | ||
| ThrowCompletionOr<Value> MathObject::exp_impl(VM& vm, Value x)
 | ||
| {
 | ||
|     // 1. Let n be ? ToNumber(x).
 | ||
|     auto number = TRY(x.to_number(vm));
 | ||
| 
 | ||
|     // 2. If n is either NaN or +∞𝔽, return n.
 | ||
|     if (number.is_nan() || number.is_positive_infinity())
 | ||
|         return number;
 | ||
| 
 | ||
|     // 3. If n is either +0𝔽 or -0𝔽, return 1𝔽.
 | ||
|     if (number.as_double() == 0)
 | ||
|         return Value(1);
 | ||
| 
 | ||
|     // 4. If n is -∞𝔽, return +0𝔽.
 | ||
|     if (number.is_negative_infinity())
 | ||
|         return Value(0);
 | ||
| 
 | ||
|     // 5. Return an implementation-approximated Number value representing the result of the exponential function of ℝ(n).
 | ||
|     return Value(::exp(number.as_double()));
 | ||
| }
 | ||
| 
 | ||
| // 21.3.2.14 Math.exp ( x ), https://tc39.es/ecma262/#sec-math.exp
 | ||
| JS_DEFINE_NATIVE_FUNCTION(MathObject::exp)
 | ||
| {
 | ||
|     return exp_impl(vm, vm.argument(0));
 | ||
| }
 | ||
| 
 | ||
| // 21.3.2.15 Math.expm1 ( x ), https://tc39.es/ecma262/#sec-math.expm1
 | ||
| JS_DEFINE_NATIVE_FUNCTION(MathObject::expm1)
 | ||
| {
 | ||
|     // 1. Let n be ? ToNumber(x).
 | ||
|     auto number = TRY(vm.argument(0).to_number(vm));
 | ||
| 
 | ||
|     // 2. If n is one of NaN, +0𝔽, -0𝔽, or +∞𝔽, return n.
 | ||
|     if (number.is_nan() || number.as_double() == 0 || number.is_positive_infinity())
 | ||
|         return number;
 | ||
| 
 | ||
|     // 3. If n is -∞𝔽, return -1𝔽.
 | ||
|     if (number.is_negative_infinity())
 | ||
|         return Value(-1);
 | ||
| 
 | ||
|     // 4. Return an implementation-approximated Number value representing the result of subtracting 1 from the exponential function of ℝ(n).
 | ||
|     return Value(::expm1(number.as_double()));
 | ||
| }
 | ||
| 
 | ||
| // 21.3.2.16 Math.floor ( x ), https://tc39.es/ecma262/#sec-math.floor
 | ||
| ThrowCompletionOr<Value> MathObject::floor_impl(VM& vm, Value x)
 | ||
| {
 | ||
|     // 1. Let n be ? ToNumber(x).
 | ||
|     auto number = TRY(x.to_number(vm));
 | ||
| 
 | ||
|     // 2. If n is not finite or n is either +0𝔽 or -0𝔽, return n.
 | ||
|     if (!number.is_finite_number() || number.as_double() == 0)
 | ||
|         return number;
 | ||
| 
 | ||
|     // 3. If n < 1𝔽 and n > +0𝔽, return +0𝔽.
 | ||
|     // 4. If n is an integral Number, return n.
 | ||
|     // 5. Return the greatest (closest to +∞) integral Number value that is not greater than n.
 | ||
|     return Value(::floor(number.as_double()));
 | ||
| }
 | ||
| 
 | ||
| // 21.3.2.16 Math.floor ( x ), https://tc39.es/ecma262/#sec-math.floor
 | ||
| JS_DEFINE_NATIVE_FUNCTION(MathObject::floor)
 | ||
| {
 | ||
|     return floor_impl(vm, vm.argument(0));
 | ||
| }
 | ||
| 
 | ||
| // 21.3.2.17 Math.fround ( x ), https://tc39.es/ecma262/#sec-math.fround
 | ||
| JS_DEFINE_NATIVE_FUNCTION(MathObject::fround)
 | ||
| {
 | ||
|     // 1. Let n be ? ToNumber(x).
 | ||
|     auto number = TRY(vm.argument(0).to_number(vm));
 | ||
| 
 | ||
|     // 2. If n is NaN, return NaN.
 | ||
|     if (number.is_nan())
 | ||
|         return js_nan();
 | ||
| 
 | ||
|     // 3. If n is one of +0𝔽, -0𝔽, +∞𝔽, or -∞𝔽, return n.
 | ||
|     if (number.as_double() == 0 || number.is_infinity())
 | ||
|         return number;
 | ||
| 
 | ||
|     // 4. Let n32 be the result of converting n to a value in IEEE 754-2019 binary32 format using roundTiesToEven mode.
 | ||
|     // 5. Let n64 be the result of converting n32 to a value in IEEE 754-2019 binary64 format.
 | ||
|     // 6. Return the ECMAScript Number value corresponding to n64.
 | ||
|     return Value((float)number.as_double());
 | ||
| }
 | ||
| 
 | ||
| // 21.3.2.18 Math.f16round ( x ), https://tc39.es/ecma262/#sec-math.f16round
 | ||
| JS_DEFINE_NATIVE_FUNCTION(MathObject::f16round)
 | ||
| {
 | ||
|     // 1. Let n be ? ToNumber(x).
 | ||
|     auto number = TRY(vm.argument(0).to_number(vm));
 | ||
| 
 | ||
|     // 2. If n is NaN, return NaN.
