Contributor
Confirmed with stackage nightly.
Note that greskell-core depends on containers but it doesn't support containers-0.6 yet. I don't think that is a problem with stackage for now, but correct me if I'm wrong.
Member
i can check whether we can re-enable doctest-discover once i get back to my compiter
Member
Some of the greskell tests are failing, is this expected?
Split typeclass
Filter -> Filter FAILED [1]
Filter -> Transform FAILED [2]
Filter -> SideEffect FAILED [3]
Transform -> Filter FAILED [4]
Transform -> Transform FAILED [5]
Transform -> SideEffect FAILED [6]
SideEffect -> Filter
SideEffect -> Transform
SideEffect -> SideEffect FAILED [7]
Lift typeclass
Filter -> Filter FAILED [8]
Filter -> Transform FAILED [9]
Filter -> SideEffect FAILED [10]
Transform -> Filter
Transform -> Transform FAILED [11]
Transform -> SideEffect FAILED [12]
SideEffect -> Filter
SideEffect -> Transform
SideEffect -> SideEffect FAILED [13]
Failures:
test/HintTest.hs:53:7:
1) Split typeclass Filter -> Filter
expected: True
but got: False
test/HintTest.hs:53:7:
2) Split typeclass Filter -> Transform
expected: True
but got: False
test/HintTest.hs:53:7:
3) Split typeclass Filter -> SideEffect
expected: True
[...]
Contributor Author
No it's not.
They test if Split and Lift typeclasses allow only for certain pairs of types. We use hint to check some pairs pass compile and other pairs don't. However, the test results you showed seems like all pairs fail to compile. I'm not sure why.
I think those tests are tricky and probably unstable. Maybe we should disable hint-test flag on greskell to disable the tests (the flag is enabled by default.)
Member
It seems like it worked now!
But I'd recommend adding manual: True to that flag, otherwise cabal-install (maybe also the stackage build server, I'm not sure) may toggle it automatically if, for example, there are dependency issues.