Hi!
The fail/panic/double_panic test stderr file contains a backtrace with many internal functions of std's panic and formatting code.
This means this test breaks every time I refactor something there. For example here and here.
I'm not exactly sure what the test is testing, but if it's not too much effort, it'd be nice if the test could be restricted to only test whatever it needed to test for miri, such that it doesn't fail if we change something in std::panicking or std::fmt.