[llvm-bugs] [Bug 48940] New: float to __int128 conversions completely broken on Mingw

via llvm-bugs llvm-bugs at lists.llvm.org
Fri Jan 29 03:11:26 PST 2021


            Bug ID: 48940
           Summary: float to __int128 conversions completely broken on
           Product: clang
           Version: 11.0
          Hardware: PC
                OS: Windows NT
            Status: NEW
          Severity: release blocker
          Priority: P
         Component: C++
          Assignee: unassignedclangbugs at nondot.org
          Reporter: john at johnmaddock.co.uk
                CC: blitzrakete at gmail.com, dgregor at apple.com,
                    erik.pilkington at gmail.com, llvm-bugs at lists.llvm.org,
                    richard-llvm at metafoo.co.uk

Converting a float (or indeed double) to an __int128 results in a garbage
result, test case:

#include <assert.h>
#include <type_traits>
#include <cmath>

template <class T>
struct wraps
   T val;
   template <class U>
   wraps(U val, typename
std::enable_if<std::is_floating_point<U>::value>::type* = 0)
   : val(static_cast<T>(std::fabs(val))){}

   T get()const { return val; }

int main()
   float f(2);
   wraps<unsigned __int128> w(f);

   assert(w.get() == 2);  // always fails for __int128

   return 0;

Tested compiler was:

$ clang++ --version
clang version 11.0.0 (https://github.com/msys2/MINGW-packages
Target: x86_64-w64-windows-gnu
Thread model: posix
InstalledDir: D:\compilers\msys64\mingw64\bin

This is a showstopper for Boost.Multiprecision on Mingw.  There are no issues
on Ubuntu Linux or with XCode clang that I can see.

You are receiving this mail because:
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-bugs/attachments/20210129/37c33176/attachment.html>

More information about the llvm-bugs mailing list