cats-effect3 migration path

[tomasherman]

So, what are your thoughts on migration to cats-effect3? I have not tried but i suspect that most of libraries should now be ready for migration. In my view, we will just need to make the jump at some point. WDYT?

adding people that might be interested - @jakubjanecek @jendakol @sideeffffect @hanny24 @Peppi-Ressler @nb-ceffa (if any of you don't care pls let me know so i don't tag you in future discussions) :) or anyone else pls chime in

[jakubjanecek]

Yes, I agree and of course I want to upgrade. I was just focusing more on Scala 3 first. I think we can have a build both for Scala 2 and Scala 3, which is good, but I'm not sure we are able to have a build for both cats-effect 2 and 3 and that would be troublesome even though most libraries are probably ready.

[tomasherman]

yeah agree that scala3 should come first. I also agree it would be very difficult to maintain both ce2 and ce3. Basically, what i had to do in datadog4s is to maintain two sets of source code versions. Its possible to do there as its a small project. Here it would be a lot more difficult. So i would suggest freezing ce2 support at some stage, bump major version and rewrite for ce3. We can always backport some huge bugs or dependency updates (thinking about log4j) if the need arise for ce2

[tomasherman]

so i took a stab at CE3 migration - there will be some obstacles. The biggest one is that Monix does not support CE3 type classes (and probably won't any time soon). We rely on monix circuit breaker so if we decided to move forward, we will need to replace it with other things (https://github.com/davenverse/circuit has worked well for us in prod) ... i don't think, realistically, there is no other way

[jakubjanecek]

I trust you so if that's the way let's do it (if you want to 😉).

[tomasherman]

at some point, i will discuss with people at avast internally aswell ... but it's gonna be quite some work i think