[llvm-bugs] [Bug 40641] New: Clang generates wrong optimized code for std::rotate followed by std::min_element with libc++

[via llvm-bugs]

https://bugs.llvm.org/show_bug.cgi?id=40641

            Bug ID: 40641
           Summary: Clang generates wrong optimized code for std::rotate
                    followed by std::min_element with libc++
           Product: clang
           Version: 7.0
          Hardware: PC
                OS: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
          Assignee: unassignedclangbugs at nondot.org
          Reporter: steffen.muething at iwr.uni-heidelberg.de
                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

Created attachment 21442
  --> https://bugs.llvm.org/attachment.cgi?id=21442&action=edit
Failing code, check return code to see if it worked

When compiling with clang 7 "-std=libc++ -O3", the attached program generates
wrong output on x86_64 (both Linux and macOS). In particular, the rotated
sequence contains one of the items twice, and one item has disappeared.

I did some testing with the following outcomes:
- The code only fails in optimized mode
- The code only fails when using libc++, when compiling with the GCC 8
libstdc++, the generated code is correct
- On the other hand, when compiling the code with GCC 8 and libc++, the
generated code is also correct
- The code only fails when the output of std::min_element is used as the middle
parameter of std:rotate

-- 
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/20190207/57dea69c/attachment.html>