 | ||
|     if (number.is_nan())
 | ||
|         return js_nan();
 | ||
| 
 | ||
|     // 3. If n is one of +0𝔽, -0𝔽, +∞𝔽, or -∞𝔽, return n.
 | ||
|     if (number.as_double() == 0 || number.is_infinity())
 | ||
|         return number;
 | ||
| 
 | ||
|     // 4. Let n16 be the result of converting n to IEEE 754-2019 binary16 format using roundTiesToEven mode.
 | ||
|     // 5. Let n64 be the result of converting n16 to IEEE 754-2019 binary64 format.
 | ||
|     // 6. Return the ECMAScript Number value corresponding to n64.
 | ||
|     return Value(static_cast<f16>(number.as_double()));
 | ||
| }
 | ||
| 
 | ||
| // 21.3.2.19 Math.hypot ( ...args ), https://tc39.es/ecma262/#sec-math.hypot
 | ||
| JS_DEFINE_NATIVE_FUNCTION(MathObject::hypot)
 | ||
| {
 | ||
|     // 1. Let coerced be a new empty List.
 | ||
|     Vector<Value> coerced;
 | ||
| 
 | ||
|     // 2. For each element arg of args, do
 | ||
|     for (size_t i = 0; i < vm.argument_count(); ++i) {
 | ||
|         // a. Let n be ? ToNumber(arg).
 | ||
|         auto number = TRY(vm.argument(i).to_number(vm));
 | ||
| 
 | ||
|         // b. Append n to coerced.
 | ||
|         coerced.append(number);
 | ||
|     }
 | ||
| 
 | ||
|     // 3. For each element number of coerced, do
 | ||
|     for (auto& number : coerced) {
 | ||
|         // a. If number is either +∞𝔽 or -∞𝔽, return +∞𝔽.
 | ||
|         if (number.is_infinity())
 | ||
|             return js_infinity();
 | ||
|     }
 | ||
| 
 | ||
|     // 4. Let onlyZero be true.
 | ||
|     auto only_zero = true;
 | ||
| 
 | ||
|     double sum_of_squares = 0;
 | ||
| 
 | ||
|     // 5. For each element number of coerced, do
 | ||
|     for (auto& number : coerced) {
 | ||
|         // a. If number is NaN, return NaN.
 | ||
|         // OPTIMIZATION: For infinities, the result will be infinity with the same sign, so we can return early.
 | ||
|         if (number.is_nan() || number.is_infinity())
 | ||
|             return number;
 | ||
| 
 | ||
|         // b. If number is neither +0𝔽 nor -0𝔽, set onlyZero to false.
 | ||
|         if (number.as_double() != 0)
 | ||
|             only_zero = false;
 | ||
| 
 | ||
|         sum_of_squares += number.as_double() * number.as_double();
 | ||
|     }
 | ||
| 
 | ||
|     // 6. If onlyZero is true, return +0𝔽.
 | ||
|     if (only_zero)
 | ||
|         return Value(0);
 | ||
| 
 | ||
|     // 7. Return an implementation-approximated Number value representing the square root of the sum of squares of the mathematical values of the elements of coerced.
 | ||
|     return Value(::sqrt(sum_of_squares));
 | ||
| }
 | ||
| 
 | ||
| // 21.3.2.20 Math.imul ( x, y ), https://tc39.es/ecma262/#sec-math.imul
 | ||
| ThrowCompletionOr<Value> MathObject::imul_impl(VM& vm, Value arg_a, Value arg_b)
 | ||
| {
 | ||
|     // 1. Let a be ℝ(? ToUint32(x)).
 | ||
|     auto const a = TRY(arg_a.to_u32(vm));
 | ||
| 
 | ||
|     // 2. Let b be ℝ(? ToUint32(y)).
 | ||
|     auto const b = TRY(arg_b.to_u32(vm));
 | ||
| 
 | ||
|     // 3. Let product be (a × b) modulo 2^32.
 | ||
|     // 4. If product ≥ 2^31, return 𝔽(product - 2^32); otherwise return 𝔽(product).
 | ||
|     return Value(static_cast<i32>(a * b));
 | ||
| }
 | ||
| 
 | ||
| // 21.3.2.20 Math.imul ( x, y ), https://tc39.es/ecma262/#sec-math.imul
 | ||
| JS_DEFINE_NATIVE_FUNCTION(MathObject::imul)
 | ||
| {
 | ||
|     return imul_impl(vm, vm.argument(0), vm.argument(1));
 | ||
| }
 | ||
| 
 | ||
| // 21.3.2.21 Math.log ( x ), https://tc39.es/ecma262/#sec-math.log
 | ||
| ThrowCompletionOr<Value> MathObject::log_impl(VM& vm, Value x)
 | ||
| {
 | ||
|     // 1. Let n be ? ToNumber(x).
 | ||
|     auto number = TRY(x.to_number(vm));
 | ||
| 
 | ||
|     // 2. If n is NaN or n is +∞𝔽, return n.
 | ||
|     if (number.is_nan() || number.is_positive_infinity())
 | ||
|         return number;
 | ||
| 
 | ||
|     // 3. If n is 1𝔽, return +0𝔽.
 | ||
|     if (number.as_double() == 1.)
 | ||
|         return Value(0);
 | ||
| 
 | ||
|     // 4. If n is +0𝔽 or n is -0𝔽, return -∞𝔽.
