[PATCH 0/6] A couple of workarounds for warnings emitted by GCC11

[Eric Pouech]

GCC11 implements new warnings. This serie of patches:
- works around a couple of false positive
- fixes a couple of bad handling of corner cases in error handling
- last patch should be reviewed by D3D folks. The fix is consistent with
  function definition, and no regression appear in make test...

A+
---

Eric Pouech (6):
      dlls/msvcrt*: ensure variable sse2_cw is set for all code paths in _control87 (GCC11)
      In tests, ensure buffers passed as const pointers are always initialized (GCC11)
      kernelbase: always return NULL in case of error in GetModuleHandleA (GCC11)
      kernel32/tests: mark pointer as volatile when reading after allocated size
      Workaround GCC11 complaining about sizeof(a)/sizeof(b) constructs when it's
      d3dx9_36/tests: fixed order of parameters


 dlls/d3dx9_36/mesh.c         | 2 +-
 dlls/d3dx9_36/tests/effect.c | 4 ++--
 dlls/d3dx9_36/tests/mesh.c   | 4 ++--
 dlls/kernel32/tests/heap.c   | 3 ++-
 dlls/kernelbase/loader.c     | 2 +-
 dlls/kernelbase/tests/path.c | 1 +
 dlls/msvcirt/tests/msvcirt.c | 5 +++--
 dlls/msvcrt/math.c           | 2 +-
 dlls/shlwapi/tests/clist.c   | 1 +
 9 files changed, 14 insertions(+), 10 deletions(-)