Zubin pushed new branch wip/inline-fs at Glasgow Haskell Compiler / GHC
--
View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/tree/wip/inline-fs
You're receiving this email because of your account on gitlab.haskell.org.
1
0
[Git][ghc/ghc][ghc-9.14] level imports: Fix infinite loop with cyclic module imports
by Zubin (＠wz1000) 05 Nov '25
by Zubin (＠wz1000) 05 Nov '25
05 Nov '25
Zubin pushed to branch ghc-9.14 at Glasgow Haskell Compiler / GHC
Commits:
0e08b9ae by Matthew Pickering at 2025-11-04T20:56:40+05:30
level imports: Fix infinite loop with cyclic module imports
I didn't anticipate that downsweep would run before we checked for
cyclic imports. Therefore we need to use the reachability function which
handles cyclic graphs.
Fixes #26087
(cherry picked from commit 8b731e3c900291655a767123bcda55eddd63920c)
- - - - -
5 changed files:
- compiler/GHC/Unit/Module/Graph.hs
- + testsuite/tests/splice-imports/T26087.stderr
- + testsuite/tests/splice-imports/T26087A.hs
- + testsuite/tests/splice-imports/T26087B.hs
- testsuite/tests/splice-imports/all.T
Changes:
=====================================
compiler/GHC/Unit/Module/Graph.hs
=====================================
@@ -866,7 +866,7 @@ mkTransZeroDeps = first graphReachability {- module graph is acyclic -} . module
-- | Transitive dependencies, but with the stage that each module is required at.
mkStageDeps :: [ModuleGraphNode] -> (ReachabilityIndex StageSummaryNode, (NodeKey, ModuleStage) -> Maybe StageSummaryNode)
-mkStageDeps = first graphReachability . moduleGraphNodesStages
+mkStageDeps = first cyclicGraphReachability . moduleGraphNodesStages
type ZeroSummaryNode = Node Int ZeroScopeKey
=====================================
testsuite/tests/splice-imports/T26087.stderr
=====================================
@@ -0,0 +1,6 @@
+./T26087B.hs: error: [GHC-92213]
+ Module graph contains a cycle:
+ module ‘main:T26087B’ (./T26087B.hs)
+ imports module ‘main:T26087A’ (T26087A.hs)
+ which imports module ‘main:T26087B’ (./T26087B.hs)
+
=====================================
testsuite/tests/splice-imports/T26087A.hs
=====================================
@@ -0,0 +1,4 @@
+{-# LANGUAGE ExplicitLevelImports #-}
+module T26087A where
+
+import quote T26087B
=====================================
testsuite/tests/splice-imports/T26087B.hs
=====================================
@@ -0,0 +1,4 @@
+{-# LANGUAGE ExplicitLevelImports, TemplateHaskell #-}
+module T26087B where
+
+import T26087A
=====================================
testsuite/tests/splice-imports/all.T
=====================================
@@ -46,7 +46,4 @@ test('SI35',
compile_and_run,
['-package ghc'])
test('SI36', [extra_files(["SI36_A.hs", "SI36_B1.hs", "SI36_B2.hs", "SI36_B3.hs", "SI36_C1.hs", "SI36_C2.hs", "SI36_C3.hs"])], multimod_compile_fail, ['SI36', '-v0'])
-test('T26090', [], multimod_compile_fail, ['T26090', '-v0'])
-test('ModuleExport', [], multimod_compile_fail, ['ModuleExport', '-v0'])
-test('LevelImportExports', [], makefile_test, [])
-test('DodgyLevelExport', [], multimod_compile, ['DodgyLevelExport', '-v0 -Wdodgy-exports'])
+test('T26087', [], multimod_compile_fail, ['T26087A', ''])
View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/0e08b9aee6ce58fe7099f14ac1d7dea…
--
View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/0e08b9aee6ce58fe7099f14ac1d7dea…
You're receiving this email because of your account on gitlab.haskell.org.
1
0
[Git][ghc/ghc][wip/marge_bot_batch_merge_job] 7 commits: rts: fix eager black holes: record mutated closure and fix assertion
by Marge Bot (＠marge-bot) 05 Nov '25
by Marge Bot (＠marge-bot) 05 Nov '25
05 Nov '25
Marge Bot pushed to branch wip/marge_bot_batch_merge_job at Glasgow Haskell Compiler / GHC
Commits:
3ba3d9f9 by Luite Stegeman at 2025-11-04T00:59:41-05:00
rts: fix eager black holes: record mutated closure and fix assertion
This fixes two problems with handling eager black holes, introduced
by a1de535f762bc23d4cf23a5b1853591dda12cdc9.
- the closure mutation must be recorded even for eager black holes,
since the mutator has mutated it before calling threadPaused
- The assertion that an unmarked eager black hole must be owned by
the TSO calling threadPaused is incorrect, since multiple threads
can race to claim the black hole.
fixes #26495
- - - - -
b5508f2c by Rodrigo Mesquita at 2025-11-04T14:10:56+00:00
build: Relax ghc/ghc-boot Cabal bound to 3.16
Fixes #26202
- - - - -
c5b3541f by Rodrigo Mesquita at 2025-11-04T14:10:56+00:00
cabal-reinstall: Use haddock-api +in-tree-ghc
Fixes #26202
- - - - -
c6d4b945 by Rodrigo Mesquita at 2025-11-04T14:10:56+00:00
cabal-reinstall: Pass --strict to Happy
This is necessary to make the generated Parser build successfully
This mimics Hadrian, which always passes --strict to happy.
Fixes #26202
- - - - -
79df1e0e by Rodrigo Mesquita at 2025-11-04T14:10:56+00:00
genprimopcode: Require higher happy version
I've bumped the happy version to forbid deprecated Happy versions which
don't successfully compile.
- - - - -
9cbfc2ce by Simon Peyton Jones at 2025-11-04T23:24:30-05:00
Add a HsWrapper optimiser
This MR addresses #26349, by introduceing optSubTypeHsWrapper.
There is a long
Note [Deep subsumption and WpSubType]
in GHC.Tc.Types.Evidence that explains what is going on.
- - - - -
4098947b by Simon Peyton Jones at 2025-11-04T23:24:30-05:00
Improve mkWpFun_FRR
This commit ensures that `mkWpFun_FRR` directly produces a `FunCo` in
the cases where it can.
(Previously called `mkWpFun` which in turn optimised to a `FunCo`, but
that made the smarts in `mkWpFun` /essential/ rather than (as they
should be) optional.
- - - - -
24 changed files:
- cabal.project-reinstall
- compiler/GHC/Core/Coercion.hs
- compiler/GHC/Hs/Syn/Type.hs
- compiler/GHC/HsToCore/Binds.hs
- compiler/GHC/HsToCore/Match.hs
- compiler/GHC/Iface/Ext/Ast.hs
- compiler/GHC/Tc/Errors/Hole.hs
- compiler/GHC/Tc/Gen/App.hs
- compiler/GHC/Tc/Gen/Head.hs
- compiler/GHC/Tc/Gen/Pat.hs
- compiler/GHC/Tc/Types/Evidence.hs
- compiler/GHC/Tc/Utils/Concrete.hs
- compiler/GHC/Tc/Utils/Unify.hs
- compiler/GHC/Tc/Zonk/Type.hs
- compiler/Setup.hs
- compiler/ghc.cabal.in
- libraries/ghc-boot/Setup.hs
- libraries/ghc-boot/ghc-boot.cabal.in
- rts/ThreadPaused.c
- + testsuite/tests/simplCore/should_compile/T26349.hs
- + testsuite/tests/simplCore/should_compile/T26349.stderr
- testsuite/tests/simplCore/should_compile/all.T
- testsuite/tests/simplCore/should_compile/rule2.stderr
- utils/genprimopcode/genprimopcode.cabal
Changes:
=====================================
cabal.project-reinstall
=====================================
@@ -59,6 +59,7 @@ constraints: ghc +internal-interpreter +dynamic-system-linke,
ghc-bin +internal-interpreter +threaded,
ghci +internal-interpreter,
haddock +in-ghc-tree,
+ haddock-api +in-ghc-tree,
any.array installed,
any.base installed,
any.deepseq installed,
@@ -68,6 +69,8 @@ constraints: ghc +internal-interpreter +dynamic-system-linke,
any.pretty installed,
any.template-haskell installed
+package *
+ happy-options: --strict
benchmarks: False
tests: False
=====================================
compiler/GHC/Core/Coercion.hs
=====================================
@@ -41,7 +41,8 @@ module GHC.Core.Coercion (
mkInstCo, mkAppCo, mkAppCos, mkTyConAppCo,
mkFunCo, mkFunCo2, mkFunCoNoFTF, mkFunResCo,
mkNakedFunCo,
- mkNakedForAllCo, mkForAllCo, mkForAllVisCos, mkHomoForAllCos,
+ mkNakedForAllCo, mkForAllCo, mkForAllVisCos,
+ mkHomoForAllCo, mkHomoForAllCos,
mkPhantomCo, mkAxiomCo,
mkHoleCo, mkUnivCo, mkSubCo,
mkProofIrrelCo,
@@ -980,7 +981,7 @@ mkForAllCo v visL visR kind_co co
= mkReflCo r (mkTyCoForAllTy v visL ty)
| otherwise
- = mkForAllCo_NoRefl v visL visR kind_co co
+ = mk_forall_co v visL visR kind_co co
-- mkForAllVisCos [tv{vis}] constructs a cast
-- forall tv. res ~R# forall tv{vis} res`.
@@ -1000,14 +1001,26 @@ mkHomoForAllCos vs orig_co
= foldr go orig_co vs
where
go :: ForAllTyBinder -> Coercion -> Coercion
- go (Bndr var vis) = mkForAllCo_NoRefl var vis vis MRefl
-
--- | Like 'mkForAllCo', but there is no need to check that the inner coercion isn't Refl;
--- the caller has done that. (For example, it is guaranteed in 'mkHomoForAllCos'.)
--- The kind of the tycovar should be the left-hand kind of the kind coercion.
-mkForAllCo_NoRefl :: TyCoVar -> ForAllTyFlag -> ForAllTyFlag
- -> KindMCoercion -> Coercion -> Coercion
-mkForAllCo_NoRefl tcv visL visR kind_co co
+ go (Bndr var vis) co = mk_forall_co var vis vis MRefl co
+
+mkHomoForAllCo :: TyVar -> Coercion -> Coercion
+-- Specialised for a single TyVar,
+-- and visibility of coreTyLamForAllTyFlag
+mkHomoForAllCo tv orig_co
+ | Just (ty, r) <- isReflCo_maybe orig_co
+ = mkReflCo r (mkForAllTy (Bndr tv vis) ty)
+ | otherwise
+ = mk_forall_co tv vis vis MRefl orig_co
+ where
+ vis = coreTyLamForAllTyFlag
+
+-- | `mk_forall_co` just builds a ForAllCo.
+-- With debug on, it checks invariants (e.g. he kind of the tycovar should
+-- be the left-hand kind of the kind coercion).
+-- Callers should have done any isReflCo short-cutting.
+mk_forall_co :: TyCoVar -> ForAllTyFlag -> ForAllTyFlag
+ -> KindMCoercion -> Coercion -> Coercion
+mk_forall_co tcv visL visR kind_co co
= assertGoodForAllCo tcv visL visR kind_co co $
assertPpr (not (isReflCo co && isReflMCo kind_co && visL == visR)) (ppr co) $
ForAllCo { fco_tcv = tcv, fco_visL = visL, fco_visR = visR
@@ -1769,7 +1782,7 @@ mkPiCos r vs co = foldr (mkPiCo r) co vs
-- | Make a forall 'Coercion', where both types related by the coercion
-- are quantified over the same variable.
mkPiCo :: Role -> Var -> Coercion -> Coercion
-mkPiCo r v co | isTyVar v = mkHomoForAllCos [Bndr v coreTyLamForAllTyFlag] co
+mkPiCo r v co | isTyVar v = mkHomoForAllCo v co
| isCoVar v = assert (not (v `elemVarSet` tyCoVarsOfCo co)) $
-- We didn't call mkForAllCo here because if v does not appear
-- in co, the argument coercion will be nominal. But here we
=====================================
compiler/GHC/Hs/Syn/Type.hs
=====================================
@@ -187,11 +187,13 @@ liftPRType :: (Type -> Type) -> PRType -> PRType
liftPRType f pty = (f (prTypeType pty), [])
hsWrapperType :: HsWrapper -> Type -> Type
+-- Return the type of (WrapExpr wrap e), given that e :: ty
hsWrapperType wrap ty = prTypeType $ go wrap (ty,[])
where
go WpHole = id
+ go (WpSubType w) = go w
go (w1 `WpCompose` w2) = go w1 . go w2
- go (WpFun _ w2 (Scaled m exp_arg)) = liftPRType $ \t ->
+ go (WpFun _ w2 (Scaled m exp_arg) _) = liftPRType $ \t ->
let act_res = funResultTy t
exp_res = hsWrapperType w2 act_res
in mkFunctionType m exp_arg exp_res
=====================================
compiler/GHC/HsToCore/Binds.hs
=====================================
@@ -1597,9 +1597,13 @@ dsHsWrapper hs_wrap thing_inside
ds_hs_wrapper :: HsWrapper
-> ((CoreExpr -> CoreExpr) -> DsM a)
-> DsM a
-ds_hs_wrapper wrap = go wrap
+ds_hs_wrapper hs_wrap
+ = go hs_wrap
where
go WpHole k = k $ \e -> e
+ go (WpSubType w) k = go (optSubTypeHsWrapper w) k
+ -- See (DSST3) in Note [Deep subsumption and WpSubType]
+ -- in GHC.Tc.Types.Evidence
go (WpTyApp ty) k = k $ \e -> App e (Type ty)
go (WpEvLam ev) k = k $ Lam ev
go (WpTyLam tv) k = k $ Lam tv
@@ -1612,13 +1616,13 @@ ds_hs_wrapper wrap = go wrap
go (WpCompose c1 c2) k = go c1 $ \w1 ->
go c2 $ \w2 ->
k (w1 . w2)
- go (WpFun c1 c2 st) k = -- See Note [Desugaring WpFun]
- do { x <- newSysLocalDs st
- ; go c1 $ \w1 ->
- go c2 $ \w2 ->
- let app f a = mkCoreApp (text "dsHsWrapper") f a
- arg = w1 (Var x)
- in k (\e -> (Lam x (w2 (app e arg)))) }
+ go (WpFun c1 c2 st _) k = -- See Note [Desugaring WpFun]
+ do { x <- newSysLocalDs st
+ ; go c1 $ \w1 ->
+ go c2 $ \w2 ->
+ let app f a = mkCoreApp (text "dsHsWrapper") f a
+ arg = w1 (Var x)
+ in k (\e -> (Lam x (w2 (app e arg)))) }
--------------------------------------
dsTcEvBinds_s :: [TcEvBinds] -> ([CoreBind] -> DsM a) -> DsM a
=====================================
compiler/GHC/HsToCore/Match.hs
=====================================
@@ -1240,7 +1240,7 @@ viewLExprEq (e1,_) (e2,_) = lexp e1 e2
-- equating different ways of writing a coercion)
wrap WpHole WpHole = True
wrap (WpCompose w1 w2) (WpCompose w1' w2') = wrap w1 w1' && wrap w2 w2'
- wrap (WpFun w1 w2 _) (WpFun w1' w2' _) = wrap w1 w1' && wrap w2 w2'
+ wrap (WpFun w1 w2 _ _) (WpFun w1' w2' _ _) = wrap w1 w1' && wrap w2 w2'
wrap (WpCast co) (WpCast co') = co `eqCoercion` co'
wrap (WpEvApp et1) (WpEvApp et2) = et1 `ev_term` et2
wrap (WpTyApp t) (WpTyApp t') = eqType t t'
=====================================
compiler/GHC/Iface/Ext/Ast.hs
=====================================
@@ -696,7 +696,7 @@ instance ToHie (LocatedA HsWrapper) where
(WpLet bs) -> toHie $ EvBindContext (mkScope osp) (getRealSpanA osp) (L osp bs)
(WpCompose a b) -> concatM $
[toHie (L osp a), toHie (L osp b)]
- (WpFun a b _) -> concatM $
+ (WpFun a b _ _) -> concatM $
[toHie (L osp a), toHie (L osp b)]
(WpEvLam a) ->
toHie $ C (EvidenceVarBind EvWrapperBind (mkScope osp) (getRealSpanA osp))
=====================================
compiler/GHC/Tc/Errors/Hole.hs
=====================================
@@ -823,9 +823,11 @@ tcFilterHoleFits limit typed_hole ht@(hole_ty, _) candidates =
unfoldWrapper :: HsWrapper -> [Type]
unfoldWrapper = reverse . unfWrp'
- where unfWrp' (WpTyApp ty) = [ty]
- unfWrp' (WpCompose w1 w2) = unfWrp' w1 ++ unfWrp' w2
- unfWrp' _ = []
+ where
+ unfWrp' (WpTyApp ty) = [ty]
+ unfWrp' (WpSubType w) = unfWrp' w
+ unfWrp' (WpCompose w1 w2) = unfWrp' w1 ++ unfWrp' w2
+ unfWrp' _ = []
-- The real work happens here, where we invoke the type checker using
=====================================
compiler/GHC/Tc/Gen/App.hs
=====================================
@@ -794,7 +794,7 @@ tcInstFun do_ql inst_final (tc_fun, fun_ctxt) fun_sigma rn_args
= do { let herald = case fun_ctxt of
VAExpansion (OrigStmt{}) _ _ -> ExpectedFunTySyntaxOp DoOrigin tc_fun
_ -> ExpectedFunTyArg (HsExprTcThing tc_fun) (unLoc arg)
- ; (wrap, arg_ty, res_ty) <-
+ ; (fun_co, arg_ty, res_ty) <-
-- NB: matchActualFunTy does the rep-poly check.
