Robin Krahl
Robin Krahl
And if we stick with `CStr`, we should use `core::ffi::CStr` to be able to benefit from [`c"..."` string literals](https://rust-lang.github.io/rfcs/3348-c-str-literal.html) instead of `cstr_core::CStr`.
We discussed this again and came to the conclusion that it makes more sense to have a `CStr`-based implementation for `Path` that can be used without conversion or allocation in...
The question is whether we need this restriction in `Path`, or if it is sufficient if the `PathBuf` and pointer conversions fail. Though they should definitely not panic.
The methods that pass `&Path` to littlefs use `as_ptr` to convert it to a pointer. We could change that method to check the length and return an error. But we...
I’m not entirely happy with `try_mount` either, but I did not see a better alternative. `mount_with_prepare` sounds like it would always call the prepare function. Maybe `mount_with_recover`?
Even if it’s just called with `format`, I would like to be able to know whether the initial mount worked. This could also encoded in the return type, though it...
@sosthene-nitrokey How about `mount_or_else`?
@sosthene-nitrokey What is the benefit of the wrapped error type over a non-exhaustive enum?
Ah, right, I misread `as_code` as only returning unknown codes. Makes sense.
I think the main problem is that v1.6.0 provides one CCID application, OpenPGP. v1.6.0-test.20231218 provides two CCID applications, OpenPGP and PIV. `opensc-tool` has the `--card-driver` option to select one: ```...