Even though this is a waste of memory it seems to trip on some UAF with
some games.

Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=54475

Chr() depends on the thread's ANSI code page which in some cases is different
from the Windows ANSI code page (e.g. hi-IN).

Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=54003

The first parameter is defined as const BLUETOOTH_DEVICE_INFO in MSDN and bluetoothapis.h.

Signed-off-by: Connor McAdams <cmcadams@codeweavers.com>

Signed-off-by: Connor McAdams <cmcadams@codeweavers.com>

Signed-off-by: Connor McAdams <cmcadams@codeweavers.com>

Signed-off-by: Connor McAdams <cmcadams@codeweavers.com>

As background information from Windows testings:
- the debug port event is closed upon thread regular exit.
- results are however inconsistent across Windows versions for forced
  termination (through TerminateThread and equiv).
- this is moreover crippled by most of the 32bit implementation on Windows
  (except Win8) which don't enforce the close on exit flag for wow64 debuggee:s
  (they enforce it for 64bit though)
- but, in the cases where the debuggee terminates, the debug port is
  always closed *after* the debuggee process termination.

So I adapted the tests to conform to "debug port is closed" when handling
the thread termination. Some tests had to be moved *after* debuggee
termination to ensure success.

Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=53144Signed-off-by: Eric Pouech <eric.pouech@gmail.com>