-- For example, suppose we have f :: forall r (a::TYPE r). a -> Int
-- In an application (f x), we need 'x' to have a fixed runtime
@@ -805,7 +805,7 @@ tcInstFun do_ql inst_final (tc_fun, fun_ctxt) fun_sigma rn_args
(n_val_args, fun_sigma) fun_ty
; arg' <- quickLookArg do_ql ctxt arg arg_ty
- ; let acc' = arg' : addArgWrap wrap acc
+ ; let acc' = arg' : addArgWrap (mkWpCastN fun_co) acc
; go (pos+1) acc' res_ty rest_args }
new_arg_ty :: LHsExpr GhcRn -> Int -> TcM (Scaled TcType)
=====================================
compiler/GHC/Tc/Gen/Head.hs
=====================================
@@ -765,13 +765,13 @@ tcInferOverLit lit@(OverLit { ol_val = val
thing = NameThing from_name
mb_thing = Just thing
herald = ExpectedFunTyArg thing (HsLit noExtField hs_lit)
- ; (wrap2, sarg_ty, res_ty) <- matchActualFunTy herald mb_thing (1, from_ty) from_ty
+ ; (co2, sarg_ty, res_ty) <- matchActualFunTy herald mb_thing (1, from_ty) from_ty
; co <- unifyType mb_thing (hsLitType hs_lit) (scaledThing sarg_ty)
-- See Note [Source locations for implicit function calls] in GHC.Iface.Ext.Ast
; let lit_expr = L (l2l loc) $ mkHsWrapCo co $
HsLit noExtField hs_lit
- from_expr = mkHsWrap (wrap2 <.> wrap1) $
+ from_expr = mkHsWrap (mkWpCastN co2 <.> wrap1) $
mkHsVar (L loc from_id)
witness = HsApp noExtField (L (l2l loc) from_expr) lit_expr
lit' = OverLit { ol_val = val
=====================================
compiler/GHC/Tc/Gen/Pat.hs
=====================================
@@ -699,7 +699,7 @@ tc_pat pat_ty penv ps_pat thing_inside = case ps_pat of
-- Expression must be a function
; let herald = ExpectedFunTyViewPat $ unLoc expr
- ; (expr_wrap1, Scaled _mult inf_arg_ty, inf_res_sigma)
+ ; (expr_co1, Scaled _mult inf_arg_ty, inf_res_sigma)
<- matchActualFunTy herald (Just . HsExprRnThing $ unLoc expr) (1,expr_rho) expr_rho
-- See Note [View patterns and polymorphism]
-- expr_wrap1 :: expr_rho "->" (inf_arg_ty -> inf_res_sigma)
@@ -720,7 +720,7 @@ tc_pat pat_ty penv ps_pat thing_inside = case ps_pat of
-- NB: pat_ty comes from matchActualFunTy, so it has a
-- fixed RuntimeRep, as needed to call mkWpFun.
- expr_wrap = expr_wrap2' <.> expr_wrap1
+ expr_wrap = expr_wrap2' <.> mkWpCastN expr_co1
; return $ (ViewPat pat_ty (mkLHsWrap expr_wrap expr') pat', res) }
=====================================
compiler/GHC/Tc/Types/Evidence.hs
=====================================
@@ -8,10 +8,11 @@ module GHC.Tc.Types.Evidence (
-- * HsWrapper
HsWrapper(..),
(<.>), mkWpTyApps, mkWpEvApps, mkWpEvVarApps, mkWpTyLams, mkWpForAllCast,
- mkWpEvLams, mkWpLet, mkWpFun, mkWpCastN, mkWpCastR, mkWpEta,
+ mkWpEvLams, mkWpLet, mkWpFun, mkWpCastN, mkWpCastR, mkWpEta, mkWpSubType,
collectHsWrapBinders,
idHsWrapper, isIdHsWrapper,
pprHsWrapper, hsWrapDictBinders,
+ optSubTypeHsWrapper,
-- * Evidence bindings
TcEvBinds(..), EvBindsVar(..),
@@ -73,7 +74,7 @@ import GHC.Types.Unique.DFM
import GHC.Types.Unique.FM
import GHC.Types.Name( isInternalName )
import GHC.Types.Var
-import GHC.Types.Id( idScaledType )
+import GHC.Types.Id( idScaledType, idType )
import GHC.Types.Var.Env
import GHC.Types.Var.Set
import GHC.Types.Basic
@@ -134,35 +135,128 @@ maybeSymCo NotSwapped co = co
************************************************************************
-}
--- We write wrap :: t1 ~> t2
--- if wrap[ e::t1 ] :: t2
+{- Note [Deep subsumption and WpSubType]
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+When making DeepSubsumption checks, we may end up with hard-to-spot identity wrappers.
+For example (#26349) suppose we have
+ (forall a. Eq a => a->a) -> Int <= (forall a. Eq a => a->a) -> Int
+The two types are equal so we should certainly get an identity wrapper. But we'll get
+tihs wrapper from `tcSubType`:
+ WpFun (WpTyLam a <.> WpEvLam dg <.> WpLet (dw=dg) <.> WpEvApp dw <.> WpTyApp a)
+ WpHole
+That elaborate wrapper is really just a no-op, but it's far from obvious. If we just
+desugar (HsWrap f wp) straightforwardly we'll get
+ \(g:forall a. Eq a => a -> a).
+ f (/\a. \(dg:Eq a). let dw=dg in g a dw)
+
+To recognise that as just `f`, we'd have to eta-reduce twice. But eta-reduction
+is not sound in general, so we'll end up retaining the lambdas. Two bad results:
+
+* Adding DeepSubsumption gratuitiously makes programs less efficient.
+
+* When the subsumption is on the LHS of a rule, or in a SPECIALISE pragma, we
+ may not be able to make a decent RULE at all, and will fail with "LHS of rule
+ is too complicated to desugar" (#26255)
+
+It'd be ideal to solve the problem at the source, by never generating those
+gruesome wrappers in the first place, but we can't do that because:
+
+* The WpTyLam and WpTyApp are introduced independently, not together, in `tcSubType`,
+ so we can't easily cancel them out. For example, even if we have
+ forall a. t1 <= forall a. t2
+ there is no guarantee that these are the "same" a. E.g.
+ forall a b. a -> b -> b <= forall x y. y -> x -> x
+ Similarly WpEvLam and WpEvApp
+
+* We have not yet done constraint solving so we don't know what evidence will
+ end up in those WpLet bindings.
+
+TL;DR we must generate the wrapper and then optimise it way if it turns out
+that it is a no-op. Here's our solution:
+
+(DSST1) Tag the wrappers generated from a subtype check with WpSubType. In normal
+ wrappers the binders of a WpTyLam or WpEvLam can scope over the "hole" of the
+ wrapper -- that is how we introduce type-lambdas and dictionary-lambda into the
+ terms! But in /subtype/ wrappers, these type/dictionary lambdas only scope over
+ the WpTyApp and WpEvApp nodes in the /same/ wrapper. That is what justifies us
+ eta-reducing the type/dictionary lambdas.
+
+ In short, (WpSubType wp) means the same as `wp`, but with the added promise that
+ the binders in `wp` do not scope over the hole.
+
+(DSST2) Avoid creating a WpSubType in the common WpHole case, using `mkWpSubType`.
+
+(DSST3) When desugaring, try eta-reduction on the payload of a WpSubType.
+ This is done in `GHC.HsToCore.Binds.dsHsWrapper` by the call to `optSubTypeHsWrapper`.
+
+ We don't attempt to optimise HsWrappers /other than/ subtype wrappers. Why not?
+ Because there aren't any useful optimsations we can do. (We could collapse
+ adjacent `WpCast`s perhaps, but that'll happen later automatically via `mkCast`.)
+
+ TL;DR:
+ * we /must/ optimise subtype-HsWrappers (that's the point of this Note!)
+ * there is little point in attempting to optimise any other HsWrappers
+
+Note [WpFun-RR-INVARIANT]
+~~~~~~~~~~~~~~~~~~~~~~~~~
+Given
+ wrap = WpFun wrap1 wrap2 sty1 ty2
+ where: wrap1 :: exp_arg ~~> act_arg
+ wrap2 :: act_res ~~> exp_res
+ wrap :: (act_arg -> act_res) ~~> (exp_arg -> exp_res)
+we have
+ WpFun-RR-INVARIANT:
+ the input (exp_arg) and output (act_arg) types of `wrap1`
+ both have a fixed runtime-rep
+
+Reason: We desugar wrap[e] into
+ \(x:exp_arg). wrap2[ e wrap1[x] ]
+And then, because of Note [Representation polymorphism invariants], we need:
+
+ * `exp_arg` must have a fixed runtime rep,
+ so that lambda obeys the the FRR rules
+
+ * `act_arg` must have a fixed runtime rep,
+ so the that application (e wrap1[x]) obeys the FRR tules
+
+Hence WpFun-INVARIANT.
+-}
+
data HsWrapper
+ -- NOTATION (~~>):
+ -- We write wrap :: t1 ~~> t2
+ -- if wrap[ e::t1 ] :: t2
= WpHole -- The identity coercion
+ | WpSubType HsWrapper
+ -- (WpSubType wp) is the same as `wp`, but with extra invariants
+ -- See Note [Deep subsumption and WpSubType] (DSST1)
+
| WpCompose HsWrapper HsWrapper
-- (wrap1 `WpCompose` wrap2)[e] = wrap1[ wrap2[ e ]]
--
-- Hence (\a. []) `WpCompose` (\b. []) = (\a b. [])
-- But ([] a) `WpCompose` ([] b) = ([] b a)
--
- -- If wrap1 :: t2 ~> t3
- -- wrap2 :: t1 ~> t2
- --- Then (wrap1 `WpCompose` wrap2) :: t1 ~> t3
-
- | WpFun HsWrapper HsWrapper (Scaled TcTypeFRR)
- -- (WpFun wrap1 wrap2 (w, t1))[e] = \(x:_w exp_arg). wrap2[ e wrap1[x] ]
- -- So note that if e :: act_arg -> act_res
- -- wrap1 :: exp_arg ~> act_arg
- -- wrap2 :: act_res ~> exp_res
- -- then WpFun wrap1 wrap2 : (act_arg -> arg_res) ~> (exp_arg -> exp_res)
+ -- If wrap1 :: t2 ~~> t3
+ -- wrap2 :: t1 ~~> t2
+ --- Then (wrap1 `WpCompose` wrap2) :: t1 ~~> t3
+
+ | WpFun HsWrapper HsWrapper (Scaled TcTypeFRR) TcType
+ -- (WpFun wrap1 wrap2 (w, t1) t2)[e] = \(x:_w exp_arg). wrap2[ e wrap1[x] ]
+ --
+ -- INVARIANT: both input and output types of `wrap1` have a fixed runtime-rep
+ -- See Note [WpFun-RR-INVARIANT]
+ --
+ -- Typing rules:
+ -- If e :: act_arg -> act_res
+ -- wrap1 :: exp_arg ~~> act_arg
+ -- wrap2 :: act_res ~~> exp_res
+ -- then WpFun wrap1 wrap2 :: (act_arg -> act_res) ~~> (exp_arg -> exp_res)
-- This isn't the same as for mkFunCo, but it has to be this way
-- because we can't use 'sym' to flip around these HsWrappers
- -- The TcType is the "from" type of the first wrapper;
- -- it always a Type, not a Constraint
--
- -- NB: a WpFun is always for a (->) function arrow
- --
- -- Use 'mkWpFun' to construct such a wrapper.
+ -- NB: a WpFun is always for a (->) function arrow, never (=>)
| WpCast TcCoercionR -- A cast: [] `cast` co
-- Guaranteed not the identity coercion
@@ -212,50 +306,48 @@ WpCast c1 <.> WpCast c2 = WpCast (c2 `mkTransCo` c1)
--
-- NB: <.> behaves like function composition:
--
- -- WpCast c1 <.> WpCast c2 :: coercionLKind c2 ~> coercionRKind c1
+ -- WpCast c1 <.> WpCast c2 :: coercionLKind c2 ~~> coercionRKind c1
--
-- This is thus the same as WpCast (c2 ; c1) and not WpCast (c1 ; c2).
c1 <.> c2 = c1 `WpCompose` c2
--- | Smart constructor to create a 'WpFun' 'HsWrapper', which avoids introducing
--- a lambda abstraction if the two supplied wrappers are either identities or
--- casts.
---
--- PRECONDITION: either:
---
--- 1. both of the 'HsWrapper's are identities or casts, or
--- 2. both the "from" and "to" types of the first wrapper have a syntactically
--- fixed RuntimeRep (see Note [Fixed RuntimeRep] in GHC.Tc.Utils.Concrete).
mkWpFun :: HsWrapper -> HsWrapper
-> Scaled TcTypeFRR -- ^ the "from" type of the first wrapper
-> TcType -- ^ Either "from" type or "to" type of the second wrapper
-- (used only when the second wrapper is the identity)
-> HsWrapper
-mkWpFun WpHole WpHole _ _ = WpHole
-mkWpFun WpHole (WpCast co2) (Scaled w t1) _ = WpCast (mk_wp_fun_co w (mkRepReflCo t1) co2)
-mkWpFun (WpCast co1) WpHole (Scaled w _) t2 = WpCast (mk_wp_fun_co w (mkSymCo co1) (mkRepReflCo t2))
-mkWpFun (WpCast co1) (WpCast co2) (Scaled w _) _ = WpCast (mk_wp_fun_co w (mkSymCo co1) co2)
-mkWpFun w_arg w_res t1 _ =
- -- In this case, we will desugar to a lambda
- --
- -- \x. w_res[ e w_arg[x] ]
- --
- -- To satisfy Note [Representation polymorphism invariants] in GHC.Core,
- -- it must be the case that both the lambda bound variable x and the function
- -- argument w_arg[x] have a fixed runtime representation, i.e. that both the
- -- "from" and "to" types of the first wrapper "w_arg" have a fixed runtime representation.
- --
- -- Unfortunately, we can't check this with an assertion here, because of
- -- [Wrinkle: Typed Template Haskell] in Note [hasFixedRuntimeRep] in GHC.Tc.Utils.Concrete.
- WpFun w_arg w_res t1
-
-mkWpEta :: [Id] -> HsWrapper -> HsWrapper
+-- ^ Smart constructor for `WpFun`
+-- Just removes clutter and optimises some common cases.
+--
+-- PRECONDITION: same as Note [WpFun-RR-INVARIANT]
+--
+-- Unfortunately, we can't check PRECONDITION with an assertion here, because of
+-- [Wrinkle: Typed Template Haskell] in Note [hasFixedRuntimeRep] in GHC.Tc.Utils.Concrete.
+mkWpFun w1 w2 st1@(Scaled m1 t1) t2
+ = case (w1,w2) of
+ (WpHole, WpHole) -> WpHole
+ (WpHole, WpCast co2) -> WpCast (mk_wp_fun_co m1 (mkRepReflCo t1) co2)
+ (WpCast co1, WpHole) -> WpCast (mk_wp_fun_co m1 (mkSymCo co1) (mkRepReflCo t2))
+ (WpCast co1, WpCast co2) -> WpCast (mk_wp_fun_co m1 (mkSymCo co1) co2)
+ (_, _) -> WpFun w1 w2 st1 t2
+
+mkWpSubType :: HsWrapper -> HsWrapper
+-- See (DSST2) in Note [Deep subsumption and WpSubType]
+mkWpSubType WpHole = WpHole
+mkWpSubType (WpCast co) = WpCast co
+mkWpSubType w = WpSubType w
+
+mkWpEta :: Type -> [Id] -> HsWrapper -> HsWrapper
-- (mkWpEta [x1, x2] wrap) [e]
-- = \x1. \x2. wrap[e x1 x2]
-- Just generates a bunch of WpFuns
-mkWpEta xs wrap = foldr eta_one wrap xs
+-- The incoming type is the type of the entire expression
+mkWpEta orig_fun_ty xs wrap = go orig_fun_ty xs
where
- eta_one x wrap = WpFun idHsWrapper wrap (idScaledType x)
+ go _ [] = wrap
+ go fun_ty (id:ids) = WpFun idHsWrapper (go res_ty ids) (idScaledType id) res_ty
+ where
+ res_ty = funResultTy fun_ty
mk_wp_fun_co :: Mult -> TcCoercionR -> TcCoercionR -> TcCoercionR
mk_wp_fun_co mult arg_co res_co
@@ -333,8 +425,9 @@ hsWrapDictBinders wrap = go wrap
where
go (WpEvLam dict_id) = unitBag dict_id
go (w1 `WpCompose` w2) = go w1 `unionBags` go w2
- go (WpFun _ w _) = go w
+ go (WpFun _ w _ _) = go w
go WpHole = emptyBag
+ go (WpSubType {}) = emptyBag -- See Note [Deep subsumption and WpSubType]
go (WpCast {}) = emptyBag
go (WpEvApp {}) = emptyBag
go (WpTyLam {}) = emptyBag
@@ -350,6 +443,7 @@ collectHsWrapBinders wrap = go wrap []
go :: HsWrapper -> [HsWrapper] -> ([Var], HsWrapper)
go (WpEvLam v) wraps = add_lam v (gos wraps)
go (WpTyLam v) wraps = add_lam v (gos wraps)
+ go (WpSubType w) wraps = go w wraps
go (WpCompose w1 w2) wraps = go w1 (w2:wraps)
go wrap wraps = ([], foldl' (<.>) wrap wraps)
@@ -358,6 +452,162 @@ collectHsWrapBinders wrap = go wrap []
add_lam v (vs,w) = (v:vs, w)
+
+optSubTypeHsWrapper :: HsWrapper -> HsWrapper
+-- This optimiser is used only on the payload of WpSubType
+-- It finds cases where the entire wrapper is a no-op
+-- See (DSST3) in Note [Deep subsumption and WpSubType]
+optSubTypeHsWrapper wrap
+ = opt wrap
+ where
+ opt :: HsWrapper -> HsWrapper
+ opt w = foldr (<.>) WpHole (opt1 w [])
+
+ opt1 :: HsWrapper -> [HsWrapper] -> [HsWrapper]
+ -- opt1 w ws = w <.> (foldr <.> WpHole ws)
+ -- INVARIANT: ws::[HsWrapper] is optimised
+ opt1 WpHole ws = ws
+ opt1 (WpSubType w) ws = opt1 w ws
+ opt1 (w1 `WpCompose` w2) ws = opt1 w1 (opt1 w2 ws)
+ opt1 (WpCast co) ws = opt_co co ws
+ opt1 (WpEvLam ev) ws = opt_ev_lam ev ws
+ opt1 (WpTyLam tv) ws = opt_ty_lam tv ws
+ opt1 (WpLet binds) ws = pushWpLet binds ws
+ opt1 (WpFun w1 w2 sty1 ty2) ws = opt_fun w1 w2 sty1 ty2 ws
+ opt1 w@(WpTyApp {}) ws = w : ws
+ opt1 w@(WpEvApp {}) ws = w : ws
+
+ -----------------
+ -- (WpTyLam a <.> WpTyApp a <.> w) = w
+ -- i.e. /\a. <hole> a --> <hole>
+ -- This is only valid if whatever fills the hole does not mention 'a'
+ -- But that's guaranteed in subtype-wrappers;
+ -- see (DSST1) in Note [Deep subsumption and WpSubType]
+ opt_ty_lam tv (WpTyApp ty : ws)
+ | Just tv' <- getTyVar_maybe ty
+ , tv==tv'
+ , all (tv `not_in`) ws
+ = ws
+
+ -- (WpTyLam a <.> WpCastCo co <.> w)
+ -- = WpCast (ForAllCo a co) (WpTyLam <.> w)
+ opt_ty_lam tv (WpCast co : ws)
+ = opt_co (mkHomoForAllCo tv co) (opt_ty_lam tv ws)
+
+ opt_ty_lam tv ws
+ = WpTyLam tv : ws
+
+ -----------------
+ -- (WpEvLam ev <.> WpEvAp ev <.> w) = w
+ -- Similar notes to WpTyLam
+ opt_ev_lam ev (WpEvApp ev_tm : ws)
+ | EvExpr (Var ev') <- ev_tm
+ , ev == ev'
+ , all (ev `not_in`) ws
+ = ws
+
+ -- (WpEvLam ev <.> WpCast co <.> w)
+ -- = WpCast (FunCo ev co) (WpEvLam <.> w)
+ opt_ev_lam ev (WpCast co : ws)
+ = opt_co fun_co (opt_ev_lam ev ws)
+ where
+ fun_co = mkFunCo Representational FTF_C_T
+ (mkNomReflCo ManyTy)
+ (mkRepReflCo (idType ev))
+ co
+
+ opt_ev_lam ev ws
+ = WpEvLam ev : ws
+
+ -----------------
+ -- WpCast co <.> WpCast co' <.> ws = WpCast (co;co') ws
+ opt_co co (WpCast co' : ws) = opt_co (co `mkTransCo` co') ws
+ opt_co co ws | isReflexiveCo co = ws
+ | otherwise = WpCast co : ws
+
+ ------------------
+ opt_fun w1 w2 sty1 ty2 ws
+ = case mkWpFun (opt w1) (opt w2) sty1 ty2 of
+ WpHole -> ws
+ WpCast co -> opt_co co ws
+ w -> w : ws
+
+ ------------------
+ -- Tiresome check that the lambda-bound type/evidence variable that we
+ -- want to eta-reduce isn't free in the rest of the wrapper
+ not_in :: TyVar -> HsWrapper -> Bool
+ not_in _ WpHole = True
+ not_in v (WpCast co) = not (anyFreeVarsOfCo (== v) co)
+ not_in v (WpTyApp ty) = not (anyFreeVarsOfType (== v) ty)
+ not_in v (WpFun w1 w2 _ _) = not_in v w1 && not_in v w2
+ not_in v (WpSubType w) = not_in v w
+ not_in v (WpCompose w1 w2) = not_in v w1 && not_in v w2
+ not_in v (WpEvApp (EvExpr e)) = not (v `elemVarSet` exprFreeVars e)
+ not_in _ (WpEvApp (EvTypeable {})) = False -- Giving up; conservative
+ not_in _ (WpEvApp (EvFun {})) = False -- Giving up; conservative
+ not_in _ (WpTyLam {}) = False -- Give up; conservative
+ not_in _ (WpEvLam {}) = False -- Ditto
+ not_in _ (WpLet {}) = False -- Ditto
+
+pushWpLet :: TcEvBinds -> [HsWrapper] -> [HsWrapper]
+-- See if we can transform
+-- WpLet binds <.> w1 <.> .. <.> wn --> w1' <.> .. <.> wn'
+-- by substitution.