 | ||
|     if (number.is_positive_zero() || number.is_negative_zero())
 | ||
|         return js_negative_infinity();
 | ||
| 
 | ||
|     // 5. If n < -0𝔽, return NaN.
 | ||
|     if (number.as_double() < -0.)
 | ||
|         return js_nan();
 | ||
| 
 | ||
|     // 6. Return an implementation-approximated Number value representing the result of the natural logarithm of ℝ(n).
 | ||
|     return Value(::log(number.as_double()));
 | ||
| }
 | ||
| 
 | ||
| // 21.3.2.21 Math.log ( x ), https://tc39.es/ecma262/#sec-math.log
 | ||
| JS_DEFINE_NATIVE_FUNCTION(MathObject::log)
 | ||
| {
 | ||
|     return log_impl(vm, vm.argument(0));
 | ||
| }
 | ||
| 
 | ||
| // 21.3.2.22 Math.log1p ( x ), https://tc39.es/ecma262/#sec-math.log1p
 | ||
| JS_DEFINE_NATIVE_FUNCTION(MathObject::log1p)
 | ||
| {
 | ||
|     // 1. Let n be ? ToNumber(x).
 | ||
|     auto number = TRY(vm.argument(0).to_number(vm));
 | ||
| 
 | ||
|     // 2. If n is NaN, n is +0𝔽, n is -0𝔽, or n is +∞𝔽, return n.
 | ||
|     if (number.is_nan() || number.is_positive_zero() || number.is_negative_zero() || number.is_positive_infinity())
 | ||
|         return number;
 | ||
| 
 | ||
|     // 3. If n is -1𝔽, return -∞𝔽.
 | ||
|     if (number.as_double() == -1.)
 | ||
|         return js_negative_infinity();
 | ||
| 
 | ||
|     // 4. If n < -1𝔽, return NaN.
 | ||
|     if (number.as_double() < -1.)
 | ||
|         return js_nan();
 | ||
| 
 | ||
|     // 5. Return an implementation-approximated Number value representing the result of the natural logarithm of 1 + ℝ(n).
 | ||
|     return Value(::log1p(number.as_double()));
 | ||
| }
 | ||
| 
 | ||
| // 21.3.2.23 Math.log10 ( x ), https://tc39.es/ecma262/#sec-math.log10
 | ||
| JS_DEFINE_NATIVE_FUNCTION(MathObject::log10)
 | ||
| {
 | ||
|     // 1. Let n be ? ToNumber(x).
 | ||
|     auto number = TRY(vm.argument(0).to_number(vm));
 | ||
| 
 | ||
|     // 2. If n is NaN or n is +∞𝔽, return n.
 | ||
|     if (number.is_nan() || number.is_positive_infinity())
 | ||
|         return number;
 | ||
| 
 | ||
|     // 3. If n is 1𝔽, return +0𝔽.
 | ||
|     if (number.as_double() == 1.)
 | ||
|         return Value(0);
 | ||
| 
 | ||
|     // 4. If n is +0𝔽 or n is -0𝔽, return -∞𝔽.
 | ||
|     if (number.is_positive_zero() || number.is_negative_zero())
 | ||
|         return js_negative_infinity();
 | ||
| 
 | ||
|     // 5. If n < -0𝔽, return NaN.
 | ||
|     if (number.as_double() < -0.)
 | ||
|         return js_nan();
 | ||
| 
 | ||
|     // 6. Return an implementation-approximated Number value representing the result of the base 10 logarithm of ℝ(n).
 | ||
|     return Value(::log10(number.as_double()));
 | ||
| }
 | ||
| 
 | ||
| // 21.3.2.24 Math.log2 ( x ), https://tc39.es/ecma262/#sec-math.log2
 | ||
| JS_DEFINE_NATIVE_FUNCTION(MathObject::log2)
 | ||
| {
 | ||
|     // 1. Let n be ? ToNumber(x).
 | ||
|     auto number = TRY(vm.argument(0).to_number(vm));
 | ||
| 
 | ||
|     // 2. If n is NaN or n is +∞𝔽, return n.
 | ||
|     if (number.is_nan() || number.is_positive_infinity())
 | ||
|         return number;
 | ||
| 
 | ||
|     // 3. If n is 1𝔽, return +0𝔽.
 | ||
|     if (number.as_double() == 1.)
 | ||
|         return Value(0);
 | ||
| 
 | ||
|     // 4. If n is +0𝔽 or n is -0𝔽, return -∞𝔽.
 | ||
|     if (number.is_positive_zero() || number.is_negative_zero())
 | ||
|         return js_negative_infinity();
 | ||
| 
 | ||
|     // 5. If n < -0𝔽, return NaN.
 | ||
|     if (number.as_double() < -0.)
 | ||
|         return js_nan();
 | ||
| 
 | ||
|     // 6. Return an implementation-approximated Number value representing the result of the base 2 logarithm of ℝ(n).
 | ||
|     return Value(::log2(number.as_double()));
 | ||
| }
 | ||
| 
 | ||
| // 21.3.2.25 Math.max ( ...args ), https://tc39.es/ecma262/#sec-math.max
 | ||
| JS_DEFINE_NATIVE_FUNCTION(MathObject::max)
 | ||
| {
 | ||
|     // 1. Let coerced be a new empty List.
 | ||
|     Vector<Value> coerced;
 | ||
| 
 | ||
|     // 2. For each element arg of args, do
 | ||
|     for (size_t i = 0; i < vm.argument_count(); ++i) {
 | ||
|         // a. Let n be ? ToNumber(arg).
