sbt-native-image VS tapir

Take a look at sbt-native-image or sbt-native-packager for building the native image.

The reason seems to be Scala 3's new encoding for lazy vals, where a reflective access to a variable called bitmap takes place. To demonstrate this, I took the http4s g8 template, set it up with Scala 3.1.0, added sbt native-image, started the nativeImageRunAgent and ran a single request against GET /joke. After stopping the server, the native-image run agent generated the reflective configs that are required for this path of execution. Comparing the generated Scala 2.13 with the Scala 3 configs, yields a bloated version for Scala 3 that contains these additional rules:

Sbt plugin: https://github.com/scalameta/sbt-native-image

IIRC, the sbt-native-image plugin can abstract that away, too: https://github.com/scalameta/sbt-native-image#nativeimagerunagent https://github.com/scalameta/sbt-native-image

Assembly, however, has some disadvantages so you may also want to look into other alternatives like sbt-native-packager. You may even want to produce a binary using sbt-native-image (which uses GraalVM under the hood) so your users do not need to install anything at all, not even a JRE.

You're not confined to the usual suggestions below (play, http4s). There's a ton of options. (I wrote test cases using a bunch of different frameworks a few years ago at https://github.com/hohonuuli/msdemos). Having written services using a variety of frameworks in production, I would strongly suggest using one that auto-generates API docs (openapi, swagger) for you. That will save you a huge amount of time later on. For heavier services, like the one at https://fathomnet.org/, I tend to the Java side (Quarkus is my current top choice, but Micronaut and Helidon are both great). For everything else I use Scala. My go-to right now is tapir using a vertx backend. See https://tapir.softwaremill.com/

Tapir is a Scala framework. (which runs on the JDK) Since the recent release of version 1.0, it's become my go to for many projects. It doens't provide much in the way of integrations with 3rd party frameworks, but I actually prefer that. It does autogenerate great swagger docs though.

I do wonder where the recommendation to use http4s for beginners came from. http4s is a very capable library (and if you care much about composition it is excellent), but I wouldn't describe the documentation as beginner friendly.

A slightly better starting point for scala 3 + type-safe server building is tapir e.g. https://github.com/softwaremill/tapir/blob/master/examples3/... . With that, you get a declarative definition of your endpoints (+ error types, auth, etc.) that you can use for both servers and clients, which comes very handy when writing integration tests of course.

> absolutely ridiculous the fetishization of extremely complex FP and type-level hacking that goes on in the ecosystem

An alternative way to look at it is that there is a lot of essential domain complexity that gets encoded via the type system to let the compiler do the hard work. That "extremely complex FP" does not arrive out of nowhere - I really recommend at least skimming through the slides from rossabaker, the http4s designer, that motivate where the core type signature comes from https://rossabaker.github.io/boston-http4s/#2

I suppose one of the "features" that I like about the (typelevel) community is that the approach of "worse is better" is not taken, and a lot of effort is expended to make things correct, modular and orthogonal. This has the drawback of increased upfront complexity, that anecdotally pays off the moment your compiler does not error and the program runs as intended.

Did you try tapir? There is a module for deploying aws lambda with Scala js. Not sure whether it is compatible with Scala 3, I am sticking with Scala 2 until Scala 3 gets more mature.

I'm aware, but it's a design decision that was made on purpose, and which I find in practice not a big problem at all.

Backend is source of truth for types on frontend (backend generated OpenAPI definition with tapir, frontend takes it with orval)

In all cases I would strongly suggest to have a look at Tapir, regardless of the server implementation that you pick.

Check the infographic below, to see this tool history, functionalities and more. Make sure, to take tAPIr for a spin here and share your feedback with us in the comment section!

Tapir is awesome, and you can pick the server backend according to your preferred ecosystem (for instance http4s + doobie, Zio + Quill, Akka + Slick, ...)

Finally I would strongly recommend having a look at Tapir. Even if you don't need to share endpoints or generate OpenAPI documentation, it provides a really neat abstraction on top of http4s.