+-- We do this just for the narrow case when
+-- - the `binds` are all just v=w, variables only
+-- - the wi are all WpTyApp, WpEvApp, or WpCast
+-- This is just enough to get us the eta-reductions that we seek
+pushWpLet tc_ev_binds ws
+ = case tc_ev_binds of
+ TcEvBinds {} -> pprPanic "pushWpLet" (ppr tc_ev_binds)
+ EvBinds binds
+ | isEmptyBag binds
+ -> ws
+ | Just env <- ev_bind_swizzle binds
+ -> case go env ws of
+ Just ws' -> ws'
+ Nothing -> bale_out
+ | otherwise
+ -> bale_out
+ where
+ bale_out = WpLet tc_ev_binds : ws
+
+ go :: IdEnv Id -> [HsWrapper] -> Maybe [HsWrapper]
+ go env (WpCast co : ws) = do { ws' <- go env ws
+ ; return (WpCast co : ws') }
+ go env (WpTyApp ty : ws) = do { ws' <- go env ws
+ ; return (WpTyApp ty : ws') }
+ go env (WpEvApp (EvExpr (Var v)) : ws)
+ = do { v' <- swizzle_id env v
+ ; ws' <- go env ws
+ ; return (WpEvApp (EvExpr (Var v')) : ws') }
+
+ go _ ws = case ws of
+ [] -> Just []
+ (_:_) -> Nothing -- Could not fully eliminate the WpLet
+
+ swizzle_id :: IdEnv Id -> Id -> Maybe Id
+ -- Nothing <=> ran out of fuel
+ -- This is just belt and braces; we should never build bottom evidence
+ swizzle_id env v = go 100 v
+ where
+ go :: Int -> EvId -> Maybe EvId
+ go fuel v
+ | fuel == 0 = Nothing
+ | Just v' <- lookupVarEnv env v = go (fuel-1) v'
+ | otherwise = Just v
+
+ ev_bind_swizzle :: Bag EvBind -> Maybe (IdEnv Id)
+ -- Succeeds only if the bindings are all var-to-var bindings
+ ev_bind_swizzle evbs = foldl' do_one (Just emptyVarEnv) evbs
+ where
+ do_one :: Maybe (IdEnv Id) -> EvBind -> Maybe (IdEnv Id)
+ do_one Nothing _ = Nothing
+ do_one (Just swizzle) (EvBind {eb_lhs = bndr, eb_rhs = rhs})
+ = case rhs of
+ EvExpr (Var v) -> Just (extendVarEnv swizzle bndr v)
+ _ -> Nothing
+
{-
************************************************************************
* *
@@ -1018,8 +1268,9 @@ pprHsWrapper wrap pp_thing_inside
-- True <=> appears in function application position
-- False <=> appears as body of let or lambda
help it WpHole = it
- help it (WpCompose f1 f2) = help (help it f2) f1
- help it (WpFun f1 f2 (Scaled w t1)) = add_parens $ text "\\(x" <> dcolon <> brackets (ppr w) <> ppr t1 <> text ")." <+>
+ help it (WpCompose w1 w2) = help (help it w2) w1
+ help it (WpSubType w) = no_parens $ text "subtype" <> braces (help it w False)
+ help it (WpFun f1 f2 (Scaled w t1) _) = add_parens $ text "\\(x" <> dcolon <> brackets (ppr w) <> ppr t1 <> text ")." <+>
help (\_ -> it True <+> help (\_ -> text "x") f1 True) f2 False
help it (WpCast co) = add_parens $ sep [it False, nest 2 (text "|>"
<+> pprParendCo co)]
=====================================
compiler/GHC/Tc/Utils/Concrete.hs
=====================================
@@ -626,8 +626,12 @@ hasFixedRuntimeRep :: HasDebugCallStack
-- @ki@ is concrete, and @co :: ty ~# ty'@.
-- That is, @ty'@ has a syntactically fixed RuntimeRep
-- in the sense of Note [Fixed RuntimeRep].
-hasFixedRuntimeRep frr_ctxt ty =
- checkFRR_with (fmap (fmap coToMCo) . unifyConcrete_kind (fsLit "cx") . ConcreteFRR) frr_ctxt ty
+hasFixedRuntimeRep frr_ctxt ty
+ = checkFRR_with unify_conc frr_ctxt ty
+ where
+ unify_conc frr_orig ki
+ = do { co <- unifyConcrete_kind (fsLit "cx") (ConcreteFRR frr_orig) ki
+ ; return (coToMCo co) }
-- | Like 'hasFixedRuntimeRep', but we perform an eager syntactic check.
--
=====================================
compiler/GHC/Tc/Utils/Unify.hs
=====================================
@@ -148,7 +148,7 @@ matchActualFunTy
-- (Both are used only for error messages)
-> TcRhoType
-- ^ Type to analyse: a TcRhoType
- -> TcM (HsWrapper, Scaled TcSigmaTypeFRR, TcSigmaType)
+ -> TcM (TcCoercion, Scaled TcSigmaTypeFRR, TcSigmaType)
-- This function takes in a type to analyse (a RhoType) and returns
-- an argument type and a result type (splitting apart a function arrow).
-- The returned argument type is a SigmaType with a fixed RuntimeRep;
@@ -157,7 +157,7 @@ matchActualFunTy
-- See Note [matchActualFunTy error handling] for the first three arguments
-- If (wrap, arg_ty, res_ty) = matchActualFunTy ... fun_ty
--- then wrap :: fun_ty ~> (arg_ty -> res_ty)
+-- then wrap :: fun_ty ~~> (arg_ty -> res_ty)
-- and NB: res_ty is an (uninstantiated) SigmaType
matchActualFunTy herald mb_thing err_info fun_ty
@@ -172,13 +172,13 @@ matchActualFunTy herald mb_thing err_info fun_ty
-- hide the forall inside a meta-variable
go :: TcRhoType -- The type we're processing, perhaps after
-- expanding type synonyms
- -> TcM (HsWrapper, Scaled TcSigmaTypeFRR, TcSigmaType)
+ -> TcM (TcCoercion, Scaled TcSigmaTypeFRR, TcSigmaType)
go ty | Just ty' <- coreView ty = go ty'
go (FunTy { ft_af = af, ft_mult = w, ft_arg = arg_ty, ft_res = res_ty })
= assert (isVisibleFunArg af) $
do { hasFixedRuntimeRep_syntactic (FRRExpectedFunTy herald 1) arg_ty
- ; return (idHsWrapper, Scaled w arg_ty, res_ty) }
+ ; return (mkNomReflCo fun_ty, Scaled w arg_ty, res_ty) }
go ty@(TyVarTy tv)
| isMetaTyVar tv
@@ -210,7 +210,7 @@ matchActualFunTy herald mb_thing err_info fun_ty
; res_ty <- newOpenFlexiTyVarTy
; let unif_fun_ty = mkScaledFunTys [arg_ty] res_ty
; co <- unifyType mb_thing fun_ty unif_fun_ty
- ; return (mkWpCastN co, arg_ty, res_ty) }
+ ; return (co, arg_ty, res_ty) }
------------
mk_ctxt :: TcType -> TidyEnv -> ZonkM (TidyEnv, ErrCtxtMsg)
@@ -249,8 +249,10 @@ matchActualFunTys :: ExpectedFunTyOrigin -- ^ See Note [Herald for matchExpected
-> Arity
-> TcSigmaType
-> TcM (HsWrapper, [Scaled TcSigmaTypeFRR], TcRhoType)
--- If matchActualFunTys n ty = (wrap, [t1,..,tn], res_ty)
--- then wrap : ty ~> (t1 -> ... -> tn -> res_ty)
+-- NB: Called only from `tcSynArgA`, and hence scheduled for destruction
+--
+-- If matchActualFunTys n fun_ty = (wrap, [t1,..,tn], res_ty)
+-- then wrap : fun_ty ~~> (t1 -> ... -> tn -> res_ty)
-- and res_ty is a RhoType
-- NB: the returned type is top-instantiated; it's a RhoType
matchActualFunTys herald ct_orig n_val_args_wanted top_ty
@@ -265,15 +267,13 @@ matchActualFunTys herald ct_orig n_val_args_wanted top_ty
go 0 _ fun_ty = return (idHsWrapper, [], fun_ty)
go n so_far fun_ty
- = do { (wrap_fun1, arg_ty1, res_ty1) <- matchActualFunTy
- herald Nothing
- (n_val_args_wanted, top_ty)
- fun_ty
- ; (wrap_res, arg_tys, res_ty) <- go (n-1) (arg_ty1:so_far) res_ty1
+ = do { (co1, arg_ty1, res_ty1) <- matchActualFunTy herald Nothing
+ (n_val_args_wanted, top_ty) fun_ty
+ ; (wrap_res, arg_tys, res_ty) <- go (n-1) (arg_ty1:so_far) res_ty1
; let wrap_fun2 = mkWpFun idHsWrapper wrap_res arg_ty1 res_ty
-- NB: arg_ty1 comes from matchActualFunTy, so it has
- -- a syntactically fixed RuntimeRep as needed to call mkWpFun.
- ; return (wrap_fun2 <.> wrap_fun1, arg_ty1:arg_tys, res_ty) }
+ -- a syntactically fixed RuntimeRep
+ ; return (wrap_fun2 <.> mkWpCastN co1, arg_ty1:arg_tys, res_ty) }
{-
************************************************************************
@@ -459,7 +459,7 @@ tcSkolemiseGeneral ds_flag ctxt top_ty expected_ty thing_inside
tcSkolemiseCompleteSig :: TcCompleteSig
-> ([ExpPatType] -> TcRhoType -> TcM result)
-> TcM (HsWrapper, result)
--- ^ The wrapper has type: spec_ty ~> expected_ty
+-- ^ The wrapper has type: spec_ty ~~> expected_ty
-- See Note [Skolemisation] for the differences between
-- tcSkolemiseCompleteSig and tcTopSkolemise
@@ -790,7 +790,7 @@ matchExpectedFunTys :: forall a.
-> ([ExpPatType] -> ExpRhoType -> TcM a)
-> TcM (HsWrapper, a)
-- If matchExpectedFunTys n ty = (wrap, _)
--- then wrap : (t1 -> ... -> tn -> ty_r) ~> ty,
+-- then wrap : (t1 -> ... -> tn -> ty_r) ~~> ty,
-- where [t1, ..., tn], ty_r are passed to the thing_inside
--
-- Unconditionally concludes by skolemising any trailing invisible
@@ -865,12 +865,13 @@ matchExpectedFunTys herald ctx arity (Check top_ty) thing_inside
, ft_arg = arg_ty, ft_res = res_ty })
= assert (isVisibleFunArg af) $
do { let arg_pos = arity - n_req + 1 -- 1 for the first argument etc
- ; (arg_co, arg_ty) <- hasFixedRuntimeRep (FRRExpectedFunTy herald arg_pos) arg_ty
+ ; (arg_co, arg_ty_frr) <- hasFixedRuntimeRep (FRRExpectedFunTy herald arg_pos) arg_ty
+ ; let arg_sty_frr = Scaled mult arg_ty_frr
; (wrap_res, result) <- check (n_req - 1)
- (mkCheckExpFunPatTy (Scaled mult arg_ty) : rev_pat_tys)
+ (mkCheckExpFunPatTy arg_sty_frr : rev_pat_tys)
res_ty
; let wrap_arg = mkWpCastN arg_co
- fun_wrap = mkWpFun wrap_arg wrap_res (Scaled mult arg_ty) res_ty
+ fun_wrap = mkWpFun wrap_arg wrap_res arg_sty_frr res_ty
; return (fun_wrap, result) }
----------------------------
@@ -1407,7 +1408,7 @@ tcSubTypePat :: CtOrigin -> UserTypeCtxt
-- Used in patterns; polarity is backwards compared
-- to tcSubType
-- If wrap = tc_sub_type_et t1 t2
--- => wrap :: t1 ~> t2
+-- => wrap :: t1 ~~> t2
tcSubTypePat inst_orig ctxt (Check ty_actual) ty_expected
= tc_sub_type unifyTypeET inst_orig ctxt ty_actual ty_expected
@@ -1427,11 +1428,12 @@ tcSubTypeDS :: HsExpr GhcRn
-- DeepSubsumption <=> when checking, this type
-- is deeply skolemised
-> TcM HsWrapper
--- Only one call site, in GHC.Tc.Gen.App.tcApp
+-- Only one call site, in GHC.Tc.Gen.App.checkResultTy
tcSubTypeDS rn_expr act_rho exp_rho
- = tc_sub_type_deep Top (unifyExprType rn_expr) orig GenSigCtxt act_rho exp_rho
- where
- orig = exprCtOrigin rn_expr
+ = do { wrap <- tc_sub_type_deep Top (unifyExprType rn_expr)
+ (exprCtOrigin rn_expr)
+ GenSigCtxt act_rho exp_rho
+ ; return (mkWpSubType wrap) }
---------------
@@ -1456,7 +1458,7 @@ tcSubTypeSigma :: CtOrigin -- where did the actual type arise / why are we
-> TcSigmaType -> TcSigmaType -> TcM HsWrapper
-- External entry point, but no ExpTypes on either side
-- Checks that actual <= expected
--- Returns HsWrapper :: actual ~ expected
+-- Returns HsWrapper :: actual ~~> expected
tcSubTypeSigma orig ctxt ty_actual ty_expected
= tc_sub_type (unifyType Nothing) orig ctxt ty_actual ty_expected
@@ -1495,7 +1497,7 @@ tc_sub_type :: (TcType -> TcType -> TcM TcCoercionN) -- How to unify
-> TcM HsWrapper
-- Checks that actual_ty is more polymorphic than expected_ty
-- If wrap = tc_sub_type t1 t2
--- => wrap :: t1 ~> t2
+-- => wrap :: t1 ~~> t2
--
-- The "how to unify argument" is always a call to `uType TypeLevel orig`,
-- but with different ways of constructing the CtOrigin `orig` from
@@ -1504,7 +1506,8 @@ tc_sub_type :: (TcType -> TcType -> TcM TcCoercionN) -- How to unify
----------------------
tc_sub_type unify inst_orig ctxt ty_actual ty_expected
= do { ds_flag <- getDeepSubsumptionFlag
- ; tc_sub_type_ds Top ds_flag unify inst_orig ctxt ty_actual ty_expected }
+ ; wrap <- tc_sub_type_ds Top ds_flag unify inst_orig ctxt ty_actual ty_expected
+ ; return (mkWpSubType wrap) }
----------------------
tc_sub_type_ds :: Position p -- ^ position in the type (for error messages only)
@@ -1753,59 +1756,59 @@ we deal with function arrows. Suppose we have:
ty_actual = act_arg -> act_res
ty_expected = exp_arg -> exp_res
-To produce fun_wrap :: (act_arg -> act_res) ~> (exp_arg -> exp_res), we use
+To produce fun_wrap :: (act_arg -> act_res) ~~> (exp_arg -> exp_res), we use
the fact that the function arrow is contravariant in its argument type and
covariant in its result type. Thus we recursively perform subtype checks
on the argument types (with actual/expected switched) and the result types,
to get:
- arg_wrap :: exp_arg ~> act_arg -- NB: expected/actual have switched sides
- res_wrap :: act_res ~> exp_res
+ arg_wrap :: exp_arg ~~> act_arg -- NB: expected/actual have switched sides
+ res_wrap :: act_res ~~> exp_res
Then fun_wrap = mkWpFun arg_wrap res_wrap.
