Remi Rousselet

This is valuable. Although I'm wondering whether we need a flag at all. Why not always make the generated class `final` if it helps? Extending a Freezed class isn't supported...

> Would be happy to remove a whole bunch of code and just hardcode sealed and final when generating subclasses, but as I said before: I would 100% support this,...

You can ignore the changeset bot btw

Cool! I'll leave this open for a few more days, but it should be good to go. I just need to try it locally and evaluate the possible danger of...

My appologies, I kind of forgot about this PR. Let me see what we can do here. Feel free to ping me if I forget again.

Sorry as you can see, I got sidetracked. I'm not sure how relevant it is anymore. To be honest, looking back at the motivation, I wonder if the iDE isn't...

Actually ignore me, I can see why the analyzer works this way. if you don't mind fixing the conflicts, we can merge this. For now, let's not enable it by...

Thinking about it, I wonder if this should be slightly generalised. We could have `@Freezed(classKeywords: ['final'])` This would support `base`/`mixin`/`sealed`/... (we'd need an assert error if the keyword is invalid)...

Ah, I kind of forgot that we wouldn't be able to mark the generated class `sealed`. Even though Freezed has inheritance now, the factory constructor still wouldn't work. So we'd...

Th refresh only happens if the provider is listened. From your code, it is unclear if your provider is listened or not. So I'll assume it isn't and that the...