Simon Peyton Jones pushed to branch wip/T23162-part2 at Glasgow Haskell Compiler / GHC Commits: 0a6db432 by Simon Peyton Jones at 2025-12-11T17:44:48+00:00 Wibble to insolubles - - - - - 1 changed file: - compiler/GHC/Tc/Solver/FunDeps.hs Changes: ===================================== compiler/GHC/Tc/Solver/FunDeps.hs ===================================== @@ -1089,7 +1089,7 @@ solveFunDeps work_ev fd_eqns = return (False, False) -- Common case no-op | otherwise - = do { traceTcS "bumping" (ppr work_ev) + = do { traceTcS "solveFunDeps {" (ppr work_ev) ; loc' <- bumpReductionDepth (ctEvLoc work_ev) (ctEvPred work_ev) -- See (CF3) in Note [Exploiting closed type families] @@ -1097,7 +1097,7 @@ solveFunDeps work_ev fd_eqns <- reportFineGrainUnifications $ nestFunDepsTcS $ TcS.pushTcLevelM_ $ - -- pushTcLevelTcM: increase the level so that unification variables + -- pushTcLevelTcM: increase the level so that unification variablesb -- allocated by the fundep-creation itself don't count as useful unifications -- See Note [Deeper TcLevel for partial improvement unification variables] do { (_, eqs) <- wrapUnifier (ctEvRewriters work_ev) loc' Nominal $ @@ -1111,7 +1111,12 @@ solveFunDeps work_ev fd_eqns -- that were unified by the fundep ; kickOutAfterUnification unifs - ; return (insolubleWC residual, not (isEmptyVarSet unifs)) } + ; let insoluble_fundeps = any insolubleCt (wc_simple residual) + -- Don't use insolubleWC, because that ignores Given constraints + -- and Given constraints are super-important when doing + -- tcCheckGivens in the pattern match overlap checker + ; traceTcS "solveFunDeps }" (ppr insoluble_fundeps <+> ppr unifs $$ ppr residual) + ; return (insoluble_fundeps, not (isEmptyVarSet unifs)) } -- insolubleWC: see (CF3) in Note [Exploiting closed type families] where do_fundeps :: UnifyEnv -> TcM () View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/0a6db4320b314ad835c8df61bced24cb... -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/0a6db4320b314ad835c8df61bced24cb... You're receiving this email because of your account on gitlab.haskell.org.