<@UEJCGUJCA> Does the new stability inference feat...

# compose

[Zach Klippenstein (he/him) [MOD]]

@Leland Richardson [G] Does the new stability inference feature mean that the Stable/Immutable annotations can eventually be deprecated? Are there still cases that the inference misses that need to be explicitly annotated?

[Leland Richardson [G]]

It means that their use should be drastically reduced, but I don’t think we will deprecate them any time soon

[Leland Richardson [G]]

for one, a stable annotation on a class is a contract, whereas inferred stability is not

[louiscad]

They could be replaced by val class if/when that becomes a thing?

[Leland Richardson [G]]

if an interface is marked @Stable, it means that subclasses of it are expected (or, “required”) to be stable.

[louiscad]

val interface? 🤔

[Leland Richardson [G]]

val class will certainly qualify for stability, but it is not a replacement. it just becomes another category of stable types

[Leland Richardson [G]]

we also can’t expect the inference to be perfect. you will be able to write classes that have state but have a public API that can be considered stateful, but the compiler won’t be able to prove that

[Leland Richardson [G]]

the goal here was to make this something that was “good enough” that most of the time you just don’t have to think about it anymore

[Zach Klippenstein (he/him) [MOD]]

Cool, that makes perfect sense