Okay, I've given it a casual skim. I'll be rebasing my own stuff early next week so I'll probably come with more but I don't want to leave you empt...
It'd be good to have a smaller code example here so it's immediately clear what does it do.
Hmm... I don't think PathBuf is marked #[must_use]. Could you check if this works:...
Is this line necessary?
/// Sets the file mode of the socket to `666` so that all users on the system have access.This is slightly problematic since if the tests fails before this line is reached then the the subprocess will not be killed. Sadly, depending on t...
expect_err will do that for you:...
Can user take any other values than "root"?
Assert may panic. If you think it's a good idea to have it here I'd at least insert a greppable message:...
I wonder if these constraints should be part of the documentation (e.g. "creates a password consisting of 30 alphanumeric characters").
Is there a reason you don't shorten this stuff into:...
Hmm... I wonder if you want to explicitly assign values to variants to avoid AdminCredentialsConfigFromToml being assigned 0 etc.:...
What's our current standing on mut variables scope? A couple of months ago there was a suggestion to wrap them in a block to make them immutable?
Maybe that's just me but I think it'd be more obvious if the function was called fail_if_non_root
pub fn store<P>(&self, path: impl AsRef<Path>) -> Result<(), crate::Error>There's quite a lot of interesting code down there but the test is marked as no_run. Is that intentional?
context: "after decrypting".to_string(),This equivalent syntax is shorter:...