On Mon, 2 Mar 2020, Marvin wrote: [...]
=== w1064v1507 (32 bit report) ===
Report validation errors: kernel32:process is missing some failure messages
That's one of the occasional timeouts that was identified before, only in this case it raced with the child process so that the test failure message was garbled and unrecognizable.
At some point I'll patch this test with timing information and longer timeouts to see just how long some of these waits are (all I know for now is that raising them to 2s is not enough).