https://bugs.winehq.org/show_bug.cgi?id=54240
--- Comment #3 from Sidney Kelley memory.thrasher@gmail.com --- (In reply to Zeb Figura from comment #1)
Wineserver should never crash, even if wine programs crash, meaning there's a separate bug here. Can you get a backtrace for the server crash?
Unfortunately no, I was never able to get much of a backtrace. I've attached the entire output of the combined wineserver and app. The line ending in "aborting" is what prompted me to implement the rpc stub.
I launches them both like so `wineserver -f -p60 &>vanillawine.err & wine MaLauncher.exe &>>vanillawine.err` to generate the attached log, and a few seconds later, bash reports "done" on the wineserver job ('job' in the bash sense of backgrounded process), meaning it closed before it's 60 second idle time, which is the only reason I say it crashed. It's possible the bug is with the handling of the 60 second persistence flag, but my money is on the "internal crash" of winedbg breaking an assumption/contract with wineserver. Note that I'm mostly an outsider with only a vague idea how these processes fit together.