https://bugs.winehq.org/show_bug.cgi?id=32515
--- Comment #51 from Christian Inci chris.pcguy.inci@gmail.com --- Created attachment 57690 --> https://bugs.winehq.org/attachment.cgi?id=57690 Testcase
There might be a problem in FindNextFile, which results in a discrepancy at the generated executable. I don't know if fixing it will resolve this bug.
I've reverse-engineered the part of CEG which is used when generating the executable. This executable is a nearly one-to-one copy of it and a testcase too. That's the closest I can get without removing some MingW parts. Which means that the output of the "algorithm" isn't a bitexact copy.
I can provide the source code on request.
Btw: Valve blacklisted the MachineGUID of one of my debugging VMs. It were simply evaded by just changing the MachineGUID registry value.