<xlocnum>: Incorrect rounding when parsing long hexadecimal floating point numbers just above midpoints #3375
Description
Describe the bug
The hexadecimal code path of num_get::do_get parses the initial 768 significant digits and discards the remaining. When the remaining digits could affect rounding, it returns an incorrect result.
Command-line test case
D:\Temp>type test.cpp
#include <iostream>
#include <sstream>
#include <string>
using namespace std;
int main() {
istringstream stream("0x1.00000000000008" + string(1000, '0') + "1p+0");
stream >> hexfloat; // GH-1259
double x;
if (stream >> x) {
cout << "actual: " << hexfloat << x << "\n";
} else {
cout << "actual: (failure)\n";
}
cout << "expected: 0x1.0000000000001p+0\n";
return 0;
}D:\Temp>cl /EHsc /W4 test.cpp
用于 x64 的 Microsoft (R) C/C++ 优化编译器 19.35.32213 版
版权所有(C) Microsoft Corporation。保留所有权利。
test.cpp
Microsoft (R) Incremental Linker Version 14.35.32213.0
Copyright (C) Microsoft Corporation. All rights reserved.
/out:test.exe
test.obj
D:\Temp>.\test
actual: 0x1.0000000000000p+0
expected: 0x1.0000000000001p+0
STL version
Microsoft Visual Studio Community 2022 (64 位) - Preview
版本 17.5.0 Preview 5.0