better-monadic-for | Desugaring scala ` for ` without implicit ` withFilter ` s | Compiler library

The README.md of scala-newtype says:

This expands into a type and companion object definition, so newtypes must be defined in an object or package object.

Macros are allowed to expand classes into other classes with the same name and companion objects, but from what I can tell, the newtype annotation turns your case class into an object of the same name (along with a type alias like type DbUrl = DbUrl.Type). This behavior (turning a top-level annottee into a tree of some other kind) isn't allowed. If the annotation had generated a class DbUrl, and maybe an object of the same name, though, it would have been all right, but pretty much anything else won't work.

To fix your problem, all you need to do is move this into a package object (or some other scope, as long as it isn't top-level).

Edit: As Dmytro Mitin pointed out, the created type is not the type of DbUrl but rather something like type DbUrl = DbUrl.Type with an uppercase "T", where the definition of DbUrl.Type looks something like this (I'm just copying this from the README):

There is a bug in documentation - see https://github.com/sbt/sbt/issues/4490

Try using dedicated sbt plugin - xsbt-web-plugin - instead of reading that website. According to current docs you need to add to project/plugins.sbt

You don't have a main class, the entrypoint to your program defined. On JVM you have to have at least one class which defines

I just hit this exception in a test, and it was caused by trying to run a Spark action in a thread that was different from the thread where I started the SparkSession. You might want to disable parallelExecution in Test (this is recommended for Spark integration tests anyway).

Specifically, I was trying to execute multiple Spark actions in parallel, and I tried doing that in Scala's ExecutionContext.global thread pool. When I created a newFixedPoolExecutor instead, everything started working fine.

AFAICT this is because in DataSource.scala:610, Spark gets the thread's ContextClassLoader:

In the Official Gitter Channel for ZIO Core, Adam Fraser suggested

You would want to have you environment just have a Queue[String] and then you would want to use a method like provideM with Queue.unbounded to create one queue and provide it to your whole application. That's where provideM as opposed to provide comes in. It let's you satisfy an environment that requires an A by providing a ZIO[A].

A little digging into the ZIO source revealed a helpful example in DefaultTestReporterSpec.scala.

By defining the Environment as

Put libraryDependencies ++= dependencies into settings.

global, core and serversupervisor are three different subprojects. They can have different library dependencies. Currently you add them to global but not to core and serversupervisor.

Alternatively you can move libraryDependencies ++= dependencies to Global or ThisBuild scope rather than specific subproject scope. You can add at top

You use this in the Compile scope (i.e. sources in src/main), so you need to remove % "test" from the ScalaCheck dependency. Or move that source to the Test scope (i.e. in src/test)