-Wrinkle [Representation-polymorphism checking during subtyping]
+Note [Representation-polymorphism checking during subtyping]
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+When doing deep subsumption in `tc_sub_type_deep`, looking under function arrows,
+we would usually build a `WpFun` HsWrapper. When desugared, we get eta-expansion:
- Inserting a WpFun HsWrapper amounts to impedance matching in deep subsumption
- via eta-expansion:
+ f ==> \(x :: exp_arg). res_wrap [ f (arg_wrap [x]) ]
- f ==> \ (x :: exp_arg) -> res_wrap [ f (arg_wrap [x]) ]
+Since we produce a lambda, we must enforce the representation polymorphism
+invariants described in Note [Representation polymorphism invariants] in GHC.Core.
+That is, we must ensure that both
+ - x (the lambda binder), and
+ - (arg_wrap [x]) (the function argument)
+have a fixed runtime representation.
- As we produce a lambda, we must enforce the representation polymorphism
- invariants described in Note [Representation polymorphism invariants] in GHC.Core.
- That is, we must ensure that both x (the lambda binder) and (arg_wrap [x]) (the function argument)
- have a fixed runtime representation.
+But we don't /always/ need to produce a `WpFun`: if both argument and result wrappers
+are merely coercions, we can produce a `WpCast co` instead of a `WpFun`. In that
+case there is no eta-expansion, and hence no need for FRR checks.
- Note however that desugaring mkWpFun does not always introduce a lambda: if
- both the argument and result HsWrappers are casts, then a FunCo cast suffices,
- in which case we should not perform representation-polymorphism checking.
+Here's a contrived example (there are undoubtedly more natural examples)
+(see testsuite/tests/rep-poly/NoEtaRequired):
- This means that, in the FunTy/FunTy case of tc_sub_type_deep, we can skip
- the representation-polymorphism checks if the produced argument and result
- wrappers are identities or casts.
- It is important to do so, otherwise we reject valid programs.
+ type Id :: k -> k
+ type family Id a where
- Here's a contrived example (there are undoubtedly more natural examples)
- (see testsuite/tests/rep-poly/NoEtaRequired):
+ type T :: TYPE r -> TYPE (Id r)
+ type family T a where
- type Id :: k -> k
- type family Id a where
+ test :: forall r (a :: TYPE r). a :~~: T a -> ()
+ test HRefl =
+ let
+ f :: (a -> a) -> ()
+ f _ = ()
+ g :: T a -> T a
+ g = undefined
+ in f g
- type T :: TYPE r -> TYPE (Id r)
- type family T a where
+We don't need to eta-expand `g` to make `f g` typecheck; a cast
+suffices. Hence we should not perform representation-polymorphism
+checks; they would fail here.
- test :: forall r (a :: TYPE r). a :~~: T a -> ()
- test HRefl =
- let
- f :: (a -> a) -> ()
- f _ = ()
- g :: T a -> T a
- g = undefined
- in f g
-
- We don't need to eta-expand `g` to make `f g` typecheck; a cast suffices.
- Hence we should not perform representation-polymorphism checks; they would
- fail here.
+All this is done by `mkWpFun_FRR`, which checks for the cast/cast case and
+returns a `FunCo` if so.
Note [Setting the argument context]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
@@ -1947,7 +1950,7 @@ getDeepSubsumptionFlag = do { ds <- xoptM LangExt.DeepSubsumption
-- | 'tc_sub_type_deep' is where the actual work happens for deep subsumption.
--
-- Given @ty_actual@ (a sigma-type) and @ty_expected@ (deeply skolemised, i.e.
--- a deep rho type), it returns an 'HsWrapper' @wrap :: ty_actual ~> ty_expected@.
+-- a deep rho type), it returns an 'HsWrapper' @wrap :: ty_actual ~~> ty_expected@.
tc_sub_type_deep :: HasDebugCallStack
=> Position p -- ^ Position in the type (for error messages only)
-> (TcType -> TcType -> TcM TcCoercionN) -- ^ How to unify
@@ -1958,7 +1961,7 @@ tc_sub_type_deep :: HasDebugCallStack
-> TcM HsWrapper
-- If wrap = tc_sub_type_deep t1 t2
--- => wrap :: t1 ~> t2
+-- => wrap :: t1 ~~> t2
-- Here is where the work actually happens!
-- Precondition: ty_expected is deeply skolemised
@@ -2015,8 +2018,8 @@ tc_sub_type_deep pos unify inst_orig ctxt ty_actual ty_expected
; unify_wrap <- just_unify exp_funTy ty_e
; fun_wrap <- go_fun af1 act_mult act_arg act_res af1 exp_mult exp_arg exp_res
; return $ unify_wrap <.> fun_wrap
- -- unify_wrap :: exp_funTy ~> ty_e
- -- fun_wrap :: ty_a ~> exp_funTy
+ -- unify_wrap :: exp_funTy ~~> ty_e
+ -- fun_wrap :: ty_a ~~> exp_funTy
}
go1 ty_a (FunTy { ft_af = af2, ft_mult = exp_mult, ft_arg = exp_arg, ft_res = exp_res })
| isVisibleFunArg af2
@@ -2028,8 +2031,8 @@ tc_sub_type_deep pos unify inst_orig ctxt ty_actual ty_expected
; unify_wrap <- just_unify ty_a act_funTy
; fun_wrap <- go_fun af2 act_mult act_arg act_res af2 exp_mult exp_arg exp_res
; return $ fun_wrap <.> unify_wrap
- -- unify_wrap :: ty_a ~> act_funTy
- -- fun_wrap :: act_funTy ~> ty_e
+ -- unify_wrap :: ty_a ~~> act_funTy
+ -- fun_wrap :: act_funTy ~~> ty_e
}
-- Otherwise, revert to unification.
@@ -2064,17 +2067,28 @@ mkWpFun_FRR
-> Position p
-> FunTyFlag -> Type -> TcType -> Type -- actual FunTy
-> FunTyFlag -> Type -> TcType -> Type -- expected FunTy
- -> HsWrapper -- ^ exp_arg ~> act_arg
- -> HsWrapper -- ^ act_res ~> exp_res
- -> TcM HsWrapper -- ^ act_funTy ~> exp_funTy
+ -> HsWrapper -- ^ exp_arg ~~> act_arg
+ -> HsWrapper -- ^ act_res ~~> exp_res
+ -> TcM HsWrapper -- ^ (act_arg->act_res) ~~> (exp_arg->exp_res)
mkWpFun_FRR unify pos act_af act_mult act_arg act_res exp_af exp_mult exp_arg exp_res arg_wrap res_wrap
- = do { ((exp_arg_co, exp_arg_frr), (act_arg_co, _act_arg_frr)) <-
- if needs_frr_checks
- -- See Wrinkle [Representation-polymorphism checking during subtyping]
- then do { exp_frr_wrap <- hasFixedRuntimeRep (frr_ctxt True ) exp_arg
- ; act_frr_wrap <- hasFixedRuntimeRep (frr_ctxt False) act_arg
- ; return (exp_frr_wrap, act_frr_wrap) }
- else return ((mkNomReflCo exp_arg, exp_arg), (mkNomReflCo act_arg, act_arg))
+ | Just arg_co <- getWpCo_maybe arg_wrap act_arg -- arg_co :: exp_arg ~R# act_arg
+ , Just res_co <- getWpCo_maybe res_wrap act_res -- res_co :: act_res ~R# exp_res
+ = -- The argument and result wrappers are both hole or cast;
+ -- so we can make do with a FunCo
+ -- See Note [Representation-polymorphism checking during subtyping]
+ do { mult_co <- unify act_mult exp_mult
+ ; let the_co = mkFunCo2 Representational act_af exp_af mult_co (mkSymCo arg_co) res_co
+ ; return (mkWpCastR the_co) }
+
+ | otherwise
+ = -- We need a full WpFun, with the eta-expansion that it entails
+ -- And hence we must add fixed-runtime-rep checks so that the eta-expansion is OK
+ -- See Note [Representation-polymorphism checking during subtyping]
+ do { (exp_arg_co, exp_arg_frr) <- hasFixedRuntimeRep (frr_ctxt True ) exp_arg
+ ; (act_arg_co, _act_arg_frr) <- hasFixedRuntimeRep (frr_ctxt False) act_arg
+ -- exp_arg_frr, act_arg_frr :: Type have fixed runtime-reps
+ -- exp_arg_co :: exp_arg ~ exp_arg_frr Usually Refl
+ -- act_arg_co :: act_arg ~ act_arg_frr Usually Refl
-- Enforce equality of multiplicities (not the more natural sub-multiplicity).
-- See Note [Multiplicity in deep subsumption]
@@ -2083,46 +2097,36 @@ mkWpFun_FRR unify pos act_af act_mult act_arg act_res exp_af exp_mult exp_arg ex
-- equality to be Refl, but it might well not be (#26332).
; let
- exp_arg_fun_co =
+ exp_arg_fun_co = -- (exp_arg_frr -> exp_res) ~ (exp_arg -> exp_res)
mkFunCo Nominal exp_af
- (mkReflCo Nominal exp_mult)
+ (mkNomReflCo exp_mult)
(mkSymCo exp_arg_co)
- (mkReflCo Nominal exp_res)
- act_arg_fun_co =
+ (mkNomReflCo exp_res)
+ act_arg_fun_co = -- (act_arg -> act_res) ~ (act_arg_frr -> act_res)
mkFunCo Nominal act_af
act_arg_mult_co
act_arg_co
- (mkReflCo Nominal act_res)
- arg_wrap_frr =
+ (mkNomReflCo act_res)
+ arg_wrap_frr = -- exp_arg_frr ~~> act_arg_frr
mkWpCastN (mkSymCo exp_arg_co) <.> arg_wrap <.> mkWpCastN act_arg_co
- -- exp_arg_co :: exp_arg ~> exp_arg_frr
- -- act_arg_co :: act_arg ~> act_arg_frr
- -- arg_wrap :: exp_arg ~> act_arg
- -- arg_wrap_frr :: exp_arg_frr ~> act_arg_frr
- ; return $
- mkWpCastN exp_arg_fun_co
+ ; return $ -- Whole thing :: (act_arg->act_res) ~~> (exp_arg->exp_ress)
+ mkWpCastN exp_arg_fun_co -- (exp_ar_frr->exp_res) ~~> (exp_arg->exp_res)
<.>
mkWpFun arg_wrap_frr res_wrap (Scaled exp_mult exp_arg_frr) exp_res
- <.>
- mkWpCastN act_arg_fun_co
+ <.> -- (act_arg_frr->act_res) ~~> (exp_arg_frr->exp_res)
+ mkWpCastN act_arg_fun_co -- (act_arg->act_res) ~~> (act_arg_frr->act_res)
}
where
- needs_frr_checks :: Bool
- needs_frr_checks =
- not (hole_or_cast arg_wrap)
- ||
- not (hole_or_cast res_wrap)
- hole_or_cast :: HsWrapper -> Bool
- hole_or_cast WpHole = True
- hole_or_cast (WpCast {}) = True
- hole_or_cast _ = False
+ getWpCo_maybe :: HsWrapper -> Type -> Maybe CoercionR
+ -- See if a HsWrapper is just a coercion
+ getWpCo_maybe WpHole ty = Just (mkRepReflCo ty)
+ getWpCo_maybe (WpCast co) _ = Just co
+ getWpCo_maybe _ _ = Nothing
+
frr_ctxt :: Bool -> FixedRuntimeRepContext
- frr_ctxt is_exp_ty =
- FRRDeepSubsumption
- { frrDSExpected = is_exp_ty
- , frrDSPosition = pos
- }
+ frr_ctxt is_exp_ty = FRRDeepSubsumption { frrDSExpected = is_exp_ty
+ , frrDSPosition = pos }
-----------------------
deeplySkolemise :: SkolemInfo -> TcSigmaType
@@ -2146,9 +2150,9 @@ deeplySkolemise skol_info ty
; let tvs = binderVars bndrs
tvs1 = binderVars bndrs1
tv_prs1 = map tyVarName tvs `zip` bndrs1
- ; return ( mkWpEta ids1 (mkWpTyLams tvs1
- <.> mkWpEvLams ev_vars1
- <.> wrap)
+ ; return ( mkWpEta ty ids1 (mkWpTyLams tvs1
+ <.> mkWpEvLams ev_vars1
+ <.> wrap)
, tv_prs1 ++ tvs_prs2
, ev_vars1 ++ ev_vars2
, mkScaledFunTys arg_tys' rho ) }
@@ -2182,7 +2186,7 @@ deeplyInstantiate orig ty
; ids1 <- newSysLocalIds (fsLit "di") arg_tys'
; wrap1 <- instCall orig (mkTyVarTys tvs') theta'
; (wrap2, rho2) <- go subst' rho
- ; return (mkWpEta ids1 (wrap2 <.> wrap1),
+ ; return (mkWpEta ty ids1 (wrap2 <.> wrap1),
mkScaledFunTys arg_tys' rho2) }
| otherwise
=====================================
compiler/GHC/Tc/Zonk/Type.hs
=====================================
@@ -1233,13 +1233,16 @@ zonk_cmd_top (HsCmdTop (CmdTopTc stack_tys ty ids) cmd)
-------------------------------------------------------------------------
zonkCoFn :: HsWrapper -> ZonkBndrTcM HsWrapper
zonkCoFn WpHole = return WpHole
+zonkCoFn (WpSubType w) = do { w' <- zonkCoFn w
+ ; return (WpSubType w') }
zonkCoFn (WpCompose c1 c2) = do { c1' <- zonkCoFn c1
; c2' <- zonkCoFn c2
; return (WpCompose c1' c2') }
-zonkCoFn (WpFun c1 c2 t1) = do { c1' <- zonkCoFn c1
- ; c2' <- zonkCoFn c2
- ; t1' <- noBinders $ zonkScaledTcTypeToTypeX t1
- ; return (WpFun c1' c2' t1') }
+zonkCoFn (WpFun c1 c2 t1 t2) = do { c1' <- zonkCoFn c1
+ ; c2' <- zonkCoFn c2
+ ; t1' <- noBinders $ zonkScaledTcTypeToTypeX t1
+ ; t2' <- noBinders $ zonkTcTypeToTypeX t2
+ ; return (WpFun c1' c2' t1' t2') }
zonkCoFn (WpCast co) = WpCast <$> noBinders (zonkCoToCo co)
zonkCoFn (WpEvLam ev) = WpEvLam <$> zonkEvBndrX ev
zonkCoFn (WpEvApp arg) = WpEvApp <$> noBinders (zonkEvTerm arg)
=====================================
compiler/Setup.hs
=====================================
@@ -1,4 +1,5 @@
{-# LANGUAGE NamedFieldPuns #-}
+{-# LANGUAGE CPP #-}
module Main where
import Distribution.Simple
@@ -12,6 +13,8 @@ import Distribution.Simple.Program
import Distribution.Simple.Utils
import Distribution.Simple.Setup
import Distribution.Simple.PackageIndex
+import qualified Distribution.Simple.LocalBuildInfo as LBI
+
import System.IO
import System.Process
@@ -59,8 +62,9 @@ primopIncls =
ghcAutogen :: Verbosity -> LocalBuildInfo -> IO ()
ghcAutogen verbosity lbi@LocalBuildInfo{pkgDescrFile,withPrograms,componentNameMap,installedPkgs}
= do
+ let i = LBI.interpretSymbolicPathLBI lbi
-- Get compiler/ root directory from the cabal file
- let Just compilerRoot = takeDirectory <$> pkgDescrFile
+ let Just compilerRoot = takeDirectory . i <$> pkgDescrFile
-- Require the necessary programs
(gcc ,withPrograms) <- requireProgram normal gccProgram withPrograms
@@ -80,15 +84,19 @@ ghcAutogen verbosity lbi@LocalBuildInfo{pkgDescrFile,withPrograms,componentNameM
-- Call genprimopcode to generate *.hs-incl
forM_ primopIncls $ \(file,command) -> do
contents <- readProcess "genprimopcode" [command] primopsStr
- rewriteFileEx verbosity (buildDir lbi </> file) contents
+ rewriteFileEx verbosity (i (buildDir lbi) </> file) contents
-- Write GHC.Platform.Constants
- let platformConstantsPath = autogenPackageModulesDir lbi </> "GHC/Platform/Constants.hs"
+ let platformConstantsPath = i (autogenPackageModulesDir lbi) </> "GHC/Platform/Constants.hs"
targetOS = case lookup "target os" settings of
Nothing -> error "no target os in settings"
Just os -> os
createDirectoryIfMissingVerbose verbosity True (takeDirectory platformConstantsPath)
+#if MIN_VERSION_Cabal(3,14,0)
+ withTempFile "Constants_tmp.hs" $ \tmp h -> do
+#else
withTempFile (takeDirectory platformConstantsPath) "Constants_tmp.hs" $ \tmp h -> do
+#endif
hClose h
callProcess "deriveConstants" ["--gen-haskell-type","-o",tmp,"--target-os",targetOS]
renameFile tmp platformConstantsPath
@@ -103,7 +111,7 @@ ghcAutogen verbosity lbi@LocalBuildInfo{pkgDescrFile,withPrograms,componentNameM
_ -> error "Couldn't find unique ghc-internal library when building ghc"
-- Write GHC.Settings.Config
- configHsPath = autogenPackageModulesDir lbi </> "GHC/Settings/Config.hs"
+ configHsPath = i (autogenPackageModulesDir lbi) </> "GHC/Settings/Config.hs"
configHs = generateConfigHs cProjectUnitId cGhcInternalUnitId settings
createDirectoryIfMissingVerbose verbosity True (takeDirectory configHsPath)
rewriteFileEx verbosity configHsPath configHs
=====================================
compiler/ghc.cabal.in
=====================================
@@ -50,7 +50,7 @@ extra-source-files:
custom-setup
- setup-depends: base >= 3 && < 5, Cabal >= 1.6 && <3.14, directory, process, filepath, containers
+ setup-depends: base >= 3 && < 5, Cabal >= 1.6 && <3.16, directory, process, filepath, containers
Flag internal-interpreter
Description: Build with internal interpreter support.
