Georgi Krastev
Georgi Krastev
Let me know if I can help with something 👍
> @joroKr21 I had to drop the whole PR history, but on the bright side it was an ez code backport. Commits history is preserved here: https://github.com/pomadchin/cats-tagless/tree/feature/scala-3-init-backup Alright, sounds good...
Thank you for driving this @pomadchin! It's been a massive effort.
So basically we're saying that decomposing refined types is more important than dealiasing. Need to think through the implications because there might be situations we want the opposite. Also there...
`A isSubType Nothing` should be false right?
> Where can I put such a test? Can you provide a link to tests that I could used as a reference? You can add a positive "partest": https://github.com/scala/scala/blob/2.13.x/CONTRIBUTING.md#partest >...
> It's very strange, but [this](https://scastie.scala-lang.org/CBBIQNmkSfyV3bN1gzPMmQ) also cannot be compiled in Scala 2. Ok, so this would indicate it has nothing to do with abstract types.
That is more or less on purpose, see #9101 - if we use streams with > 1 element then combining multiple generators would behave like `ZStream.flatMap` which means not really...
> The example with filtering to constant Int was just for illustration of runHead and performance issues that I experience. I think the performance issues you are experiencing come from...
@disordered what if I write it like this? Should it be regenerating `first` or not? ```scala test("gen filter doesn't backtrack") { val filteredGen = for { first