 | ||
|         auto number = TRY(vm.argument(i).to_number(vm));
 | ||
| 
 | ||
|         // b. Append n to coerced.
 | ||
|         coerced.append(number);
 | ||
|     }
 | ||
| 
 | ||
|     // 3. Let highest be -∞𝔽.
 | ||
|     auto highest = js_negative_infinity();
 | ||
| 
 | ||
|     // 4. For each element number of coerced, do
 | ||
|     for (auto& number : coerced) {
 | ||
|         // a. If number is NaN, return NaN.
 | ||
|         if (number.is_nan())
 | ||
|             return js_nan();
 | ||
| 
 | ||
|         // b. If number is +0𝔽 and highest is -0𝔽, set highest to +0𝔽.
 | ||
|         // c. If number > highest, set highest to number.
 | ||
|         if ((number.is_positive_zero() && highest.is_negative_zero()) || number.as_double() > highest.as_double())
 | ||
|             highest = number;
 | ||
|     }
 | ||
| 
 | ||
|     // 5. Return highest.
 | ||
|     return highest;
 | ||
| }
 | ||
| 
 | ||
| // 21.3.2.26 Math.min ( ...args ), https://tc39.es/ecma262/#sec-math.min
 | ||
| JS_DEFINE_NATIVE_FUNCTION(MathObject::min)
 | ||
| {
 | ||
|     // 1. Let coerced be a new empty List.
 | ||
|     Vector<Value> coerced;
 | ||
| 
 | ||
|     // 2. For each element arg of args, do
 | ||
|     for (size_t i = 0; i < vm.argument_count(); ++i) {
 | ||
|         // a. Let n be ? ToNumber(arg).
 | ||
|         auto number = TRY(vm.argument(i).to_number(vm));
 | ||
| 
 | ||
|         // b. Append n to coerced.
 | ||
|         coerced.append(number);
 | ||
|     }
 | ||
| 
 | ||
|     // 3. Let lowest be +∞𝔽.
 | ||
|     auto lowest = js_infinity();
 | ||
| 
 | ||
|     // 4. For each element number of coerced, do
 | ||
|     for (auto& number : coerced) {
 | ||
|         // a. If number is NaN, return NaN.
 | ||
|         if (number.is_nan())
 | ||
|             return js_nan();
 | ||
| 
 | ||
|         // b. If number is -0𝔽 and lowest is +0𝔽, set lowest to -0𝔽.
 | ||
|         // c. If number < lowest, set lowest to number.
 | ||
|         if ((number.is_negative_zero() && lowest.is_positive_zero()) || number.as_double() < lowest.as_double())
 | ||
|             lowest = number;
 | ||
|     }
 | ||
| 
 | ||
|     // 5. Return lowest.
 | ||
|     return lowest;
 | ||
| }
 | ||
| 
 | ||
| // 21.3.2.27 Math.pow ( base, exponent ), https://tc39.es/ecma262/#sec-math.pow
 | ||
| ThrowCompletionOr<Value> MathObject::pow_impl(VM& vm, Value base, Value exponent)
 | ||
| {
 | ||
|     // Set base to ? ToNumber(base).
 | ||
|     base = TRY(base.to_number(vm));
 | ||
| 
 | ||
|     // 2. Set exponent to ? ToNumber(exponent).
 | ||
|     exponent = TRY(exponent.to_number(vm));
 | ||
| 
 | ||
|     // 3. Return Number::exponentiate(base, exponent).
 | ||
|     return JS::exp(vm, base, exponent);
 | ||
| }
 | ||
| 
 | ||
| // 21.3.2.27 Math.pow ( base, exponent ), https://tc39.es/ecma262/#sec-math.pow
 | ||
| JS_DEFINE_NATIVE_FUNCTION(MathObject::pow)
 | ||
| {
 | ||
|     return pow_impl(vm, vm.argument(0), vm.argument(1));
 | ||
| }
 | ||
| 
 | ||
| class XorShift128PlusPlusRNG {
 | ||
| public:
 | ||
|     XorShift128PlusPlusRNG()
 | ||
|     {
 | ||
|         u64 seed = get_random<u32>();
 | ||
|         m_low = splitmix64(seed);
 | ||
|         m_high = splitmix64(seed);
 | ||
|     }
 | ||
| 
 | ||
|     double get()
 | ||
|     {
 | ||
|         u64 value = advance() & ((1ULL << 53) - 1);
 | ||
|         return value * (1.0 / (1ULL << 53));
 | ||
|     }
 | ||
| 
 | ||
| private:
 | ||
|     u64 splitmix64(u64& state)
 | ||
|     {
 | ||
|         u64 z = (state += 0x9e3779b97f4a7c15ULL);
 | ||
|         z = (z ^ (z >> 30)) * 0xbf58476d1ce4e5b9ULL;
 | ||
|         z = (z ^ (z >> 27)) * 0x94d049bb133111ebULL;
 | ||
|         return z ^ (z >> 31);
 | ||
|     }
 | ||
| 
 | ||
|     u64 advance()
 | ||
|     {
 | ||
|         u64 s1 = m_low;
 | ||
|         u64 const s0 = m_high;
 | ||
|         u64 const result = s0 + s1;
 | ||
|         m_low = s0;
 | ||
|         s1 ^= s1 << 23;
 | ||
|         s1 ^= s1 >> 17;
 | ||
|         s1 ^= s0 ^ (s0 >> 26);
 | ||
|         m_high = s1;
 | ||
|         return result + s1;
 | ||
|     }
 | ||
| 
 | ||
|     u64 m_low { 0 };
 | ||
|     u64 m_high { 0 };
 | ||
| };
 | ||
| 
 | ||
| Value MathObject::random_impl()
 | ||
| {
 | ||
|     // This function returns a Number value with positive sign, greater than or equal to +0𝔽 but strictly less than 1𝔽,
 | ||
|     // chosen randomly or pseudo randomly with approximately uniform distribution over that range, using an
 | ||
|     // implementation-defined algorithm or strategy.