=====================================
libraries/ghc-boot/Setup.hs
=====================================
@@ -10,6 +10,7 @@ import Distribution.Verbosity
import Distribution.Simple.Program
import Distribution.Simple.Utils
import Distribution.Simple.Setup
+import qualified Distribution.Simple.LocalBuildInfo as LBI
import System.IO
import System.Directory
@@ -32,12 +33,13 @@ main = defaultMainWithHooks ghcHooks
ghcAutogen :: Verbosity -> LocalBuildInfo -> IO ()
ghcAutogen verbosity lbi@LocalBuildInfo{..} = do
-- Get compiler/ root directory from the cabal file
- let Just compilerRoot = takeDirectory <$> pkgDescrFile
+ let Just compilerRoot = takeDirectory . i <$> pkgDescrFile
- let platformHostFile = "GHC/Platform/Host.hs"
- platformHostPath = autogenPackageModulesDir lbi </> platformHostFile
+ i = LBI.interpretSymbolicPathLBI lbi
+ platformHostFile = "GHC/Platform/Host.hs"
+ platformHostPath = i (autogenPackageModulesDir lbi) </> platformHostFile
ghcVersionFile = "GHC/Version.hs"
- ghcVersionPath = autogenPackageModulesDir lbi </> ghcVersionFile
+ ghcVersionPath = i (autogenPackageModulesDir lbi) </> ghcVersionFile
-- Get compiler settings
settings <- lookupEnv "HADRIAN_SETTINGS" >>= \case
=====================================
libraries/ghc-boot/ghc-boot.cabal.in
=====================================
@@ -28,7 +28,7 @@ build-type: Custom
extra-source-files: changelog.md
custom-setup
- setup-depends: base >= 3 && < 5, Cabal >= 1.6 && <3.14, directory, filepath
+ setup-depends: base >= 3 && < 5, Cabal >= 1.6 && <3.16, directory, filepath
source-repository head
type: git
=====================================
rts/ThreadPaused.c
=====================================
@@ -321,11 +321,12 @@ threadPaused(Capability *cap, StgTSO *tso)
if(frame_info == &stg_bh_upd_frame_info) {
// eager black hole: we do nothing
- // it should be a black hole that we own
+ // it should be a black hole (but we may not own it, as another
+ // thread could have raced us to claim it)
ASSERT(bh_info == &stg_BLACKHOLE_info ||
bh_info == &__stg_EAGER_BLACKHOLE_info ||
bh_info == &stg_CAF_BLACKHOLE_info);
- ASSERT(blackHoleOwner(bh) == tso || blackHoleOwner(bh) == NULL);
+
} else {
// lazy black hole
@@ -368,13 +369,15 @@ threadPaused(Capability *cap, StgTSO *tso)
// The payload of the BLACKHOLE points to the TSO
RELEASE_STORE(&((StgInd *)bh)->indirectee, (StgClosure *)tso);
SET_INFO_RELEASE(bh,&stg_BLACKHOLE_info);
+ }
- // .. and we need a write barrier, since we just mutated the closure:
- recordClosureMutated(cap,bh);
+ // We need a write barrier, since the closure was mutated (by
+ // threadPaused for lazy black holes, or the mutator for eager
+ // black holes).
+ recordClosureMutated(cap,bh);
- // We pretend that bh has just been created.
- LDV_RECORD_CREATE(bh);
- }
+ // We pretend that bh has just been created.
+ LDV_RECORD_CREATE(bh);
frame = (StgClosure *) ((StgUpdateFrame *)frame + 1);
if (prev_was_update_frame) {
=====================================
testsuite/tests/simplCore/should_compile/T26349.hs
=====================================
@@ -0,0 +1,11 @@
+{-# LANGUAGE DeepSubsumption, RankNTypes #-}
+module T26349 where
+
+{-# SPECIALIZE INLINE mapTCMT :: (forall b. IO b -> IO b) -> IO a -> IO a #-}
+mapTCMT :: (forall b. m b -> n b) -> m a -> n a
+mapTCMT f m = f m
+
+{-
+ We'll check
+ tcExpr (mapTCMT) (Check ((forall b. IO b -> IO b) -> IO a_sk -> IO a_sk))
+-}
=====================================
testsuite/tests/simplCore/should_compile/T26349.stderr
=====================================
@@ -0,0 +1,3 @@
+==================== Tidy Core rules ====================
+"USPEC mapTCMT @(*) @IO @IO @_"
+ forall (@a). mapTCMT @(*) @IO @IO @a = mapTCMT_$smapTCMT @a
=====================================
testsuite/tests/simplCore/should_compile/all.T
=====================================
@@ -559,3 +559,4 @@ test('T26051', [ grep_errmsg(r'\$wspecMe')
test('T26115', [grep_errmsg(r'DFun')], compile, ['-O -ddump-simpl -dsuppress-uniques'])
test('T26116', normal, compile, ['-O -ddump-rules'])
test('T26117', [grep_errmsg(r'==')], compile, ['-O -ddump-simpl -dsuppress-uniques'])
+test('T26349', normal, compile, ['-O -ddump-rules'])
=====================================
testsuite/tests/simplCore/should_compile/rule2.stderr
=====================================
@@ -10,18 +10,15 @@
==================== Grand total simplifier statistics ====================
-Total ticks: 13
+Total ticks: 11
-2 PreInlineUnconditionally
- 1 ds
- 1 f
+1 PreInlineUnconditionally 1 f
2 UnfoldingDone
1 GHC.Internal.Base.id
1 Roman.bar
1 RuleFired 1 foo/bar
1 LetFloatFromLet 1
-7 BetaReduction
- 1 ds
+6 BetaReduction
1 f
1 a
1 m
=====================================
utils/genprimopcode/genprimopcode.cabal
=====================================
@@ -32,4 +32,4 @@ Executable genprimopcode
Build-Depends: base >= 4 && < 5,
array
if flag(build-tool-depends)
- build-tool-depends: alex:alex >= 3.2.6, happy:happy >= 1.20.0
+ build-tool-depends: alex:alex >= 3.2.6, happy:happy >= 2.1.5 || == 1.20.0 || == 1.20.1.1
View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/dc7cff8bff7d777f53eb8da22bc0f7…
--
View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/dc7cff8bff7d777f53eb8da22bc0f7…
You're receiving this email because of your account on gitlab.haskell.org.
1
0
[Git][ghc/ghc][wip/T26425] Be a little less eager to inline
by Simon Peyton Jones (＠simonpj) 04 Nov '25
by Simon Peyton Jones (＠simonpj) 04 Nov '25
04 Nov '25
Simon Peyton Jones pushed to branch wip/T26425 at Glasgow Haskell Compiler / GHC
Commits:
ec080d05 by Simon Peyton Jones at 2025-11-04T23:45:25+00:00
Be a little less eager to inline
- - - - -
1 changed file:
- compiler/GHC/Core/Opt/Simplify/Utils.hs
Changes:
=====================================
compiler/GHC/Core/Opt/Simplify/Utils.hs
=====================================
@@ -1032,8 +1032,8 @@ interestingArg env e = go env 0 e
| n > 0 = NonTrivArg -- Saturated or unknown call
| otherwise -- n==0, no value arguments; look for an interesting unfolding
= case idUnfolding v of
- OtherCon [] -> NonTrivArg -- It's evaluated, but that's all we know
- OtherCon _ -> ValueArg -- Evaluated and we know it isn't these constructors
+ OtherCon [] -> TrivArg -- It's evaluated, but that's all we know
+ OtherCon _ -> NonTrivArg -- Evaluated and we know it isn't these constructors
-- See Note [OtherCon and interestingArg]
DFunUnfolding {} -> ValueArg -- We konw that idArity=0
CoreUnfolding{ uf_cache = cache }
View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/ec080d05d53af029018ee11db9aeab1…
--
View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/ec080d05d53af029018ee11db9aeab1…
You're receiving this email because of your account on gitlab.haskell.org.
1
0
[Git][ghc/ghc][wip/T26349] 2 commits: Add a HsWrapper optimiser
by Simon Peyton Jones (＠simonpj) 04 Nov '25
by Simon Peyton Jones (＠simonpj) 04 Nov '25
04 Nov '25
Simon Peyton Jones pushed to branch wip/T26349 at Glasgow Haskell Compiler / GHC
Commits:
ea55d5c5 by Simon Peyton Jones at 2025-11-04T22:56:38+00:00
Add a HsWrapper optimiser
This MR addresses #26349, by introduceing optSubTypeHsWrapper.
There is a long
Note [Deep subsumption and WpSubType]
in GHC.Tc.Types.Evidence that explains what is going on.
- - - - -
cf191e55 by Simon Peyton Jones at 2025-11-04T22:59:33+00:00
Improve mkWpFun_FRR
This commit ensures that `mkWpFun_FRR` directly produces a `FunCo` in
the cases where it can.
(Previously called `mkWpFun` which in turn optimised to a `FunCo`, but
that made the smarts in `mkWpFun` /essential/ rather than (as they
should be) optional.
- - - - -
17 changed files:
- compiler/GHC/Core/Coercion.hs
- compiler/GHC/Hs/Syn/Type.hs
- compiler/GHC/HsToCore/Binds.hs
- compiler/GHC/HsToCore/Match.hs
- compiler/GHC/Iface/Ext/Ast.hs
- compiler/GHC/Tc/Errors/Hole.hs
- compiler/GHC/Tc/Gen/App.hs
- compiler/GHC/Tc/Gen/Head.hs
- compiler/GHC/Tc/Gen/Pat.hs
- compiler/GHC/Tc/Types/Evidence.hs
- compiler/GHC/Tc/Utils/Concrete.hs
- compiler/GHC/Tc/Utils/Unify.hs
- compiler/GHC/Tc/Zonk/Type.hs
- + testsuite/tests/simplCore/should_compile/T26349.hs
- + testsuite/tests/simplCore/should_compile/T26349.stderr
- testsuite/tests/simplCore/should_compile/all.T
- testsuite/tests/simplCore/should_compile/rule2.stderr
Changes:
=====================================
compiler/GHC/Core/Coercion.hs
=====================================
@@ -41,7 +41,8 @@ module GHC.Core.Coercion (
mkInstCo, mkAppCo, mkAppCos, mkTyConAppCo,
mkFunCo, mkFunCo2, mkFunCoNoFTF, mkFunResCo,
mkNakedFunCo,
- mkNakedForAllCo, mkForAllCo, mkForAllVisCos, mkHomoForAllCos,
+ mkNakedForAllCo, mkForAllCo, mkForAllVisCos,
+ mkHomoForAllCo, mkHomoForAllCos,
mkPhantomCo, mkAxiomCo,
mkHoleCo, mkUnivCo, mkSubCo,
mkProofIrrelCo,
@@ -980,7 +981,7 @@ mkForAllCo v visL visR kind_co co
= mkReflCo r (mkTyCoForAllTy v visL ty)
| otherwise
- = mkForAllCo_NoRefl v visL visR kind_co co
+ = mk_forall_co v visL visR kind_co co
-- mkForAllVisCos [tv{vis}] constructs a cast
-- forall tv. res ~R# forall tv{vis} res`.
@@ -1000,14 +1001,26 @@ mkHomoForAllCos vs orig_co
= foldr go orig_co vs
where
go :: ForAllTyBinder -> Coercion -> Coercion
- go (Bndr var vis) = mkForAllCo_NoRefl var vis vis MRefl
-
--- | Like 'mkForAllCo', but there is no need to check that the inner coercion isn't Refl;
--- the caller has done that. (For example, it is guaranteed in 'mkHomoForAllCos'.)
--- The kind of the tycovar should be the left-hand kind of the kind coercion.
-mkForAllCo_NoRefl :: TyCoVar -> ForAllTyFlag -> ForAllTyFlag
- -> KindMCoercion -> Coercion -> Coercion
-mkForAllCo_NoRefl tcv visL visR kind_co co
+ go (Bndr var vis) co = mk_forall_co var vis vis MRefl co
+
+mkHomoForAllCo :: TyVar -> Coercion -> Coercion
+-- Specialised for a single TyVar,
+-- and visibility of coreTyLamForAllTyFlag
+mkHomoForAllCo tv orig_co
+ | Just (ty, r) <- isReflCo_maybe orig_co
+ = mkReflCo r (mkForAllTy (Bndr tv vis) ty)
+ | otherwise
+ = mk_forall_co tv vis vis MRefl orig_co
+ where
+ vis = coreTyLamForAllTyFlag
+
+-- | `mk_forall_co` just builds a ForAllCo.
+-- With debug on, it checks invariants (e.g. he kind of the tycovar should
+-- be the left-hand kind of the kind coercion).
+-- Callers should have done any isReflCo short-cutting.
+mk_forall_co :: TyCoVar -> ForAllTyFlag -> ForAllTyFlag
+ -> KindMCoercion -> Coercion -> Coercion
+mk_forall_co tcv visL visR kind_co co
= assertGoodForAllCo tcv visL visR kind_co co $
assertPpr (not (isReflCo co && isReflMCo kind_co && visL == visR)) (ppr co) $
ForAllCo { fco_tcv = tcv, fco_visL = visL, fco_visR = visR
@@ -1769,7 +1782,7 @@ mkPiCos r vs co = foldr (mkPiCo r) co vs
-- | Make a forall 'Coercion', where both types related by the coercion
-- are quantified over the same variable.
mkPiCo :: Role -> Var -> Coercion -> Coercion
-mkPiCo r v co | isTyVar v = mkHomoForAllCos [Bndr v coreTyLamForAllTyFlag] co
+mkPiCo r v co | isTyVar v = mkHomoForAllCo v co
| isCoVar v = assert (not (v `elemVarSet` tyCoVarsOfCo co)) $
-- We didn't call mkForAllCo here because if v does not appear
-- in co, the argument coercion will be nominal. But here we
=====================================
compiler/GHC/Hs/Syn/Type.hs
=====================================
@@ -187,11 +187,13 @@ liftPRType :: (Type -> Type) -> PRType -> PRType
liftPRType f pty = (f (prTypeType pty), [])
hsWrapperType :: HsWrapper -> Type -> Type
+-- Return the type of (WrapExpr wrap e), given that e :: ty
hsWrapperType wrap ty = prTypeType $ go wrap (ty,[])
where
go WpHole = id
+ go (WpSubType w) = go w
go (w1 `WpCompose` w2) = go w1 . go w2
- go (WpFun _ w2 (Scaled m exp_arg)) = liftPRType $ \t ->
+ go (WpFun _ w2 (Scaled m exp_arg) _) = liftPRType $ \t ->
let act_res = funResultTy t
exp_res = hsWrapperType w2 act_res
in mkFunctionType m exp_arg exp_res
=====================================
compiler/GHC/HsToCore/Binds.hs
=====================================
@@ -1597,9 +1597,13 @@ dsHsWrapper hs_wrap thing_inside
ds_hs_wrapper :: HsWrapper
-> ((CoreExpr -> CoreExpr) -> DsM a)
-> DsM a
-ds_hs_wrapper wrap = go wrap
+ds_hs_wrapper hs_wrap
+ = go hs_wrap
where
go WpHole k = k $ \e -> e
+ go (WpSubType w) k = go (optSubTypeHsWrapper w) k
+ -- See (DSST3) in Note [Deep subsumption and WpSubType]
+ -- in GHC.Tc.Types.Evidence
go (WpTyApp ty) k = k $ \e -> App e (Type ty)
go (WpEvLam ev) k = k $ Lam ev
go (WpTyLam tv) k = k $ Lam tv
@@ -1612,13 +1616,13 @@ ds_hs_wrapper wrap = go wrap
go (WpCompose c1 c2) k = go c1 $ \w1 ->
go c2 $ \w2 ->
k (w1 . w2)
- go (WpFun c1 c2 st) k = -- See Note [Desugaring WpFun]
- do { x <- newSysLocalDs st
- ; go c1 $ \w1 ->
- go c2 $ \w2 ->
- let app f a = mkCoreApp (text "dsHsWrapper") f a
- arg = w1 (Var x)
- in k (\e -> (Lam x (w2 (app e arg)))) }
+ go (WpFun c1 c2 st _) k = -- See Note [Desugaring WpFun]
+ do { x <- newSysLocalDs st
+ ; go c1 $ \w1 ->
+ go c2 $ \w2 ->
+ let app f a = mkCoreApp (text "dsHsWrapper") f a
+ arg = w1 (Var x)
+ in k (\e -> (Lam x (w2 (app e arg)))) }
--------------------------------------
dsTcEvBinds_s :: [TcEvBinds] -> ([CoreBind] -> DsM a) -> DsM a
=====================================
compiler/GHC/HsToCore/Match.hs
=====================================
@@ -1240,7 +1240,7 @@ viewLExprEq (e1,_) (e2,_) = lexp e1 e2
-- equating different ways of writing a coercion)
wrap WpHole WpHole = True
wrap (WpCompose w1 w2) (WpCompose w1' w2') = wrap w1 w1' && wrap w2 w2'
- wrap (WpFun w1 w2 _) (WpFun w1' w2' _) = wrap w1 w1' && wrap w2 w2'
+ wrap (WpFun w1 w2 _ _) (WpFun w1' w2' _ _) = wrap w1 w1' && wrap w2 w2'
wrap (WpCast co) (WpCast co') = co `eqCoercion` co'
wrap (WpEvApp et1) (WpEvApp et2) = et1 `ev_term` et2
wrap (WpTyApp t) (WpTyApp t') = eqType t t'
=====================================
compiler/GHC/Iface/Ext/Ast.hs
=====================================
@@ -696,7 +696,7 @@ instance ToHie (LocatedA HsWrapper) where
(WpLet bs) -> toHie $ EvBindContext (mkScope osp) (getRealSpanA osp) (L osp bs)
(WpCompose a b) -> concatM $
[toHie (L osp a), toHie (L osp b)]
- (WpFun a b _) -> concatM $
+ (WpFun a b _ _) -> concatM $
[toHie (L osp a), toHie (L osp b)]
(WpEvLam a) ->
toHie $ C (EvidenceVarBind EvWrapperBind (mkScope osp) (getRealSpanA osp))
=====================================
compiler/GHC/Tc/Errors/Hole.hs
=====================================
@@ -823,9 +823,11 @@ tcFilterHoleFits limit typed_hole ht@(hole_ty, _) candidates =
unfoldWrapper :: HsWrapper -> [Type]
unfoldWrapper = reverse . unfWrp'
- where unfWrp' (WpTyApp ty) = [ty]
- unfWrp' (WpCompose w1 w2) = unfWrp' w1 ++ unfWrp' w2
- unfWrp' _ = []
+ where
+ unfWrp' (WpTyApp ty) = [ty]
+ unfWrp' (WpSubType w) = unfWrp' w
+ unfWrp' (WpCompose w1 w2) = unfWrp' w1 ++ unfWrp' w2
+ unfWrp' _ = []
-- The real work happens here, where we invoke the type checker using
=====================================
compiler/GHC/Tc/Gen/App.hs
=====================================
@@ -794,7 +794,7 @@ tcInstFun do_ql inst_final (tc_fun, fun_ctxt) fun_sigma rn_args
= do { let herald = case fun_ctxt of
VAExpansion (OrigStmt{}) _ _ -> ExpectedFunTySyntaxOp DoOrigin tc_fun
_ -> ExpectedFunTyArg (HsExprTcThing tc_fun) (unLoc arg)
- ; (wrap, arg_ty, res_ty) <-
+ ; (fun_co, arg_ty, res_ty) <-
-- NB: matchActualFunTy does the rep-poly check.
