#Rust famously runs unit tests in parallel. This is usually great, but I have a peculiar situation.
I have a test A, that checks if function A works correctly. I have several other tests B, C, …, that check if functions B, C, … work correctly.
Functions B, C, … call function A. This is a problem because function A could potentially trash the developer's $HOME if it works incorrectly, so it really needs to be tested *first*, and functions B, C, … must not be called if test A fails.