 | ||
|     static XorShift128PlusPlusRNG rng;
 | ||
|     return Value(rng.get());
 | ||
| }
 | ||
| 
 | ||
| // 21.3.2.28 Math.random ( ), https://tc39.es/ecma262/#sec-math.random
 | ||
| JS_DEFINE_NATIVE_FUNCTION(MathObject::random)
 | ||
| {
 | ||
|     return random_impl();
 | ||
| }
 | ||
| 
 | ||
| // 21.3.2.29 Math.round ( x ), https://tc39.es/ecma262/#sec-math.round
 | ||
| ThrowCompletionOr<Value> MathObject::round_impl(VM& vm, Value x)
 | ||
| {
 | ||
|     // 1. Let n be ? ToNumber(x).
 | ||
|     auto number = TRY(x.to_number(vm));
 | ||
| 
 | ||
|     // 2. If n is not finite or n is an integral Number, return n.
 | ||
|     if (!number.is_finite_number() || number.as_double() == ::trunc(number.as_double()))
 | ||
|         return number;
 | ||
| 
 | ||
|     // 3. If n < 0.5𝔽 and n > +0𝔽, return +0𝔽.
 | ||
|     // 4. If n < -0𝔽 and n ≥ -0.5𝔽, return -0𝔽.
 | ||
|     // 5. Return the integral Number closest to n, preferring the Number closer to +∞ in the case of a tie.
 | ||
|     double integer = ::ceil(number.as_double());
 | ||
|     if (integer - 0.5 > number.as_double())
 | ||
|         integer--;
 | ||
|     return Value(integer);
 | ||
| }
 | ||
| 
 | ||
| // 21.3.2.29 Math.round ( x ), https://tc39.es/ecma262/#sec-math.round
 | ||
| JS_DEFINE_NATIVE_FUNCTION(MathObject::round)
 | ||
| {
 | ||
|     return round_impl(vm, vm.argument(0));
 | ||
| }
 | ||
| 
 | ||
| // 21.3.2.30 Math.sign ( x ), https://tc39.es/ecma262/#sec-math.sign
 | ||
| JS_DEFINE_NATIVE_FUNCTION(MathObject::sign)
 | ||
| {
 | ||
|     // 1. Let n be ? ToNumber(x).
 | ||
|     auto number = TRY(vm.argument(0).to_number(vm));
 | ||
| 
 | ||
|     // 2. If n is one of NaN, +0𝔽, or -0𝔽, return n.
 | ||
|     if (number.is_nan() || number.as_double() == 0)
 | ||
|         return number;
 | ||
| 
 | ||
|     // 3. If n < -0𝔽, return -1𝔽.
 | ||
|     if (number.as_double() < 0)
 | ||
|         return Value(-1);
 | ||
| 
 | ||
|     // 4. Return 1𝔽.
 | ||
|     return Value(1);
 | ||
| }
 | ||
| 
 | ||
| // 21.3.2.31 Math.sin ( x ), https://tc39.es/ecma262/#sec-math.sin
 | ||
| JS_DEFINE_NATIVE_FUNCTION(MathObject::sin)
 | ||
| {
 | ||
|     // 1. Let n be ? ToNumber(x).
 | ||
|     auto number = TRY(vm.argument(0).to_number(vm));
 | ||
| 
 | ||
|     // 2. If n is NaN, n is +0𝔽, or n is -0𝔽, return n.
 | ||
|     if (number.is_nan() || number.is_positive_zero() || number.is_negative_zero())
 | ||
|         return number;
 | ||
| 
 | ||
|     // 3. If n is +∞𝔽 or n is -∞𝔽, return NaN.
 | ||
|     if (number.is_infinity())
 | ||
|         return js_nan();
 | ||
| 
 | ||
|     // 4. Return an implementation-approximated Number value representing the result of the sine of ℝ(n).
 | ||
|     return Value(::sin(number.as_double()));
 | ||
| }
 | ||
| 
 | ||
| // 21.3.2.32 Math.sinh ( x ), https://tc39.es/ecma262/#sec-math.sinh
 | ||
| JS_DEFINE_NATIVE_FUNCTION(MathObject::sinh)
 | ||
| {
 | ||
|     // 1. Let n be ? ToNumber(x).
 | ||
|     auto number = TRY(vm.argument(0).to_number(vm));
 | ||
| 
 | ||
|     // 2. If n is not finite or n is either +0𝔽 or -0𝔽, return n.
 | ||
|     if (!number.is_finite_number() || number.is_positive_zero() || number.is_negative_zero())
 | ||
|         return number;
 | ||
| 
 | ||
|     // 3. Return an implementation-approximated Number value representing the result of the hyperbolic sine of ℝ(n).
