Codecov Report

Merging #3246 into master will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##           master    #3246   +/-   ##
=======================================
  Coverage   93.05%   93.05%           
=======================================
  Files         376      376           
  Lines        7413     7413           
  Branches      201      201           
=======================================
  Hits         6898     6898           
  Misses        515      515

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update d2f8a96...eeecbc3. Read the comment docs.

This effectively means that sbt release will not work anymore, since the release needs to be split across two separate sbt runs. I don't know another way to do this, but it also means that, in order to keep the release documented and reproducible, we should toss a shell script in here somewhere which replaces the functionality we're currently getting from sbt-release while juggling the environment variables.

@djspiewak Sure, I'll add a script. My preference would be to keep this fairly low-key, though—I don't want to commit to cross-building in the long term, or to make CI builds even longer by testing 1.0.0-RC2 on every PR now.

Perhaps you could switch to a system property, and do something like this:

releaseProcess := Seq[ReleaseStep](
  ...
  publishArtifacts,
  releaseStepCommand("""eval sys.prop("scalajs.version") = "1.0.0-RC2"; reload"),
  publishArtifacts,
  ...
)

@dwijnand Noooooo, please, no. Every time I've ever done anything like that in any build I've regretted it. All of this stuff is already a fragile pile of crap on the edge of disaster. I want to unblock people who want 1.0.0 primarily because I want us to be in a position to drop 0.6 the minute 1.0.0 final is out, and that's unlikely to be more than one or two Cats releases away. I think we should stick to the absolute minimum amount of extra process.

WFM 😄

@dwijnand In any case that would publish the JVM artifacts twice.

Yeah, but I think that's fine if you're using the new sonatype bundle thing.

And here's the 2.1.0 backport branch, which also passes SCALAJS_VERSION=1.0.0-RC2 sbt +validateJS: https://github.com/typelevel/cats/tree/backport/scala.js-1.0/2.1.0

@djspiewak Any objection to merging now?

@travisbrown Thanks for this. it would help unlock a big part of the ecosystem that can't be ported to RC2 just yet