Elastic4s VS doobie

Here's an example from a library I've used: https://github.com/sksamuel/elastic4s/blob/master/elastic4s-core/src/main/scala/com/sksamuel/elastic4s/Executor.scala. Then you can provide your own Executor implementation and it will be used instead of the FutureExecutor. You can check out the top level of that repo to see how much complexity this introduces.

Seems like the accepted solution is to have a separate subproject for each of the ~5 major json libraries. For example see elastic4s (elastic4s-jackson, elastic4s-circe, ...) or akka-http-json.

Most important/known: * https://http4s.org/ - an HTTP client/server * https://github.com/typelevel/fs2 - streaming * https://github.com/tpolecat/doobie - JDBC

So you need to use PGgeography scala type and import another set of implicits: https://github.com/tpolecat/doobie/blob/main/modules/postgres/src/main/scala/doobie/postgres/pgisgeographyinstances.scala

Developer removing a library integration due to above politics

Doobie removes Quill integration in order to make some sort of political protest. The Quill developers are harassed on Twitter.

The libraries doobie and skunk are more closely associated with cats. They both use cats-effect and fs2 for implementing database connectiona and input-output operations. The doobie library is a wrapper on JDBC, and as such is compatible with many DBMS, such as MySQL, PostgreSQL, H2, Oracle... Whereas skunk is specific to PostgreSQL, and is based on using the server protocol of that database.

You may find this issue helpful. tl;dr you can easily make a Transactor with the void Strategy that just ignores the null Connection. But I’m with Rob: at that point, you’re not actually testing anything.