 | ||
|     return Value(::sinh(number.as_double()));
 | ||
| }
 | ||
| 
 | ||
| // 21.3.2.33 Math.sqrt ( x ), https://tc39.es/ecma262/#sec-math.sqrt
 | ||
| ThrowCompletionOr<Value> MathObject::sqrt_impl(VM& vm, Value x)
 | ||
| {
 | ||
|     // Let n be ? ToNumber(x).
 | ||
|     auto number = TRY(x.to_number(vm));
 | ||
| 
 | ||
|     // 2. If n is one of NaN, +0𝔽, -0𝔽, or +∞𝔽, return n.
 | ||
|     if (number.is_nan() || number.as_double() == 0 || number.is_positive_infinity())
 | ||
|         return number;
 | ||
| 
 | ||
|     // 3. If n < -0𝔽, return NaN.
 | ||
|     if (number.as_double() < 0)
 | ||
|         return js_nan();
 | ||
| 
 | ||
|     // 4. Return an implementation-approximated Number value representing the result of the square root of ℝ(n).
 | ||
|     return Value(::sqrt(number.as_double()));
 | ||
| }
 | ||
| 
 | ||
| // 21.3.2.33 Math.sqrt ( x ), https://tc39.es/ecma262/#sec-math.sqrt
 | ||
| JS_DEFINE_NATIVE_FUNCTION(MathObject::sqrt)
 | ||
| {
 | ||
|     return sqrt_impl(vm, vm.argument(0));
 | ||
| }
 | ||
| 
 | ||
| // 21.3.2.34 Math.tan ( x ), https://tc39.es/ecma262/#sec-math.tan
 | ||
| JS_DEFINE_NATIVE_FUNCTION(MathObject::tan)
 | ||
| {
 | ||
|     // Let n be ? ToNumber(x).
 | ||
|     auto number = TRY(vm.argument(0).to_number(vm));
 | ||
| 
 | ||
|     // 2. If n is NaN, n is +0𝔽, or n is -0𝔽, return n.
 | ||
|     if (number.is_nan() || number.is_positive_zero() || number.is_negative_zero())
 | ||
|         return number;
 | ||
| 
 | ||
|     // 3. If n is +∞𝔽, or n is -∞𝔽, return NaN.
 | ||
|     if (number.is_infinity())
 | ||
|         return js_nan();
 | ||
| 
 | ||
|     // 4. Return an implementation-approximated Number value representing the result of the tangent of ℝ(n).
 | ||
|     return Value(::tan(number.as_double()));
 | ||
| }
 | ||
| 
 | ||
| // 21.3.2.35 Math.tanh ( x ), https://tc39.es/ecma262/#sec-math.tanh
 | ||
| JS_DEFINE_NATIVE_FUNCTION(MathObject::tanh)
 | ||
| {
 | ||
|     // 1. Let n be ? ToNumber(x).
 | ||
|     auto number = TRY(vm.argument(0).to_number(vm));
 | ||
| 
 | ||
|     // 2. If n is NaN, n is +0𝔽, or n is -0𝔽, return n.
 | ||
|     if (number.is_nan() || number.is_positive_zero() || number.is_negative_zero())
 | ||
|         return number;
 | ||
| 
 | ||
|     // 3. If n is +∞𝔽, return 1𝔽.
 | ||
|     if (number.is_positive_infinity())
 | ||
|         return Value(1);
 | ||
| 
 | ||
|     // 4. If n is -∞𝔽, return -1𝔽.
 | ||
|     if (number.is_negative_infinity())
 | ||
|         return Value(-1);
 | ||
| 
 | ||
|     // 5. Return an implementation-approximated Number value representing the result of the hyperbolic tangent of ℝ(n).
 | ||
|     return Value(::tanh(number.as_double()));
 | ||
| }
 | ||
| 
 | ||
| // 21.3.2.36 Math.trunc ( x ), https://tc39.es/ecma262/#sec-math.trunc
 | ||
| JS_DEFINE_NATIVE_FUNCTION(MathObject::trunc)
 | ||
| {
 | ||
|     // 1. Let n be ? ToNumber(x).
 | ||
|     auto number = TRY(vm.argument(0).to_number(vm));
 | ||
| 
 | ||
|     // 2. If n is not finite or n is either +0𝔽 or -0𝔽, return n.
 | ||
|     if (number.is_nan() || number.is_infinity() || number.as_double() == 0)
 | ||
|         return number;
 | ||
| 
 | ||
|     // 3. If n < 1𝔽 and n > +0𝔽, return +0𝔽.
 | ||
|     // 4. If n < -0𝔽 and n > -1𝔽, return -0𝔽.
 | ||
|     // 5. Return the integral Number nearest n in the direction of +0𝔽.
