Doesn't this just mask issues? I don't remember if we have nullability annotations on, but I think if you want to add a check here, it should throw ArgumentNullException, and instead null should be handled at the call site.

It does mask issues, but in a non-failing way which is a lot better than raising during startup (which effectively renders this library useless). This is an adaptation of what fixes #2405. Throwing an exception here is (apart from having a slightly better stacktrace) effectively the same as leaving the code unchanged.

I mean I suggest to check for null at the call site which is what #2405 (comment) does, and optionally add an argument check in ShouldBindMethod to throw something better than NullReferenceException.

Simply returning false for null inputs is error-prone.

I don't agree, but I will also not put any more work into this. If you want to propose a different solution (in code), be my guest.

@filmor opened #2409