[Bug 58234] New: llvm-mingw build makes ddraw:ddraw4 test crash because of a ceilf recursion.

[WineHQ Bugzilla]

http://bugs.winehq.org/show_bug.cgi?id=58234

            Bug ID: 58234
           Summary: llvm-mingw build makes ddraw:ddraw4 test crash because
                    of a ceilf recursion.
           Product: Wine
           Version: 10.7
          Hardware: x86-64
                OS: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: -unknown
          Assignee: wine-bugs(a)winehq.org
          Reporter: bernhardu(a)mailbox.org
      Distribution: ---

Created attachment 78561
  --> http://bugs.winehq.org/attachment.cgi?id=78561
ddraw_test_llvm-ceilf-ceil-recursion.txt

Hello, trying to look at a failure of test ddraw:ddraw4
I just found an issue with my build.

The issue is a recursive call to ceilf.

This affects just the 32-bit build, where it looks like ceilf is just
an inline wrapper in the header file for the function ceil.

But for some reason llvm-mingw/clang replaces the call to ceil by ceilf,
therefore causing the recursion.

Adding `EXTRADEFS = -fno-builtin` to dlls/ddraw/tests/Makefile.in
fixes this issue for here.

Or is this caused by me using the `{i686,186_64}-w64-mingw32-gcc`
wrappers instead of the clang binary?


Can someone else confirm this?
Is this an issue that should be fixed in wine,
or is there something wrong in llvm-mingw/llvm-project,
or I am just using an unsupported way to configure?

-- 
Do not reply to this email, post in Bugzilla using the
above URL to reply.
You are receiving this mail because:
You are watching all bug changes.

[Bug 58234] llvm-mingw build makes ddraw:ddraw4 test crash because of a ceilf recursion.

[WineHQ Bugzilla]

http://bugs.winehq.org/show_bug.cgi?id=58234

--- Comment #1 from Bernhard Übelacker <bernhardu(a)mailbox.org> ---
Currently I was using a line like this:
./configure \
        x86_64_CC='ccache x86_64-w64-mingw32-gcc' \
          i386_CC='ccache i686-w64-mingw32-gcc' \
...

I tested now to use this:
./configure \
        x86_64_CC='ccache clang' \
          i386_CC='ccache clang' \
...

And indeed I cannot see this issue in this build,
so I assume I was suppoed to use clang?

(A few months ago I think to remember there was
some other issue blocking with this way,
therefore I used the wrappers, but today it works like a charm.)

-- 
Do not reply to this email, post in Bugzilla using the
above URL to reply.
You are receiving this mail because:
You are watching all bug changes.