Looks like quickcheck breaks due to a false positive in the overlapping instance check.

Ah, well, aside from it not fully working it looks really nice 😉

It's much simpler now 😄

How are the false positives arising?

Not sure yet, but I'll look into it.

Hah, it's failing because monadStateStateT and monadStateStateT1 overlap now, i.e. because it should 😄

It just means we need to patch transformers after this is released.

What on earth is that? 😉

Apparently needed by ParserT, but I assume that will work after this too?

I haven't tested parsers yet, but I will do now.

parsers compiles fine with the patch on transformers.

• Merge and release Updates for next major release purescript-transformers#49, verify core-tests
• Pick simplest instance when superclasses are involved.

Looks great! +:100: :shipit:

Although, there are non-transformers test failures in there too, is that because the tests are wrong now?

Oops, you're right. Looks like it's matching the suArray instance too eagerly in Superclasses2.purs. I'll look into it.

Hmm, GHC gives

/tmp/1.hs:21:13:
    Could not deduce (a ~ [a])
    from the context (Cl a)
      bound by the type signature for test :: Cl a => a -> a
      at /tmp/1.hs:20:9-24
      ‘a’ is a rigid type variable bound by
          the type signature for test :: Cl a => a -> a at /tmp/1.hs:20:9
    Relevant bindings include
      x :: a (bound at /tmp/1.hs:21:6)
      test :: a -> a (bound at /tmp/1.hs:21:1)
    In the first argument of ‘su’, namely ‘[cl x x]’
    In the expression: su [cl x x]
    In an equation for ‘test’: test x = su [cl x x]

So maybe we should let that test fail now.

Should we turn that into a failing test then, or do you think it's something that could be addressed? I guess if the new behaviour agrees with GHC then perhaps it is right 😉

I think it should probably be a failing test, yeah. I can't get to it today, but hopefully can make this mergeable tomorrow.

Oops, merging #1410 means this now conflicts.

I'm more than happy for this to go in whenever you're ready, by the way.

core-tests is breaking, but should work again once transformers is merged and released. I need to also release 0.7.4 to make that transformers release build, so I just need to coordinate a few things. If you think it's all good to go, I'll merge and release everything this afternoon.

Could you please run your tool and find reverse dependencies for transformers?

I'm currently working on a fix for #1244, I'd really like to get that in for 0.7.4, hopefully it shouldn't take too much longer.

I'll run off that list once I'm done :)

Also suffering from internet problems today so not actually present in IRC, posting this from my phone...

I can figure out the dependent modules, that's ok. I can make a transformers release now, but make it pre-release?

• purescript-quickcheck
• purescript-maps
• purescript-catenable-lists
• purescript-free
• purescript-strongcheck
• purescript-sets
• purescript-graphs
• purescript-parallel

maps, really?

Oh devDependencies though.

Yeah, via quickcheck I assume

No need for a version bump on those then, but we can update the dependency version. It's not going to break any code to pull in maps, since transformers won't be pulled in too, right?