[llvm-bugs] [Bug 43996] New: bad code produced... sometimes??

[via llvm-bugs]

https://bugs.llvm.org/show_bug.cgi?id=43996

            Bug ID: 43996
           Summary: bad code produced... sometimes??
           Product: clang
           Version: 9.0
          Hardware: PC
                OS: Linux
            Status: NEW
          Severity: enhancement
          Priority: P
         Component: C++
          Assignee: unassignedclangbugs at nondot.org
          Reporter: stsp2 at yandex.ru
                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 22810
  --> https://bugs.llvm.org/attachment.cgi?id=22810&action=edit
test case, you must see it to believe

Well, this is probably the weirdest bug I've
ever seen. The attached is a fully automated
test-case (without which no one would even
try it out, as people will likely think I am
crazy).

The thing is, if you compile this test-case
_multiple times_, sometimes you get the bad
binary that will crash!
No no, read on, I have a proof! :)
The script that I created, will run the compilation
in a loop. It will also record the generated
asm from every compilation cycle. As soon as
the bad build is detected, the script will
output the asm diff between good and bad
builds, and will stop execution. If the problem
happens on the first loop, then no diff is printed
for obvious reasons, and you'll need to restart
the test to get the diff output.

And yes, I tested this on multiple machines.

-- 
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/20191113/0cce8c7e/attachment.html>