Lázaro Clapp

@msridhar FYI, we also want to consume those stubs in the local version of the RxChecker, so we can just have a generic infrastructure for getting CF stubs into EP,...

@Stephan202 The RxChecker (technically: the Rx Thread/Effect Checker) is an internal tool for the time being. It is more about threading issues in RxAndroid than any sort of general toolbox...

NullAway, by design, assumes that "no annotation" implies non-null for code being directly checked by it. Any method argument, return value, or field that can have a `null` value must...

Unfortunately, as far as I know, there isn't currently an option to reverse that default in NullAway, nor is a single place of the codebase that would need to be...

> Thanks a lot for linking https://github.com/uber/NullAway/issues/453! Agreed! (And sorry to see we kinda dropped the ball on that issue!) I actually ran into this due to our test suite...

NullAway is saying that the class constructor (the only initializer I see there) is not setting these three fields to anything: > language (line 22), moduleType (line 24), project (line...

Oh, right, sorry. I saw this fairly late last night and I missed the fact that you were setting `picocli.CommandLine.Option` to be an `@Initializer` annotation. Note that, for the three...

> Basically, the meaning of having multiple `@Initializer` methods is that all of them should be called before the class is used. I honestly don't remember why we chose to...

> Could you subclass `Supplier` and annotate the subclass' return value with `@Nonnull`? > > eg: > > ``` > interface NonNullSupplier extends Supplier { > @Nonnull T get(); >...

@msridhar The ask is around making `Supplier.get()` default to returning `@Nullable`, rather than `@NonNull`, under the current NullAway type system. This is certainly sounder, given our current limitations, but at...