https://bugs.winehq.org/show_bug.cgi?id=53464
--- Comment #3 from Eric Pouech eric.pouech@gmail.com --- Well, w1064v1909 is so bogus regarding console tests, I'd avoid it (I'm even considering dropping it from conhost tests...) Also you noted a potential crash, that may explain the 88% ratio.
Tweaking the tests (*), I isolated two spots where the timeout (not the crash) happens: 1) the one you listed in read_control_result 2) another one in expect_input
Both boil down to string sent as input not being (fully) processed. This could be a race (as when pushing another string, the first appears being processed).
More investigation (and flooding on Marvin required...)
(*) by replacing synchronous read with asynchronous reads + timeout; I get the places where it times out. However, need to run heavily the tests to grasp this kind of event :-(