rmsone / slick-cats Goto Github PK

"org.typelevel" %% "cats-laws" % "1.0.1", is declared without % "Test", and it depends on cats-kernel-laws which depends on scalacheck.

This is causing my builds to somewhat non-deterministic in which scalacheck version makes it onto my classpath, so I need to work around it with exclude.

Assuming nothing from cats-laws is used in non-test code paths, fixing this should only take adding % "Test"

Hi,

The docs mention a 0.10.2 version which I couldn't find in Maven Central or Nexus. Has this been released yet?

Thanks.

I can't see 0.10.5 release in mvn repo

https://repo1.maven.org/maven2/com/rms/miu/slick-cats_2.13/

Hi,

It seems that an old version 0.7-RC1 is published in the Maven repository.
Would it be possible to upload the right (latest) version of slick-cats to Maven central

Thanks a lot.
Roll

Hi!

First of all - thanks for making this great library available to all of us. I was wondering if there's any chance of releasing a new version with Slick 3.4.0?

Thanks!

Hi, neat project!

I'm wondering if you have any ideas about making the instances compatible with a DBIOAction instead of a DBIO?

I'm trying to use the library in a project where we make use of the effect type (the E in DBIOAction[R, S, E]) in order to permit/deny actions via specific database handles. Similar to this example.

Using DBIO[A] means we lose track of the effect type when using monad transformers. For example, if I pass a DBIOAction[Option[Int], NoStream, Effect.Read] into an OptionT then it comes out as a DBIOAction[Option[Int], NoStream, Effect.All].

Curious if you have any thoughts on how one might preserve the effect type? I have a feeling it might not be possible, since F[_] only has one slot for a type parameter.

This would allow for us to combine multiple DBIOs in an applicative way.

dbioA |@| dbioB |@| dbioC should have the type of DBIO[(A, B, C)]

A simple implementation of the requisite function:

product[A, B]( fa: DBIO[A], fb: DBIO[B]): DBIO[(A, B)] = 
   (fa zip fb)

I think it was renamed from Cartesian to Semigroupal btw.

It would be good to use a custom release process to simplify the release, i.e. make sure tut and cross build steps aren't missed out

We'd like to use this but prefer to use ivy-managed dependencies in our builds.

Thanks!