invariant case object encoding

[fommil]

close #8

[fommil]

@jdegoes encoding you may not have been aware of ^^^

[fommil]

(I think we can easily create a compiler plugin that does this optimisation automatically)

[jdegoes]

@fommil This looks good. I'll fix this in 8.

[xuwei-k]

[info] Compiling 12 Scala sources to /home/travis/build/scalaz/effect/target/scala-2.12/classes ...
[error] /home/travis/build/scalaz/effect/src/main/scala/scalaz/ioeffect/RTS.scala:535:34: unreachable code
[error]                             eval = false
[error]                                  ^
[error] 5 errors found
[error] (Compile / compileIncremental) Compilation failed

🤔

https://travis-ci.org/scalaz/ioeffect/builds/364861651#L515

[fommil]

@xuwei-k I have no idea what is going on there... looks like a compiler bug.

@jdegoes is the Later codepath covered in the tests? If so, I'll turn off the broken compiler warning.

[fommil]

I raised a bug report with scalac. I suspect a bug in the interplay between this encoding, var and try / catch.

If ever there was a way to start the day depressed, I recommend discovering yet another scalac cliff edge.

[emilypi]

@fommil I was confused about that bug as well when I reviewed the code and saw that it was failing for dead code. It seemed to be a total pattern match, so i'm left wondering why it's a problem. I'd be interested to see why scalac things that's dead code.

[jbgi]

in the meantime @silent can be use to suppress the warning.

[hrhino]

https://github.com/fommil/effect/pull/1

[hrhino]

src/main/scala/scalaz/ioeffect/AsyncReturn.scala:34:38: error: [DisableSyntax.covariant] Covariant types could lead to error-prone situations.

hahahahahahahahaha.

[fommil]

you can suppress scalafix warnings https://scalacenter.github.io/scalafix/docs/users/configuration#suppressing-rules

[fommil]

(I'll do it later)