-- For example, suppose we have f :: forall r (a::TYPE r). a -> Int
-- In an application (f x), we need 'x' to have a fixed runtime
@@ -805,7 +805,7 @@ tcInstFun do_ql inst_final (tc_fun, fun_ctxt) fun_sigma rn_args
(n_val_args, fun_sigma) fun_ty
; arg' <- quickLookArg do_ql ctxt arg arg_ty
- ; let acc' = arg' : addArgWrap wrap acc
+ ; let acc' = arg' : addArgWrap (mkWpCastN fun_co) acc
; go (pos+1) acc' res_ty rest_args }
new_arg_ty :: LHsExpr GhcRn -> Int -> TcM (Scaled TcType)
=====================================
compiler/GHC/Tc/Gen/Head.hs
=====================================
@@ -765,13 +765,13 @@ tcInferOverLit lit@(OverLit { ol_val = val
thing = NameThing from_name
mb_thing = Just thing
herald = ExpectedFunTyArg thing (HsLit noExtField hs_lit)
- ; (wrap2, sarg_ty, res_ty) <- matchActualFunTy herald mb_thing (1, from_ty) from_ty
+ ; (co2, sarg_ty, res_ty) <- matchActualFunTy herald mb_thing (1, from_ty) from_ty
; co <- unifyType mb_thing (hsLitType hs_lit) (scaledThing sarg_ty)
-- See Note [Source locations for implicit function calls] in GHC.Iface.Ext.Ast
; let lit_expr = L (l2l loc) $ mkHsWrapCo co $
HsLit noExtField hs_lit
- from_expr = mkHsWrap (wrap2 <.> wrap1) $
+ from_expr = mkHsWrap (mkWpCastN co2 <.> wrap1) $
mkHsVar (L loc from_id)
witness = HsApp noExtField (L (l2l loc) from_expr) lit_expr
lit' = OverLit { ol_val = val
=====================================
compiler/GHC/Tc/Gen/Pat.hs
=====================================
@@ -699,7 +699,7 @@ tc_pat pat_ty penv ps_pat thing_inside = case ps_pat of
-- Expression must be a function
; let herald = ExpectedFunTyViewPat $ unLoc expr
- ; (expr_wrap1, Scaled _mult inf_arg_ty, inf_res_sigma)
+ ; (expr_co1, Scaled _mult inf_arg_ty, inf_res_sigma)
<- matchActualFunTy herald (Just . HsExprRnThing $ unLoc expr) (1,expr_rho) expr_rho
-- See Note [View patterns and polymorphism]
-- expr_wrap1 :: expr_rho "->" (inf_arg_ty -> inf_res_sigma)
@@ -720,7 +720,7 @@ tc_pat pat_ty penv ps_pat thing_inside = case ps_pat of
-- NB: pat_ty comes from matchActualFunTy, so it has a
-- fixed RuntimeRep, as needed to call mkWpFun.
- expr_wrap = expr_wrap2' <.> expr_wrap1
+ expr_wrap = expr_wrap2' <.> mkWpCastN expr_co1
; return $ (ViewPat pat_ty (mkLHsWrap expr_wrap expr') pat', res) }
=====================================
compiler/GHC/Tc/Types/Evidence.hs
=====================================
@@ -8,10 +8,11 @@ module GHC.Tc.Types.Evidence (
-- * HsWrapper
HsWrapper(..),
(<.>), mkWpTyApps, mkWpEvApps, mkWpEvVarApps, mkWpTyLams, mkWpForAllCast,
- mkWpEvLams, mkWpLet, mkWpFun, mkWpCastN, mkWpCastR, mkWpEta,
+ mkWpEvLams, mkWpLet, mkWpFun, mkWpCastN, mkWpCastR, mkWpEta, mkWpSubType,
collectHsWrapBinders,
idHsWrapper, isIdHsWrapper,
pprHsWrapper, hsWrapDictBinders,
+ optSubTypeHsWrapper,
-- * Evidence bindings
TcEvBinds(..), EvBindsVar(..),
@@ -73,7 +74,7 @@ import GHC.Types.Unique.DFM
import GHC.Types.Unique.FM
import GHC.Types.Name( isInternalName )
import GHC.Types.Var
-import GHC.Types.Id( idScaledType )
+import GHC.Types.Id( idScaledType, idType )
import GHC.Types.Var.Env
import GHC.Types.Var.Set
import GHC.Types.Basic
@@ -134,35 +135,128 @@ maybeSymCo NotSwapped co = co
************************************************************************
-}
--- We write wrap :: t1 ~> t2
--- if wrap[ e::t1 ] :: t2
+{- Note [Deep subsumption and WpSubType]
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+When making DeepSubsumption checks, we may end up with hard-to-spot identity wrappers.
+For example (#26349) suppose we have
+ (forall a. Eq a => a->a) -> Int <= (forall a. Eq a => a->a) -> Int
+The two types are equal so we should certainly get an identity wrapper. But we'll get
+tihs wrapper from `tcSubType`:
+ WpFun (WpTyLam a <.> WpEvLam dg <.> WpLet (dw=dg) <.> WpEvApp dw <.> WpTyApp a)
+ WpHole
+That elaborate wrapper is really just a no-op, but it's far from obvious. If we just
+desugar (HsWrap f wp) straightforwardly we'll get
+ \(g:forall a. Eq a => a -> a).
+ f (/\a. \(dg:Eq a). let dw=dg in g a dw)
+
+To recognise that as just `f`, we'd have to eta-reduce twice. But eta-reduction
+is not sound in general, so we'll end up retaining the lambdas. Two bad results:
+
+* Adding DeepSubsumption gratuitiously makes programs less efficient.
+
+* When the subsumption is on the LHS of a rule, or in a SPECIALISE pragma, we
+ may not be able to make a decent RULE at all, and will fail with "LHS of rule
+ is too complicated to desugar" (#26255)
+
+It'd be ideal to solve the problem at the source, by never generating those
+gruesome wrappers in the first place, but we can't do that because:
+
+* The WpTyLam and WpTyApp are introduced independently, not together, in `tcSubType`,
+ so we can't easily cancel them out. For example, even if we have
+ forall a. t1 <= forall a. t2
+ there is no guarantee that these are the "same" a. E.g.
+ forall a b. a -> b -> b <= forall x y. y -> x -> x
+ Similarly WpEvLam and WpEvApp
+
+* We have not yet done constraint solving so we don't know what evidence will
+ end up in those WpLet bindings.
+
+TL;DR we must generate the wrapper and then optimise it way if it turns out
+that it is a no-op. Here's our solution:
+
+(DSST1) Tag the wrappers generated from a subtype check with WpSubType. In normal
+ wrappers the binders of a WpTyLam or WpEvLam can scope over the "hole" of the
+ wrapper -- that is how we introduce type-lambdas and dictionary-lambda into the
+ terms! But in /subtype/ wrappers, these type/dictionary lambdas only scope over
+ the WpTyApp and WpEvApp nodes in the /same/ wrapper. That is what justifies us
+ eta-reducing the type/dictionary lambdas.
+
+ In short, (WpSubType wp) means the same as `wp`, but with the added promise that
+ the binders in `wp` do not scope over the hole.
+
+(DSST2) Avoid creating a WpSubType in the common WpHole case, using `mkWpSubType`.
+
+(DSST3) When desugaring, try eta-reduction on the payload of a WpSubType.
+ This is done in `GHC.HsToCore.Binds.dsHsWrapper` by the call to `optSubTypeHsWrapper`.
+
+ We don't attempt to optimise HsWrappers /other than/ subtype wrappers. Why not?
+ Because there aren't any useful optimsations we can do. (We could collapse
+ adjacent `WpCast`s perhaps, but that'll happen later automatically via `mkCast`.)
+
+ TL;DR:
+ * we /must/ optimise subtype-HsWrappers (that's the point of this Note!)
+ * there is little point in attempting to optimise any other HsWrappers
+
+Note [WpFun-RR-INVARIANT]
+~~~~~~~~~~~~~~~~~~~~~~~~~
+Given
+ wrap = WpFun wrap1 wrap2 sty1 ty2
+ where: wrap1 :: exp_arg ~~> act_arg
+ wrap2 :: act_res ~~> exp_res
+ wrap :: (act_arg -> act_res) ~~> (exp_arg -> exp_res)
+we have
+ WpFun-RR-INVARIANT:
+ the input (exp_arg) and output (act_arg) types of `wrap1`
+ both have a fixed runtime-rep
+
+Reason: We desugar wrap[e] into
+ \(x:exp_arg). wrap2[ e wrap1[x] ]
+And then, because of Note [Representation polymorphism invariants], we need:
+
+ * `exp_arg` must have a fixed runtime rep,
+ so that lambda obeys the the FRR rules
+
+ * `act_arg` must have a fixed runtime rep,
+ so the that application (e wrap1[x]) obeys the FRR tules
+
+Hence WpFun-INVARIANT.
+-}
+
data HsWrapper
+ -- NOTATION (~~>):
+ -- We write wrap :: t1 ~~> t2
+ -- if wrap[ e::t1 ] :: t2
= WpHole -- The identity coercion
+ | WpSubType HsWrapper
+ -- (WpSubType wp) is the same as `wp`, but with extra invariants
+ -- See Note [Deep subsumption and WpSubType] (DSST1)
+
| WpCompose HsWrapper HsWrapper
-- (wrap1 `WpCompose` wrap2)[e] = wrap1[ wrap2[ e ]]
--
-- Hence (\a. []) `WpCompose` (\b. []) = (\a b. [])
-- But ([] a) `WpCompose` ([] b) = ([] b a)
--
- -- If wrap1 :: t2 ~> t3
- -- wrap2 :: t1 ~> t2
- --- Then (wrap1 `WpCompose` wrap2) :: t1 ~> t3
-
- | WpFun HsWrapper HsWrapper (Scaled TcTypeFRR)
- -- (WpFun wrap1 wrap2 (w, t1))[e] = \(x:_w exp_arg). wrap2[ e wrap1[x] ]
- -- So note that if e :: act_arg -> act_res
- -- wrap1 :: exp_arg ~> act_arg
- -- wrap2 :: act_res ~> exp_res
- -- then WpFun wrap1 wrap2 : (act_arg -> arg_res) ~> (exp_arg -> exp_res)
+ -- If wrap1 :: t2 ~~> t3
+ -- wrap2 :: t1 ~~> t2
+ --- Then (wrap1 `WpCompose` wrap2) :: t1 ~~> t3
+
+ | WpFun HsWrapper HsWrapper (Scaled TcTypeFRR) TcType
+ -- (WpFun wrap1 wrap2 (w, t1) t2)[e] = \(x:_w exp_arg). wrap2[ e wrap1[x] ]
+ --
+ -- INVARIANT: both input and output types of `wrap1` have a fixed runtime-rep
+ -- See Note [WpFun-RR-INVARIANT]
+ --
+ -- Typing rules:
+ -- If e :: act_arg -> act_res
+ -- wrap1 :: exp_arg ~~> act_arg
+ -- wrap2 :: act_res ~~> exp_res
+ -- then WpFun wrap1 wrap2 :: (act_arg -> act_res) ~~> (exp_arg -> exp_res)
-- This isn't the same as for mkFunCo, but it has to be this way
-- because we can't use 'sym' to flip around these HsWrappers
- -- The TcType is the "from" type of the first wrapper;
- -- it always a Type, not a Constraint
--
- -- NB: a WpFun is always for a (->) function arrow
- --
- -- Use 'mkWpFun' to construct such a wrapper.
+ -- NB: a WpFun is always for a (->) function arrow, never (=>)
| WpCast TcCoercionR -- A cast: [] `cast` co
-- Guaranteed not the identity coercion
@@ -212,50 +306,48 @@ WpCast c1 <.> WpCast c2 = WpCast (c2 `mkTransCo` c1)
--
-- NB: <.> behaves like function composition:
--
- -- WpCast c1 <.> WpCast c2 :: coercionLKind c2 ~> coercionRKind c1
+ -- WpCast c1 <.> WpCast c2 :: coercionLKind c2 ~~> coercionRKind c1
--
-- This is thus the same as WpCast (c2 ; c1) and not WpCast (c1 ; c2).
c1 <.> c2 = c1 `WpCompose` c2
--- | Smart constructor to create a 'WpFun' 'HsWrapper', which avoids introducing
--- a lambda abstraction if the two supplied wrappers are either identities or
--- casts.
---
--- PRECONDITION: either:
---
--- 1. both of the 'HsWrapper's are identities or casts, or
--- 2. both the "from" and "to" types of the first wrapper have a syntactically
--- fixed RuntimeRep (see Note [Fixed RuntimeRep] in GHC.Tc.Utils.Concrete).
mkWpFun :: HsWrapper -> HsWrapper
-> Scaled TcTypeFRR -- ^ the "from" type of the first wrapper
-> TcType -- ^ Either "from" type or "to" type of the second wrapper
-- (used only when the second wrapper is the identity)
-> HsWrapper
-mkWpFun WpHole WpHole _ _ = WpHole
-mkWpFun WpHole (WpCast co2) (Scaled w t1) _ = WpCast (mk_wp_fun_co w (mkRepReflCo t1) co2)
-mkWpFun (WpCast co1) WpHole (Scaled w _) t2 = WpCast (mk_wp_fun_co w (mkSymCo co1) (mkRepReflCo t2))
-mkWpFun (WpCast co1) (WpCast co2) (Scaled w _) _ = WpCast (mk_wp_fun_co w (mkSymCo co1) co2)
-mkWpFun w_arg w_res t1 _ =
- -- In this case, we will desugar to a lambda
- --
- -- \x. w_res[ e w_arg[x] ]
- --
- -- To satisfy Note [Representation polymorphism invariants] in GHC.Core,
- -- it must be the case that both the lambda bound variable x and the function
- -- argument w_arg[x] have a fixed runtime representation, i.e. that both the
- -- "from" and "to" types of the first wrapper "w_arg" have a fixed runtime representation.
- --
- -- Unfortunately, we can't check this with an assertion here, because of
- -- [Wrinkle: Typed Template Haskell] in Note [hasFixedRuntimeRep] in GHC.Tc.Utils.Concrete.
- WpFun w_arg w_res t1
-
-mkWpEta :: [Id] -> HsWrapper -> HsWrapper
+-- ^ Smart constructor for `WpFun`
+-- Just removes clutter and optimises some common cases.
+--
+-- PRECONDITION: same as Note [WpFun-RR-INVARIANT]
+--
+-- Unfortunately, we can't check PRECONDITION with an assertion here, because of
+-- [Wrinkle: Typed Template Haskell] in Note [hasFixedRuntimeRep] in GHC.Tc.Utils.Concrete.