 | ||
|     return Value(number.as_double() < 0
 | ||
|             ? ::ceil(number.as_double())
 | ||
|             : ::floor(number.as_double()));
 | ||
| }
 | ||
| 
 | ||
| struct TwoSumResult {
 | ||
|     double hi;
 | ||
|     double lo;
 | ||
| };
 | ||
| 
 | ||
| static TwoSumResult two_sum(double x, double y)
 | ||
| {
 | ||
|     double hi = x + y;
 | ||
|     double lo = y - (hi - x);
 | ||
|     return { hi, lo };
 | ||
| }
 | ||
| 
 | ||
| // 2 Math.sumPrecise ( items ), https://tc39.es/proposal-math-sum/#sec-math.sumprecise
 | ||
| JS_DEFINE_NATIVE_FUNCTION(MathObject::sumPrecise)
 | ||
| {
 | ||
|     static constexpr double MAX_DOUBLE = 1.79769313486231570815e+308;         // std::numeric_limits<double>::max()
 | ||
|     static constexpr double PENULTIMATE_DOUBLE = 1.79769313486231550856e+308; // std::nextafter(DBL_MAX, 0)
 | ||
|     static constexpr double MAX_ULP = MAX_DOUBLE - PENULTIMATE_DOUBLE;
 | ||
|     static constexpr double POW_2_1023 = 8.98846567431158e+307; // 2^1023
 | ||
| 
 | ||
|     enum class State {
 | ||
|         MinusZero,
 | ||
|         PlusInfinity,
 | ||
|         MinusInfinity,
 | ||
|         NotANumber,
 | ||
|         Finite
 | ||
|     };
 | ||
| 
 | ||
|     auto items = vm.argument(0);
 | ||
| 
 | ||
|     // 1. Perform ? RequireObjectCoercible(items).
 | ||
|     TRY(require_object_coercible(vm, items));
 | ||
| 
 | ||
|     // 2. Let iteratorRecord be ? GetIterator(items, SYNC).
 | ||
|     auto iterator_record = TRY(get_iterator(vm, items, IteratorHint::Sync));
 | ||
| 
 | ||
|     // 3. Let state be MINUS-ZERO.
 | ||
|     State state = State::MinusZero;
 | ||
| 
 | ||
|     // 4. Let sum be 0.
 | ||
|     // 5. Let count be 0.
 | ||
|     double overflow = 0.0;
 | ||
|     u64 count = 0;
 | ||
|     Vector<double> partials;
 | ||
| 
 | ||
|     // 6. Let next be NOT-STARTED.
 | ||
|     // 7. Repeat, while next is not DONE
 | ||
|     for (;;) {
 | ||
|         // a. Set next to ? IteratorStepValue(iteratorRecord).
 | ||
|         auto next = TRY(iterator_step_value(vm, iterator_record));
 | ||
|         if (!next.has_value())
 | ||
|             break;
 | ||
| 
 | ||
|         auto next_value = next.value();
 | ||
| 
 | ||
|         // b. If next is not DONE, then
 | ||
|         // i. Set count to count + 1.
 | ||
|         ++count;
 | ||
| 
 | ||
|         // ii. If count ≥ 2**53, then
 | ||
|         if (count >= (1ULL << 53)) {
 | ||
|             // 1. Let error be ThrowCompletion(a newly created RangeError object).
 | ||
|             auto error = vm.throw_completion<RangeError>(ErrorType::ArrayMaxSize);
 | ||
| 
 | ||
|             // 2. Return ? IteratorClose(iteratorRecord, error).
 | ||
|             return iterator_close(vm, iterator_record, error);
 | ||
|         }
 | ||
| 
 | ||
|         // iii. NOTE: The above case is not expected to be reached in practice and is included only so that implementations
 | ||
|         //      may rely on inputs being "reasonably sized" without violating this specification.
 | ||
| 
 | ||
|         // iv. If next is not a Number, then
 | ||
|         if (!next_value.is_number()) {
 | ||
|             // 1. Let error be ThrowCompletion(a newly created TypeError object).
 | ||
|             auto error = vm.throw_completion<TypeError>(ErrorType::IsNotA, next_value, "number");
 | ||
| 
 | ||
|             // 2. Return ? IteratorClose(iteratorRecord, error).
 | ||
|             return iterator_close(vm, iterator_record, error);
 | ||
|         }
 | ||
| 
 | ||
|         // v. Let n be next.
 | ||
|         auto n = next_value.as_double();
 | ||
| 
 | ||
|         // vi. If state is not NOT-A-NUMBER, then
 | ||
|         if (state != State::NotANumber) {
 | ||
|             // 1. If n is NaN, then
 | ||
|             if (next_value.is_nan()) {
 | ||
|                 // a. Set state to NOT-A-NUMBER.
 | ||
|                 state = State::NotANumber;
 | ||
|             }
 | ||
|             // 2. Else if n is +∞𝔽, then
 | ||
|             else if (next_value.is_positive_infinity()) {
 | ||
|                 // a. If state is MINUS-INFINITY, set state to NOT-A-NUMBER.
 | ||
|                 // b. Else, set state to PLUS-INFINITY.
 | ||
|                 state = state == State::MinusInfinity ? State::NotANumber : State::PlusInfinity;
 | ||
|             }
 | ||
|             // 3. Else if n is -∞𝔽, then
 | ||
|             else if (next_value.is_negative_infinity()) {
 | ||
|                 // a. If state is PLUS-INFINITY, set state to NOT-A-NUMBER.
 | ||
|                 // b. Else, set state to MINUS-INFINITY.
 | ||
|                 state = state == State::PlusInfinity ? State::NotANumber : State::MinusInfinity;
 | ||
|             }
 | ||
|             // 4. Else if n is not -0𝔽 and state is either MINUS-ZERO or FINITE, then
 | ||
|             else if (!next_value.is_negative_zero() && (state == State::MinusZero || state == State::Finite)) {
 | ||
|                 // a. Set state to FINITE.
 | ||
|                 state = State::Finite;
 | ||
| 
 | ||
|                 // b. Set sum to sum + ℝ(n).
