https://bugs.kde.org/show_bug.cgi?id=491667
Bug ID: 491667
Summary: std::fmin(val, NaN) should returns NaN on arm.
Classification: Developer tools
Product: valgrind
Version: unspecified
Platform: Other
OS: Other
Status: REPORTED
Severity: normal
Priority: NOR
Component: memcheck
Assignee: jsew...@acm.org
Reporter: meirav.grimb...@redis.com
Target Milestone: ---
SUMMARY
fmin(NaN,-1) returns NaN on linux arm.
According to fmin documentation:
https://en.cppreference.com/w/cpp/numeric/math/fmin
Error handling
This function is not subject to any of the error conditions specified in
math_errhandling.
If the implementation supports IEEE floating-point arithmetic (IEC 60559),
* If one of the two arguments is NaN, the value of the other argument is
returned.
* Only if both arguments are NaN, NaN is returned.
STEPS TO REPRODUCE
program test.cpp
``
#include <cmath>
#include <iostream>
int main()
{
std::cout << "fmin(NaN,-1) = " << std::fmin(NAN, -1) << '\n';
}
``
g++ test.cpp
OBSERVED RESULT
valgrind ./a.out
==9822== Memcheck, a memory error detector
==9822== Copyright (C) 2002-2017, and GNU GPL'd, by Julian Seward et al.
==9822== Using Valgrind-3.18.1 and LibVEX; rerun with -h for copyright info
==9822== Command: ./a.out
==9822==
fmin(NaN,-1) = nan
EXPECTED RESULT
./a.out
fmin(NaN,-1) = -1
SOFTWARE/OS VERSIONS
Was tested on dockers.
arch: armV8
1. linux jammy. valgrind version 18.1
2. linux apline. valgrind version 23
ADDITIONAL INFORMATION
--
You are receiving this mail because:
You are watching all bug changes.