+mkWpFun w1 w2 st1@(Scaled m1 t1) t2
+ = case (w1,w2) of
+ (WpHole, WpHole) -> WpHole
+ (WpHole, WpCast co2) -> WpCast (mk_wp_fun_co m1 (mkRepReflCo t1) co2)
+ (WpCast co1, WpHole) -> WpCast (mk_wp_fun_co m1 (mkSymCo co1) (mkRepReflCo t2))
+ (WpCast co1, WpCast co2) -> WpCast (mk_wp_fun_co m1 (mkSymCo co1) co2)
+ (_, _) -> WpFun w1 w2 st1 t2
+
+mkWpSubType :: HsWrapper -> HsWrapper
+-- See (DSST2) in Note [Deep subsumption and WpSubType]
+mkWpSubType WpHole = WpHole
+mkWpSubType (WpCast co) = WpCast co
+mkWpSubType w = WpSubType w
+
+mkWpEta :: Type -> [Id] -> HsWrapper -> HsWrapper
-- (mkWpEta [x1, x2] wrap) [e]
-- = \x1. \x2. wrap[e x1 x2]
-- Just generates a bunch of WpFuns
-mkWpEta xs wrap = foldr eta_one wrap xs
+-- The incoming type is the type of the entire expression
+mkWpEta orig_fun_ty xs wrap = go orig_fun_ty xs
where
- eta_one x wrap = WpFun idHsWrapper wrap (idScaledType x)
+ go _ [] = wrap
+ go fun_ty (id:ids) = WpFun idHsWrapper (go res_ty ids) (idScaledType id) res_ty
+ where
+ res_ty = funResultTy fun_ty
mk_wp_fun_co :: Mult -> TcCoercionR -> TcCoercionR -> TcCoercionR
mk_wp_fun_co mult arg_co res_co
@@ -333,8 +425,9 @@ hsWrapDictBinders wrap = go wrap
where
go (WpEvLam dict_id) = unitBag dict_id
go (w1 `WpCompose` w2) = go w1 `unionBags` go w2
- go (WpFun _ w _) = go w
+ go (WpFun _ w _ _) = go w
go WpHole = emptyBag
+ go (WpSubType {}) = emptyBag -- See Note [Deep subsumption and WpSubType]
go (WpCast {}) = emptyBag
go (WpEvApp {}) = emptyBag
go (WpTyLam {}) = emptyBag
@@ -350,6 +443,7 @@ collectHsWrapBinders wrap = go wrap []
go :: HsWrapper -> [HsWrapper] -> ([Var], HsWrapper)
go (WpEvLam v) wraps = add_lam v (gos wraps)
go (WpTyLam v) wraps = add_lam v (gos wraps)
+ go (WpSubType w) wraps = go w wraps
go (WpCompose w1 w2) wraps = go w1 (w2:wraps)
go wrap wraps = ([], foldl' (<.>) wrap wraps)
@@ -358,6 +452,162 @@ collectHsWrapBinders wrap = go wrap []
add_lam v (vs,w) = (v:vs, w)
+
+optSubTypeHsWrapper :: HsWrapper -> HsWrapper
+-- This optimiser is used only on the payload of WpSubType
+-- It finds cases where the entire wrapper is a no-op
+-- See (DSST3) in Note [Deep subsumption and WpSubType]
+optSubTypeHsWrapper wrap
+ = opt wrap
+ where
+ opt :: HsWrapper -> HsWrapper
+ opt w = foldr (<.>) WpHole (opt1 w [])
+
+ opt1 :: HsWrapper -> [HsWrapper] -> [HsWrapper]
+ -- opt1 w ws = w <.> (foldr <.> WpHole ws)
+ -- INVARIANT: ws::[HsWrapper] is optimised
+ opt1 WpHole ws = ws
+ opt1 (WpSubType w) ws = opt1 w ws
+ opt1 (w1 `WpCompose` w2) ws = opt1 w1 (opt1 w2 ws)
+ opt1 (WpCast co) ws = opt_co co ws
+ opt1 (WpEvLam ev) ws = opt_ev_lam ev ws
+ opt1 (WpTyLam tv) ws = opt_ty_lam tv ws
+ opt1 (WpLet binds) ws = pushWpLet binds ws
+ opt1 (WpFun w1 w2 sty1 ty2) ws = opt_fun w1 w2 sty1 ty2 ws
+ opt1 w@(WpTyApp {}) ws = w : ws
+ opt1 w@(WpEvApp {}) ws = w : ws
+
+ -----------------
+ -- (WpTyLam a <.> WpTyApp a <.> w) = w
+ -- i.e. /\a. <hole> a --> <hole>
+ -- This is only valid if whatever fills the hole does not mention 'a'
+ -- But that's guaranteed in subtype-wrappers;
+ -- see (DSST1) in Note [Deep subsumption and WpSubType]
+ opt_ty_lam tv (WpTyApp ty : ws)
+ | Just tv' <- getTyVar_maybe ty
+ , tv==tv'
+ , all (tv `not_in`) ws
+ = ws
+
+ -- (WpTyLam a <.> WpCastCo co <.> w)
+ -- = WpCast (ForAllCo a co) (WpTyLam <.> w)
+ opt_ty_lam tv (WpCast co : ws)
+ = opt_co (mkHomoForAllCo tv co) (opt_ty_lam tv ws)
+
+ opt_ty_lam tv ws
+ = WpTyLam tv : ws
+
+ -----------------
+ -- (WpEvLam ev <.> WpEvAp ev <.> w) = w
+ -- Similar notes to WpTyLam
+ opt_ev_lam ev (WpEvApp ev_tm : ws)
+ | EvExpr (Var ev') <- ev_tm
+ , ev == ev'
+ , all (ev `not_in`) ws
+ = ws
+
+ -- (WpEvLam ev <.> WpCast co <.> w)
+ -- = WpCast (FunCo ev co) (WpEvLam <.> w)
+ opt_ev_lam ev (WpCast co : ws)
+ = opt_co fun_co (opt_ev_lam ev ws)
+ where
+ fun_co = mkFunCo Representational FTF_C_T
+ (mkNomReflCo ManyTy)
+ (mkRepReflCo (idType ev))
+ co
+
+ opt_ev_lam ev ws
+ = WpEvLam ev : ws
+
+ -----------------
+ -- WpCast co <.> WpCast co' <.> ws = WpCast (co;co') ws
+ opt_co co (WpCast co' : ws) = opt_co (co `mkTransCo` co') ws
+ opt_co co ws | isReflexiveCo co = ws
+ | otherwise = WpCast co : ws
+
+ ------------------
+ opt_fun w1 w2 sty1 ty2 ws
+ = case mkWpFun (opt w1) (opt w2) sty1 ty2 of
+ WpHole -> ws
+ WpCast co -> opt_co co ws
+ w -> w : ws
+
+ ------------------
+ -- Tiresome check that the lambda-bound type/evidence variable that we
+ -- want to eta-reduce isn't free in the rest of the wrapper
+ not_in :: TyVar -> HsWrapper -> Bool
+ not_in _ WpHole = True
+ not_in v (WpCast co) = not (anyFreeVarsOfCo (== v) co)
+ not_in v (WpTyApp ty) = not (anyFreeVarsOfType (== v) ty)
+ not_in v (WpFun w1 w2 _ _) = not_in v w1 && not_in v w2
+ not_in v (WpSubType w) = not_in v w
+ not_in v (WpCompose w1 w2) = not_in v w1 && not_in v w2
+ not_in v (WpEvApp (EvExpr e)) = not (v `elemVarSet` exprFreeVars e)
+ not_in _ (WpEvApp (EvTypeable {})) = False -- Giving up; conservative
+ not_in _ (WpEvApp (EvFun {})) = False -- Giving up; conservative
+ not_in _ (WpTyLam {}) = False -- Give up; conservative
+ not_in _ (WpEvLam {}) = False -- Ditto
+ not_in _ (WpLet {}) = False -- Ditto
+
+pushWpLet :: TcEvBinds -> [HsWrapper] -> [HsWrapper]
+-- See if we can transform
+-- WpLet binds <.> w1 <.> .. <.> wn --> w1' <.> .. <.> wn'
+-- by substitution.
+-- We do this just for the narrow case when
+-- - the `binds` are all just v=w, variables only
+-- - the wi are all WpTyApp, WpEvApp, or WpCast
+-- This is just enough to get us the eta-reductions that we seek
+pushWpLet tc_ev_binds ws
+ = case tc_ev_binds of
+ TcEvBinds {} -> pprPanic "pushWpLet" (ppr tc_ev_binds)
+ EvBinds binds
+ | isEmptyBag binds
+ -> ws
+ | Just env <- ev_bind_swizzle binds
+ -> case go env ws of
+ Just ws' -> ws'
+ Nothing -> bale_out
+ | otherwise
+ -> bale_out
+ where
+ bale_out = WpLet tc_ev_binds : ws
+
+ go :: IdEnv Id -> [HsWrapper] -> Maybe [HsWrapper]
+ go env (WpCast co : ws) = do { ws' <- go env ws
+ ; return (WpCast co : ws') }
+ go env (WpTyApp ty : ws) = do { ws' <- go env ws
+ ; return (WpTyApp ty : ws') }
+ go env (WpEvApp (EvExpr (Var v)) : ws)
+ = do { v' <- swizzle_id env v
+ ; ws' <- go env ws
+ ; return (WpEvApp (EvExpr (Var v')) : ws') }
+
+ go _ ws = case ws of
+ [] -> Just []
+ (_:_) -> Nothing -- Could not fully eliminate the WpLet
+
+ swizzle_id :: IdEnv Id -> Id -> Maybe Id
+ -- Nothing <=> ran out of fuel
+ -- This is just belt and braces; we should never build bottom evidence
+ swizzle_id env v = go 100 v
+ where
+ go :: Int -> EvId -> Maybe EvId
+ go fuel v
+ | fuel == 0 = Nothing
+ | Just v' <- lookupVarEnv env v = go (fuel-1) v'
+ | otherwise = Just v
+
+ ev_bind_swizzle :: Bag EvBind -> Maybe (IdEnv Id)
+ -- Succeeds only if the bindings are all var-to-var bindings
+ ev_bind_swizzle evbs = foldl' do_one (Just emptyVarEnv) evbs
+ where
+ do_one :: Maybe (IdEnv Id) -> EvBind -> Maybe (IdEnv Id)
+ do_one Nothing _ = Nothing
+ do_one (Just swizzle) (EvBind {eb_lhs = bndr, eb_rhs = rhs})
+ = case rhs of
+ EvExpr (Var v) -> Just (extendVarEnv swizzle bndr v)
+ _ -> Nothing
+
{-
************************************************************************
* *
@@ -1018,8 +1268,9 @@ pprHsWrapper wrap pp_thing_inside
-- True <=> appears in function application position
-- False <=> appears as body of let or lambda
help it WpHole = it
- help it (WpCompose f1 f2) = help (help it f2) f1
- help it (WpFun f1 f2 (Scaled w t1)) = add_parens $ text "\\(x" <> dcolon <> brackets (ppr w) <> ppr t1 <> text ")." <+>
+ help it (WpCompose w1 w2) = help (help it w2) w1
+ help it (WpSubType w) = no_parens $ text "subtype" <> braces (help it w False)
+ help it (WpFun f1 f2 (Scaled w t1) _) = add_parens $ text "\\(x" <> dcolon <> brackets (ppr w) <> ppr t1 <> text ")." <+>
help (\_ -> it True <+> help (\_ -> text "x") f1 True) f2 False
help it (WpCast co) = add_parens $ sep [it False, nest 2 (text "|>"
<+> pprParendCo co)]
=====================================
compiler/GHC/Tc/Utils/Concrete.hs
=====================================
@@ -626,8 +626,12 @@ hasFixedRuntimeRep :: HasDebugCallStack
-- @ki@ is concrete, and @co :: ty ~# ty'@.
-- That is, @ty'@ has a syntactically fixed RuntimeRep
-- in the sense of Note [Fixed RuntimeRep].
-hasFixedRuntimeRep frr_ctxt ty =
- checkFRR_with (fmap (fmap coToMCo) . unifyConcrete_kind (fsLit "cx") . ConcreteFRR) frr_ctxt ty
+hasFixedRuntimeRep frr_ctxt ty
+ = checkFRR_with unify_conc frr_ctxt ty
+ where
+ unify_conc frr_orig ki
+ = do { co <- unifyConcrete_kind (fsLit "cx") (ConcreteFRR frr_orig) ki
+ ; return (coToMCo co) }
-- | Like 'hasFixedRuntimeRep', but we perform an eager syntactic check.
--
=====================================
compiler/GHC/Tc/Utils/Unify.hs
=====================================
@@ -148,7 +148,7 @@ matchActualFunTy
-- (Both are used only for error messages)
-> TcRhoType
-- ^ Type to analyse: a TcRhoType
- -> TcM (HsWrapper, Scaled TcSigmaTypeFRR, TcSigmaType)
+ -> TcM (TcCoercion, Scaled TcSigmaTypeFRR, TcSigmaType)
-- This function takes in a type to analyse (a RhoType) and returns
-- an argument type and a result type (splitting apart a function arrow).
-- The returned argument type is a SigmaType with a fixed RuntimeRep;
@@ -157,7 +157,7 @@ matchActualFunTy
-- See Note [matchActualFunTy error handling] for the first three arguments
-- If (wrap, arg_ty, res_ty) = matchActualFunTy ... fun_ty
--- then wrap :: fun_ty ~> (arg_ty -> res_ty)
+-- then wrap :: fun_ty ~~> (arg_ty -> res_ty)
-- and NB: res_ty is an (uninstantiated) SigmaType
matchActualFunTy herald mb_thing err_info fun_ty
@@ -172,13 +172,13 @@ matchActualFunTy herald mb_thing err_info fun_ty
-- hide the forall inside a meta-variable
go :: TcRhoType -- The type we're processing, perhaps after
-- expanding type synonyms
- -> TcM (HsWrapper, Scaled TcSigmaTypeFRR, TcSigmaType)
+ -> TcM (TcCoercion, Scaled TcSigmaTypeFRR, TcSigmaType)
go ty | Just ty' <- coreView ty = go ty'
go (FunTy { ft_af = af, ft_mult = w, ft_arg = arg_ty, ft_res = res_ty })
= assert (isVisibleFunArg af) $
do { hasFixedRuntimeRep_syntactic (FRRExpectedFunTy herald 1) arg_ty
- ; return (idHsWrapper, Scaled w arg_ty, res_ty) }
+ ; return (mkNomReflCo fun_ty, Scaled w arg_ty, res_ty) }
go ty@(TyVarTy tv)
| isMetaTyVar tv
@@ -210,7 +210,7 @@ matchActualFunTy herald mb_thing err_info fun_ty
; res_ty <- newOpenFlexiTyVarTy
; let unif_fun_ty = mkScaledFunTys [arg_ty] res_ty
; co <- unifyType mb_thing fun_ty unif_fun_ty
- ; return (mkWpCastN co, arg_ty, res_ty) }
+ ; return (co, arg_ty, res_ty) }
------------
mk_ctxt :: TcType -> TidyEnv -> ZonkM (TidyEnv, ErrCtxtMsg)
@@ -249,8 +249,10 @@ matchActualFunTys :: ExpectedFunTyOrigin -- ^ See Note [Herald for matchExpected
-> Arity
-> TcSigmaType
-> TcM (HsWrapper, [Scaled TcSigmaTypeFRR], TcRhoType)
--- If matchActualFunTys n ty = (wrap, [t1,..,tn], res_ty)
--- then wrap : ty ~> (t1 -> ... -> tn -> res_ty)
+-- NB: Called only from `tcSynArgA`, and hence scheduled for destruction
+--
+-- If matchActualFunTys n fun_ty = (wrap, [t1,..,tn], res_ty)
+-- then wrap : fun_ty ~~> (t1 -> ... -> tn -> res_ty)
-- and res_ty is a RhoType
-- NB: the returned type is top-instantiated; it's a RhoType
matchActualFunTys herald ct_orig n_val_args_wanted top_ty
@@ -265,15 +267,13 @@ matchActualFunTys herald ct_orig n_val_args_wanted top_ty
go 0 _ fun_ty = return (idHsWrapper, [], fun_ty)
go n so_far fun_ty
- = do { (wrap_fun1, arg_ty1, res_ty1) <- matchActualFunTy
- herald Nothing
- (n_val_args_wanted, top_ty)
- fun_ty
- ; (wrap_res, arg_tys, res_ty) <- go (n-1) (arg_ty1:so_far) res_ty1
+ = do { (co1, arg_ty1, res_ty1) <- matchActualFunTy herald Nothing
+ (n_val_args_wanted, top_ty) fun_ty
+ ; (wrap_res, arg_tys, res_ty) <- go (n-1) (arg_ty1:so_far) res_ty1
; let wrap_fun2 = mkWpFun idHsWrapper wrap_res arg_ty1 res_ty
-- NB: arg_ty1 comes from matchActualFunTy, so it has
- -- a syntactically fixed RuntimeRep as needed to call mkWpFun.
- ; return (wrap_fun2 <.> wrap_fun1, arg_ty1:arg_tys, res_ty) }
+ -- a syntactically fixed RuntimeRep
+ ; return (wrap_fun2 <.> mkWpCastN co1, arg_ty1:arg_tys, res_ty) }
{-
************************************************************************
@@ -459,7 +459,7 @@ tcSkolemiseGeneral ds_flag ctxt top_ty expected_ty thing_inside
tcSkolemiseCompleteSig :: TcCompleteSig
-> ([ExpPatType] -> TcRhoType -> TcM result)
-> TcM (HsWrapper, result)
--- ^ The wrapper has type: spec_ty ~> expected_ty
+-- ^ The wrapper has type: spec_ty ~~> expected_ty
-- See Note [Skolemisation] for the differences between
-- tcSkolemiseCompleteSig and tcTopSkolemise
@@ -790,7 +790,7 @@ matchExpectedFunTys :: forall a.
-> ([ExpPatType] -> ExpRhoType -> TcM a)
-> TcM (HsWrapper, a)
-- If matchExpectedFunTys n ty = (wrap, _)
--- then wrap : (t1 -> ... -> tn -> ty_r) ~> ty,
+-- then wrap : (t1 -> ... -> tn -> ty_r) ~~> ty,
-- where [t1, ..., tn], ty_r are passed to the thing_inside
--
-- Unconditionally concludes by skolemising any trailing invisible
@@ -865,12 +865,13 @@ matchExpectedFunTys herald ctx arity (Check top_ty) thing_inside
, ft_arg = arg_ty, ft_res = res_ty })
= assert (isVisibleFunArg af) $
do { let arg_pos = arity - n_req + 1 -- 1 for the first argument etc
- ; (arg_co, arg_ty) <- hasFixedRuntimeRep (FRRExpectedFunTy herald arg_pos) arg_ty
+ ; (arg_co, arg_ty_frr) <- hasFixedRuntimeRep (FRRExpectedFunTy herald arg_pos) arg_ty
+ ; let arg_sty_frr = Scaled mult arg_ty_frr
; (wrap_res, result) <- check (n_req - 1)
- (mkCheckExpFunPatTy (Scaled mult arg_ty) : rev_pat_tys)
+ (mkCheckExpFunPatTy arg_sty_frr : rev_pat_tys)
res_ty
; let wrap_arg = mkWpCastN arg_co
- fun_wrap = mkWpFun wrap_arg wrap_res (Scaled mult arg_ty) res_ty
+ fun_wrap = mkWpFun wrap_arg wrap_res arg_sty_frr res_ty
; return (fun_wrap, result) }
----------------------------
@@ -1407,7 +1408,7 @@ tcSubTypePat :: CtOrigin -> UserTypeCtxt
-- Used in patterns; polarity is backwards compared
-- to tcSubType
-- If wrap = tc_sub_type_et t1 t2
--- => wrap :: t1 ~> t2
+-- => wrap :: t1 ~~> t2
tcSubTypePat inst_orig ctxt (Check ty_actual) ty_expected
= tc_sub_type unifyTypeET inst_orig ctxt ty_actual ty_expected
@@ -1427,11 +1428,12 @@ tcSubTypeDS :: HsExpr GhcRn
-- DeepSubsumption <=> when checking, this type
-- is deeply skolemised
-> TcM HsWrapper
--- Only one call site, in GHC.Tc.Gen.App.tcApp
+-- Only one call site, in GHC.Tc.Gen.App.checkResultTy
tcSubTypeDS rn_expr act_rho exp_rho
- = tc_sub_type_deep Top (unifyExprType rn_expr) orig GenSigCtxt act_rho exp_rho
- where
- orig = exprCtOrigin rn_expr
+ = do { wrap <- tc_sub_type_deep Top (unifyExprType rn_expr)
+ (exprCtOrigin rn_expr)
+ GenSigCtxt act_rho exp_rho
+ ; return (mkWpSubType wrap) }
---------------
@@ -1456,7 +1458,7 @@ tcSubTypeSigma :: CtOrigin -- where did the actual type arise / why are we
-> TcSigmaType -> TcSigmaType -> TcM HsWrapper
-- External entry point, but no ExpTypes on either side
-- Checks that actual <= expected
--- Returns HsWrapper :: actual ~ expected
+-- Returns HsWrapper :: actual ~~> expected
tcSubTypeSigma orig ctxt ty_actual ty_expected
= tc_sub_type (unifyType Nothing) orig ctxt ty_actual ty_expected
@@ -1495,7 +1497,7 @@ tc_sub_type :: (TcType -> TcType -> TcM TcCoercionN) -- How to unify
-> TcM HsWrapper
-- Checks that actual_ty is more polymorphic than expected_ty
-- If wrap = tc_sub_type t1 t2
--- => wrap :: t1 ~> t2
+-- => wrap :: t1 ~~> t2
--
-- The "how to unify argument" is always a call to `uType TypeLevel orig`,
-- but with different ways of constructing the CtOrigin `orig` from
@@ -1504,7 +1506,8 @@ tc_sub_type :: (TcType -> TcType -> TcM TcCoercionN) -- How to unify
----------------------
tc_sub_type unify inst_orig ctxt ty_actual ty_expected
= do { ds_flag <- getDeepSubsumptionFlag
- ; tc_sub_type_ds Top ds_flag unify inst_orig ctxt ty_actual ty_expected }
+ ; wrap <- tc_sub_type_ds Top ds_flag unify inst_orig ctxt ty_actual ty_expected
+ ; return (mkWpSubType wrap) }
----------------------
tc_sub_type_ds :: Position p -- ^ position in the type (for error messages only)
@@ -1753,59 +1756,59 @@ we deal with function arrows. Suppose we have:
ty_actual = act_arg -> act_res
ty_expected = exp_arg -> exp_res
-To produce fun_wrap :: (act_arg -> act_res) ~> (exp_arg -> exp_res), we use
+To produce fun_wrap :: (act_arg -> act_res) ~~> (exp_arg -> exp_res), we use
the fact that the function arrow is contravariant in its argument type and
covariant in its result type. Thus we recursively perform subtype checks
on the argument types (with actual/expected switched) and the result types,
to get:
- arg_wrap :: exp_arg ~> act_arg -- NB: expected/actual have switched sides
- res_wrap :: act_res ~> exp_res
+ arg_wrap :: exp_arg ~~> act_arg -- NB: expected/actual have switched sides
+ res_wrap :: act_res ~~> exp_res
Then fun_wrap = mkWpFun arg_wrap res_wrap.
