Fix internal methods using ReadOnlySpans as writable

[TimonLukas]

(fixes #103)

I've only changed the methods that read directly into ReadOnlySpans. It might be worth it to update the ones that read into WCharTStrings if the library is adapted as well - I'll take a look at the implementation in the coming days, it sounds really interesting! :)

The change is implemented as discussed, but my IDE (Jetbrains Rider) suggests using the object initializer syntax for setting the report ID as well. This would mean there are 3 possible implementations (using the example of GetFeatureReport again):

Option 1 (currently implemented):

public ReadOnlySpan<byte> GetFeatureReport(byte reportId, int maxLength)
{
    if (maxLength < 1)
        throw new ArgumentOutOfRangeException(nameof(maxLength), maxLength, "Please provide a value greater than 1");

    var data = new Span<byte>(new byte[maxLength]);
    data[0] = reportId;
    var result = NativeMethods.GetFeatureReport(handle, data);

    if (result == -1)
        HidException.Throw(handle);

    return data[..result];
}

Option 2 (with object initializer):

public ReadOnlySpan<byte> GetFeatureReport(byte reportId, int maxLength)
{
    if (maxLength < 1)
        throw new ArgumentOutOfRangeException(nameof(maxLength), maxLength, "Please provide a value greater than 1");

    var data = new Span<byte>(new byte[maxLength]) { [0] = reportId };
    var result = NativeMethods.GetFeatureReport(handle, data);

    if (result == -1)
        HidException.Throw(handle);

    return data[..result];
}

Option 3 (implicit conversion):

public ReadOnlySpan<byte> GetFeatureReport(byte reportId, int maxLength)
{
    if (maxLength < 1)
        throw new ArgumentOutOfRangeException(nameof(maxLength), maxLength, "Please provide a value greater than 1");

    Span<byte> data = new byte[maxLength];
    data[0] = reportId;
    var result = NativeMethods.GetFeatureReport(handle, data);

    if (result == -1)
        HidException.Throw(handle);

    return data[..result];
}

Which do you prefer? I think all options are readable, though 2 reads the worst IMO (and I'm not sure it's correctly formatted, it's what my IDE suggested).

[badcel]

I think avoiding implicit casts makes the code easier to grasp especially for newcomers (option 1).

Nevertheless I think if a developer knows about C# implicit conversions option 3 is easier to read as the types are more separated but it hides the constructor call.

I still tend to option 1 even if it is a bit harder to read for myself than option 3. What do you think?

Option 2 is too noisy for me. Do you know if there is some compiler optimization possible if object initializers are used?

[TimonLukas]

Option 1 is totally fine! I'd also have gone with 1 or 3, maybe tending a little towards 3 - mostly because I mentally replace the Span<byte> with byte[]. But it really doesn't matter, all three seem to lead to the same compiled output. So, I'd say it's ready for review!