 | ||
|                 double x = n;
 | ||
|                 size_t used_partials = 0;
 | ||
| 
 | ||
|                 for (size_t i = 0; i < partials.size(); i++) {
 | ||
|                     double y = partials[i];
 | ||
| 
 | ||
|                     if (AK::abs(x) < AK::abs(y))
 | ||
|                         swap(x, y);
 | ||
| 
 | ||
|                     TwoSumResult result = two_sum(x, y);
 | ||
|                     double hi = result.hi;
 | ||
|                     double lo = result.lo;
 | ||
| 
 | ||
|                     if (isinf(hi)) {
 | ||
|                         double sign = signbit(hi) ? -1.0 : 1.0;
 | ||
|                         overflow += sign;
 | ||
| 
 | ||
|                         if (AK::abs(overflow) >= (1ULL << 53))
 | ||
|                             return vm.throw_completion<RangeError>(ErrorType::MathSumPreciseOverflow);
 | ||
| 
 | ||
|                         x = (x - sign * POW_2_1023) - sign * POW_2_1023;
 | ||
| 
 | ||
|                         if (AK::abs(x) < AK::abs(y))
 | ||
|                             swap(x, y);
 | ||
| 
 | ||
|                         result = two_sum(x, y);
 | ||
|                         hi = result.hi;
 | ||
|                         lo = result.lo;
 | ||
|                     }
 | ||
| 
 | ||
|                     if (lo != 0.0) {
 | ||
|                         partials[used_partials++] = lo;
 | ||
|                     }
 | ||
| 
 | ||
|                     x = hi;
 | ||
|                 }
 | ||
| 
 | ||
|                 partials.resize(used_partials);
 | ||
| 
 | ||
|                 if (x != 0.0) {
 | ||
|                     partials.append(x);
 | ||
|                 }
 | ||
|             }
 | ||
|         }
 | ||
|     }
 | ||
| 
 | ||
|     // 8. If state is NOT-A-NUMBER, return NaN.
 | ||
|     if (state == State::NotANumber)
 | ||
|         return js_nan();
 | ||
| 
 | ||
|     // 9. If state is PLUS-INFINITY, return +∞𝔽.
 | ||
|     if (state == State::PlusInfinity)
 | ||
|         return js_infinity();
 | ||
| 
 | ||
|     // 10. If state is MINUS-INFINITY, return -∞𝔽.
 | ||
|     if (state == State::MinusInfinity)
 | ||
|         return js_negative_infinity();
 | ||
| 
 | ||
|     // 11. If state is MINUS-ZERO, return -0𝔽.
 | ||
|     if (state == State::MinusZero)
 | ||
|         return Value(-0.0);
 | ||
| 
 | ||
|     // 12. Return 𝔽(sum).
 | ||
|     int n = partials.size() - 1;
 | ||
|     double hi = 0.0;
 | ||
|     double lo = 0.0;
 | ||
| 
 | ||
|     if (overflow != 0.0) {
 | ||
|         double next = n >= 0 ? partials[n] : 0.0;
 | ||
|         n--;
 | ||
| 
 | ||
|         if (AK::abs(overflow) > 1.0 || (overflow > 0.0 && next > 0.0) || (overflow < 0.0 && next < 0.0)) {
 | ||
|             return overflow > 0.0 ? js_infinity() : js_negative_infinity();
 | ||
|         }
 | ||
| 
 | ||
|         TwoSumResult result = two_sum(overflow * POW_2_1023, next / 2.0);
 | ||
|         hi = result.hi;
 | ||
|         lo = result.lo * 2.0;
 | ||
| 
 | ||
|         if (isinf(hi * 2.0)) {
 | ||
|             if (hi > 0.0) {
 | ||
|                 if (hi == POW_2_1023 && lo == -(MAX_ULP / 2.0) && n >= 0 && partials[n] < 0.0) {
 | ||
|                     return Value(MAX_DOUBLE);
 | ||
|                 }
 | ||
| 
 | ||
|                 return js_infinity();
 | ||
|             } else {
 | ||
|                 if (hi == -POW_2_1023 && lo == (MAX_ULP / 2.0) && n >= 0 && partials[n] > 0.0) {
 | ||
|                     return Value(-MAX_DOUBLE);
 | ||
|                 }
 | ||
| 
 | ||
|                 return js_negative_infinity();
 | ||
|             }
 | ||
|         }
 | ||
| 
 | ||
|         if (lo != 0.0) {
 | ||
|             partials[n + 1] = lo;
 | ||
|             n++;
 | ||
|             lo = 0.0;
 | ||
|         }
 | ||
| 
 | ||
|         hi *= 2.0;
 | ||
|     }
 | ||
| 
 | ||
|     while (n >= 0) {
 | ||
|         double x = hi;
 | ||
|         double y = partials[n];
 | ||
|         n--;
 | ||
| 
 | ||
|         TwoSumResult result = two_sum(x, y);
 | ||
|         hi = result.hi;
 | ||
|         lo = result.lo;
 | ||
| 
 | ||
|         if (lo != 0.0) {
 | ||
|             break;
 | ||
|         }
 | ||
|     }
 | ||
| 
 | ||
|     if (n >= 0 && ((lo < 0.0 && partials[n] < 0.0) || (lo > 0.0 && partials[n] > 0.0))) {
 | ||
|         double y = lo * 2.0;
 | ||
|         double x = hi + y;
 | ||
|         double yr = x - hi;
 | ||
| 
 | ||
|         if (y == yr) {
 | ||
|             hi = x;
 | ||
|         }
 | ||
|     }
 | ||
| 
 | ||
|     return hi;
 | ||
| }
 | ||
| 
 | ||
| }
 | 
