https://bugs.winehq.org/show_bug.cgi?id=50955
--- Comment #5 from florian.will@gmail.com --- Created attachment 71901 --> https://bugs.winehq.org/attachment.cgi?id=71901 Patch that "fixes" this issue
This is one way to "fix" this issue, tested to work in ZusiDisplay, but it feels like a terrible hack. It also introduces some test suite failures like "bind succeeded unexpectedly" and EACCESS instead of EADDRINUSE error codes (that last one could probably be fixed).
Bernhard mentioned another way to approach this, setting the SO_LINGER timeout to 0 seconds when destroying the socket in wineserver (I haven't tested this).
Maybe there are more options?