deadlock module is broken on windows

[Justin Schwartz]

As long as I can remember the deadlock module has been useless on Windows. Upon recent testing, the usleep function is a no-op as far as I can tell, which means as soon as you load the module, it cries deadlock. So the module has to be unloaded to proceed. An alternative for usleep needs to be implemented.

[Xifer]

Looks like Dr. Brain commented out the Windows implementation of usleep as a "Windows build fix": https://bitbucket.org/grelminar/asss/diff/src/include/win32compat.h?diff2=dd1378daa8c8&at=default

Uncommenting it seems fix the abort on start... I'm curious why it was commented out though. Could still be a build fix depending on the build environment on Windows (Cygwin, M$). This definitely needs to be uncommented for me using MinGW though.

[Joris]

This issue no longer occurs in 1.6.0 (msvc)