https://bugs.winehq.org/show_bug.cgi?id=52401
--- Comment #2 from Jinoh Kang jinoh.kang.kr@gmail.com --- I realised that more needs to be done to the solution #1, if the socket is in fact synchronous.
Also, testing for this bug seems quite complicated--we need to block kernel APCs until after try_recv() has been run. Can we, well, just block SIGUSR1 temporarily in winetest? Sounds like something we wouldn't do in code freeze.