Description of problem: Building the latest git-annex fails on F31: https://koji.fedoraproject.org/koji/taskinfo?taskID=38557890 and Rawhide: https://koji.fedoraproject.org/koji/taskinfo?taskID=38557895 but not F30: https://koji.fedoraproject.org/koji/taskinfo?taskID=38557889 Version-Release number of selected component (if applicable): 8.6.5-100.fc31 Steps to Reproduce: 1. Build git-annex on F31 or Rawhide Actual results: [565 of 628] Compiling Assistant.WebApp.Types ( Assistant/WebApp/Types.hs, dist/build/git-annex/git-annex-tmp/Assistant/WebApp/Types.o ) ghc: panic! (the 'impossible' happened) (GHC version 8.6.5 for arm-unknown-linux): tyThingTyCon Identifier ‘$tcT1’ Call stack: CallStack (from HasCallStack): callStackDoc, called at compiler/utils/Outputable.hs:1160:37 in ghc:Outputable pprPanic, called at compiler/main/HscTypes.hs:2165:28 in ghc:HscTypes Please report this as a GHC bug: http://www.haskell.org/ghc/reportabug Expected results: Build is fine.
Sorry, only just noticed this now... It looks like the original problem fixed itself, however: [566 of 629] Compiling Assistant.WebApp.Types ( Assistant/WebApp/Types.hs, dist/build/git-annex/git-annex-tmp/Assistant/WebApp/Types.o ) Assistant/WebApp/Types.hs:49:37: error: • Identifier ‘String’ used as a type • In the first argument of ‘IO’, namely ‘String’ In the first argument of ‘Maybe’, namely ‘(IO String)’ In the type ‘Maybe (IO String)’ | 49 | , postFirstRun :: Maybe (IO String) | ^^^^^^ However it seems to build okay with -O1 (ie with the `Production` flag set to `False`). It might better to move and rename this bug, open a new bug against git-annex, or close it now with -O1.
In the end I opened bug 1797176. So I will close out this original bug for the ghc compiler crash.