Richard Eisenberg pushed to branch wip/rae/unconditional-optimizations at Glasgow Haskell Compiler / GHC Commits: cc9227b1 by Richard Eisenberg at 2026-03-17T22:33:15-04:00 Force optimizations on where required for building - - - - - 2 changed files: - compiler/GHC/Data/TrieMap.hs - compiler/GHC/Tc/Solver/InertSet.hs Changes: ===================================== compiler/GHC/Data/TrieMap.hs ===================================== @@ -1,5 +1,8 @@ {-# LANGUAGE TypeFamilies #-} {-# LANGUAGE UndecidableInstances #-} +{-# OPTIONS_GHC -O -fno-omit-interface-pragmas #-} + -- There are SPECIALIZE pragmas in GHC.Core.Map.Type that require + -- the ability to inline some functions here {- (c) The University of Glasgow 2006 ===================================== compiler/GHC/Tc/Solver/InertSet.hs ===================================== @@ -1,4 +1,7 @@ {-# LANGUAGE MultiWayIf #-} +{-# OPTIONS_GHC -O -fno-omit-interface-pragmas #-} + -- There is a SPECIALIZE pragma in GHC.Tc.Solver.Monad that requires + -- the ability to inline forAllCycleBreakerBindings_ some functions here module GHC.Tc.Solver.InertSet ( -- * The work list @@ -2174,4 +2177,3 @@ Wrong! The level-check ensures that the inner implicit parameter wins. (Actually I think that the order in which the work-list is processed means that this chain of events won't happen, but that's very fragile.) -} - View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/cc9227b1c86063a9cbac67d9d2977830... -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/cc9227b1c86063a9cbac67d9d2977830... You're receiving this email because of your account on gitlab.haskell.org.