Closed Munksgaard closed 4 years ago
Forgot to mention that I'm using ormolu from nixos-unstable:
$ ormolu --version
ormolu 0.0.3.1 UNKNOWN UNKNOWN
using ghc-lib-parser 8.8.3.20200224
How this is different from the existing tickets?
Ah, it's probably not. I hadn't seen you also had tickets regarding comments. Thanks for your work!
I'm evaluating formatters for the Futhark project, and upon trying ormolu the files in that project, the following five files exhibited idempotency bugs in ormolu:
https://github.com/diku-dk/futhark/blob/b407b5323c4cef993f18e76cc3ada00d84679428/src/Futhark/Representation/AST/Pretty.hs https://github.com/diku-dk/futhark/blob/b407b5323c4cef993f18e76cc3ada00d84679428/src/Futhark/Analysis/HORepresentation/SOAC.hs https://github.com/diku-dk/futhark/blob/b407b5323c4cef993f18e76cc3ada00d84679428/src/Futhark/Optimise/Fusion.hs https://github.com/diku-dk/futhark/blob/b407b5323c4cef993f18e76cc3ada00d84679428/src/Language/Futhark/TypeChecker/Terms.hs https://github.com/diku-dk/futhark/blob/b407b5323c4cef993f18e76cc3ada00d84679428/src/Language/Futhark/TypeChecker/Unify.hs
Most of them seem to have to do with
<>
and<+>
operators (and thus could be related to https://github.com/tweag/ormolu/issues/451), but in SOAC.hs and Fusion.hs, there also seem to be comments that are incorrectly formatted the second time around.