-Wrinkle [Representation-polymorphism checking during subtyping]
+Note [Representation-polymorphism checking during subtyping]
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+When doing deep subsumption in `tc_sub_type_deep`, looking under function arrows,
+we would usually build a `WpFun` HsWrapper. When desugared, we get eta-expansion:
- Inserting a WpFun HsWrapper amounts to impedance matching in deep subsumption
- via eta-expansion:
+ f ==> \(x :: exp_arg). res_wrap [ f (arg_wrap [x]) ]
- f ==> \ (x :: exp_arg) -> res_wrap [ f (arg_wrap [x]) ]
+Since we produce a lambda, we must enforce the representation polymorphism
+invariants described in Note [Representation polymorphism invariants] in GHC.Core.
+That is, we must ensure that both
+ - x (the lambda binder), and
+ - (arg_wrap [x]) (the function argument)
+have a fixed runtime representation.
- As we produce a lambda, we must enforce the representation polymorphism
- invariants described in Note [Representation polymorphism invariants] in GHC.Core.
- That is, we must ensure that both x (the lambda binder) and (arg_wrap [x]) (the function argument)
- have a fixed runtime representation.
+But we don't /always/ need to produce a `WpFun`: if both argument and result wrappers
+are merely coercions, we can produce a `WpCast co` instead of a `WpFun`. In that
+case there is no eta-expansion, and hence no need for FRR checks.
- Note however that desugaring mkWpFun does not always introduce a lambda: if
- both the argument and result HsWrappers are casts, then a FunCo cast suffices,
- in which case we should not perform representation-polymorphism checking.
+Here's a contrived example (there are undoubtedly more natural examples)
+(see testsuite/tests/rep-poly/NoEtaRequired):
- This means that, in the FunTy/FunTy case of tc_sub_type_deep, we can skip
- the representation-polymorphism checks if the produced argument and result
- wrappers are identities or casts.
- It is important to do so, otherwise we reject valid programs.
+ type Id :: k -> k
+ type family Id a where
- Here's a contrived example (there are undoubtedly more natural examples)
- (see testsuite/tests/rep-poly/NoEtaRequired):
+ type T :: TYPE r -> TYPE (Id r)
+ type family T a where
- type Id :: k -> k
- type family Id a where
+ test :: forall r (a :: TYPE r). a :~~: T a -> ()
+ test HRefl =
+ let
+ f :: (a -> a) -> ()
+ f _ = ()
+ g :: T a -> T a
+ g = undefined
+ in f g
- type T :: TYPE r -> TYPE (Id r)
- type family T a where
+We don't need to eta-expand `g` to make `f g` typecheck; a cast
+suffices. Hence we should not perform representation-polymorphism
+checks; they would fail here.
- test :: forall r (a :: TYPE r). a :~~: T a -> ()
- test HRefl =
- let
- f :: (a -> a) -> ()
- f _ = ()
- g :: T a -> T a
- g = undefined
- in f g
-
- We don't need to eta-expand `g` to make `f g` typecheck; a cast suffices.
- Hence we should not perform representation-polymorphism checks; they would
- fail here.
+All this is done by `mkWpFun_FRR`, which checks for the cast/cast case and
+returns a `FunCo` if so.
Note [Setting the argument context]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
@@ -1947,7 +1950,7 @@ getDeepSubsumptionFlag = do { ds <- xoptM LangExt.DeepSubsumption
-- | 'tc_sub_type_deep' is where the actual work happens for deep subsumption.
--
-- Given @ty_actual@ (a sigma-type) and @ty_expected@ (deeply skolemised, i.e.
--- a deep rho type), it returns an 'HsWrapper' @wrap :: ty_actual ~> ty_expected@.
+-- a deep rho type), it returns an 'HsWrapper' @wrap :: ty_actual ~~> ty_expected@.
tc_sub_type_deep :: HasDebugCallStack
=> Position p -- ^ Position in the type (for error messages only)
-> (TcType -> TcType -> TcM TcCoercionN) -- ^ How to unify
@@ -1958,7 +1961,7 @@ tc_sub_type_deep :: HasDebugCallStack
-> TcM HsWrapper
-- If wrap = tc_sub_type_deep t1 t2
--- => wrap :: t1 ~> t2
+-- => wrap :: t1 ~~> t2
-- Here is where the work actually happens!
-- Precondition: ty_expected is deeply skolemised
@@ -2015,8 +2018,8 @@ tc_sub_type_deep pos unify inst_orig ctxt ty_actual ty_expected
; unify_wrap <- just_unify exp_funTy ty_e
; fun_wrap <- go_fun af1 act_mult act_arg act_res af1 exp_mult exp_arg exp_res
; return $ unify_wrap <.> fun_wrap
- -- unify_wrap :: exp_funTy ~> ty_e
- -- fun_wrap :: ty_a ~> exp_funTy
+ -- unify_wrap :: exp_funTy ~~> ty_e
+ -- fun_wrap :: ty_a ~~> exp_funTy
}
go1 ty_a (FunTy { ft_af = af2, ft_mult = exp_mult, ft_arg = exp_arg, ft_res = exp_res })
| isVisibleFunArg af2
@@ -2028,8 +2031,8 @@ tc_sub_type_deep pos unify inst_orig ctxt ty_actual ty_expected
; unify_wrap <- just_unify ty_a act_funTy
; fun_wrap <- go_fun af2 act_mult act_arg act_res af2 exp_mult exp_arg exp_res
; return $ fun_wrap <.> unify_wrap
- -- unify_wrap :: ty_a ~> act_funTy
- -- fun_wrap :: act_funTy ~> ty_e
+ -- unify_wrap :: ty_a ~~> act_funTy
+ -- fun_wrap :: act_funTy ~~> ty_e
}
-- Otherwise, revert to unification.
@@ -2064,17 +2067,28 @@ mkWpFun_FRR
-> Position p
-> FunTyFlag -> Type -> TcType -> Type -- actual FunTy
-> FunTyFlag -> Type -> TcType -> Type -- expected FunTy
- -> HsWrapper -- ^ exp_arg ~> act_arg
- -> HsWrapper -- ^ act_res ~> exp_res
- -> TcM HsWrapper -- ^ act_funTy ~> exp_funTy
+ -> HsWrapper -- ^ exp_arg ~~> act_arg
+ -> HsWrapper -- ^ act_res ~~> exp_res
+ -> TcM HsWrapper -- ^ (act_arg->act_res) ~~> (exp_arg->exp_res)
mkWpFun_FRR unify pos act_af act_mult act_arg act_res exp_af exp_mult exp_arg exp_res arg_wrap res_wrap
- = do { ((exp_arg_co, exp_arg_frr), (act_arg_co, _act_arg_frr)) <-
- if needs_frr_checks
- -- See Wrinkle [Representation-polymorphism checking during subtyping]
- then do { exp_frr_wrap <- hasFixedRuntimeRep (frr_ctxt True ) exp_arg
- ; act_frr_wrap <- hasFixedRuntimeRep (frr_ctxt False) act_arg
- ; return (exp_frr_wrap, act_frr_wrap) }
- else return ((mkNomReflCo exp_arg, exp_arg), (mkNomReflCo act_arg, act_arg))
+ | Just arg_co <- getWpCo_maybe arg_wrap act_arg -- arg_co :: exp_arg ~R# act_arg
+ , Just res_co <- getWpCo_maybe res_wrap act_res -- res_co :: act_res ~R# exp_res
+ = -- The argument and result wrappers are both hole or cast;
+ -- so we can make do with a FunCo
+ -- See Note [Representation-polymorphism checking during subtyping]
+ do { mult_co <- unify act_mult exp_mult
+ ; let the_co = mkFunCo2 Representational act_af exp_af mult_co (mkSymCo arg_co) res_co
+ ; return (mkWpCastR the_co) }
+
+ | otherwise
+ = -- We need a full WpFun, with the eta-expansion that it entails
+ -- And hence we must add fixed-runtime-rep checks so that the eta-expansion is OK
+ -- See Note [Representation-polymorphism checking during subtyping]
+ do { (exp_arg_co, exp_arg_frr) <- hasFixedRuntimeRep (frr_ctxt True ) exp_arg
+ ; (act_arg_co, _act_arg_frr) <- hasFixedRuntimeRep (frr_ctxt False) act_arg
+ -- exp_arg_frr, act_arg_frr :: Type have fixed runtime-reps
+ -- exp_arg_co :: exp_arg ~ exp_arg_frr Usually Refl
+ -- act_arg_co :: act_arg ~ act_arg_frr Usually Refl
-- Enforce equality of multiplicities (not the more natural sub-multiplicity).
-- See Note [Multiplicity in deep subsumption]
@@ -2083,46 +2097,36 @@ mkWpFun_FRR unify pos act_af act_mult act_arg act_res exp_af exp_mult exp_arg ex
-- equality to be Refl, but it might well not be (#26332).
; let
- exp_arg_fun_co =
+ exp_arg_fun_co = -- (exp_arg_frr -> exp_res) ~ (exp_arg -> exp_res)
mkFunCo Nominal exp_af
- (mkReflCo Nominal exp_mult)
+ (mkNomReflCo exp_mult)
(mkSymCo exp_arg_co)
- (mkReflCo Nominal exp_res)
- act_arg_fun_co =
+ (mkNomReflCo exp_res)
+ act_arg_fun_co = -- (act_arg -> act_res) ~ (act_arg_frr -> act_res)
mkFunCo Nominal act_af
act_arg_mult_co
act_arg_co
- (mkReflCo Nominal act_res)
- arg_wrap_frr =
+ (mkNomReflCo act_res)
+ arg_wrap_frr = -- exp_arg_frr ~~> act_arg_frr
mkWpCastN (mkSymCo exp_arg_co) <.> arg_wrap <.> mkWpCastN act_arg_co
- -- exp_arg_co :: exp_arg ~> exp_arg_frr
- -- act_arg_co :: act_arg ~> act_arg_frr
- -- arg_wrap :: exp_arg ~> act_arg
- -- arg_wrap_frr :: exp_arg_frr ~> act_arg_frr
- ; return $
- mkWpCastN exp_arg_fun_co
+ ; return $ -- Whole thing :: (act_arg->act_res) ~~> (exp_arg->exp_ress)
+ mkWpCastN exp_arg_fun_co -- (exp_ar_frr->exp_res) ~~> (exp_arg->exp_res)
<.>
mkWpFun arg_wrap_frr res_wrap (Scaled exp_mult exp_arg_frr) exp_res
- <.>
- mkWpCastN act_arg_fun_co
+ <.> -- (act_arg_frr->act_res) ~~> (exp_arg_frr->exp_res)
+ mkWpCastN act_arg_fun_co -- (act_arg->act_res) ~~> (act_arg_frr->act_res)
}
where
- needs_frr_checks :: Bool
- needs_frr_checks =
- not (hole_or_cast arg_wrap)
- ||
- not (hole_or_cast res_wrap)
- hole_or_cast :: HsWrapper -> Bool
- hole_or_cast WpHole = True
- hole_or_cast (WpCast {}) = True
- hole_or_cast _ = False
+ getWpCo_maybe :: HsWrapper -> Type -> Maybe CoercionR
+ -- See if a HsWrapper is just a coercion
+ getWpCo_maybe WpHole ty = Just (mkRepReflCo ty)
+ getWpCo_maybe (WpCast co) _ = Just co
+ getWpCo_maybe _ _ = Nothing
+
frr_ctxt :: Bool -> FixedRuntimeRepContext
- frr_ctxt is_exp_ty =
- FRRDeepSubsumption
- { frrDSExpected = is_exp_ty
- , frrDSPosition = pos
- }
+ frr_ctxt is_exp_ty = FRRDeepSubsumption { frrDSExpected = is_exp_ty
+ , frrDSPosition = pos }
-----------------------
deeplySkolemise :: SkolemInfo -> TcSigmaType
@@ -2146,9 +2150,9 @@ deeplySkolemise skol_info ty
; let tvs = binderVars bndrs
tvs1 = binderVars bndrs1
tv_prs1 = map tyVarName tvs `zip` bndrs1
- ; return ( mkWpEta ids1 (mkWpTyLams tvs1
- <.> mkWpEvLams ev_vars1
- <.> wrap)
+ ; return ( mkWpEta ty ids1 (mkWpTyLams tvs1
+ <.> mkWpEvLams ev_vars1
+ <.> wrap)
, tv_prs1 ++ tvs_prs2
, ev_vars1 ++ ev_vars2
, mkScaledFunTys arg_tys' rho ) }
@@ -2182,7 +2186,7 @@ deeplyInstantiate orig ty
; ids1 <- newSysLocalIds (fsLit "di") arg_tys'
; wrap1 <- instCall orig (mkTyVarTys tvs') theta'
; (wrap2, rho2) <- go subst' rho
- ; return (mkWpEta ids1 (wrap2 <.> wrap1),
+ ; return (mkWpEta ty ids1 (wrap2 <.> wrap1),
mkScaledFunTys arg_tys' rho2) }
| otherwise
=====================================
compiler/GHC/Tc/Zonk/Type.hs
=====================================
@@ -1233,13 +1233,16 @@ zonk_cmd_top (HsCmdTop (CmdTopTc stack_tys ty ids) cmd)
-------------------------------------------------------------------------
zonkCoFn :: HsWrapper -> ZonkBndrTcM HsWrapper
zonkCoFn WpHole = return WpHole
+zonkCoFn (WpSubType w) = do { w' <- zonkCoFn w
+ ; return (WpSubType w') }
zonkCoFn (WpCompose c1 c2) = do { c1' <- zonkCoFn c1
; c2' <- zonkCoFn c2
; return (WpCompose c1' c2') }
-zonkCoFn (WpFun c1 c2 t1) = do { c1' <- zonkCoFn c1
- ; c2' <- zonkCoFn c2
- ; t1' <- noBinders $ zonkScaledTcTypeToTypeX t1
- ; return (WpFun c1' c2' t1') }
+zonkCoFn (WpFun c1 c2 t1 t2) = do { c1' <- zonkCoFn c1
+ ; c2' <- zonkCoFn c2
+ ; t1' <- noBinders $ zonkScaledTcTypeToTypeX t1
+ ; t2' <- noBinders $ zonkTcTypeToTypeX t2
+ ; return (WpFun c1' c2' t1' t2') }
zonkCoFn (WpCast co) = WpCast <$> noBinders (zonkCoToCo co)
zonkCoFn (WpEvLam ev) = WpEvLam <$> zonkEvBndrX ev
zonkCoFn (WpEvApp arg) = WpEvApp <$> noBinders (zonkEvTerm arg)
=====================================
testsuite/tests/simplCore/should_compile/T26349.hs
=====================================
@@ -0,0 +1,11 @@
+{-# LANGUAGE DeepSubsumption, RankNTypes #-}
+module T26349 where
+
+{-# SPECIALIZE INLINE mapTCMT :: (forall b. IO b -> IO b) -> IO a -> IO a #-}
+mapTCMT :: (forall b. m b -> n b) -> m a -> n a
+mapTCMT f m = f m
+
+{-
+ We'll check
+ tcExpr (mapTCMT) (Check ((forall b. IO b -> IO b) -> IO a_sk -> IO a_sk))
+-}
=====================================
testsuite/tests/simplCore/should_compile/T26349.stderr
=====================================
@@ -0,0 +1,3 @@
+==================== Tidy Core rules ====================
+"USPEC mapTCMT @(*) @IO @IO @_"
+ forall (@a). mapTCMT @(*) @IO @IO @a = mapTCMT_$smapTCMT @a
=====================================
testsuite/tests/simplCore/should_compile/all.T
=====================================
@@ -559,3 +559,4 @@ test('T26051', [ grep_errmsg(r'\$wspecMe')
test('T26115', [grep_errmsg(r'DFun')], compile, ['-O -ddump-simpl -dsuppress-uniques'])
test('T26116', normal, compile, ['-O -ddump-rules'])
test('T26117', [grep_errmsg(r'==')], compile, ['-O -ddump-simpl -dsuppress-uniques'])
+test('T26349', normal, compile, ['-O -ddump-rules'])
=====================================
testsuite/tests/simplCore/should_compile/rule2.stderr
=====================================
@@ -10,18 +10,15 @@
==================== Grand total simplifier statistics ====================
-Total ticks: 13
+Total ticks: 11
-2 PreInlineUnconditionally
- 1 ds
- 1 f
+1 PreInlineUnconditionally 1 f
2 UnfoldingDone
1 GHC.Internal.Base.id
1 Roman.bar
1 RuleFired 1 foo/bar
1 LetFloatFromLet 1
-7 BetaReduction
- 1 ds
+6 BetaReduction
1 f
1 a
1 m
View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/a7c1c88a2d0cde2ae7fe75c14abd5d…
--
View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/a7c1c88a2d0cde2ae7fe75c14abd5d…
You're receiving this email because of your account on gitlab.haskell.org.
1
0
[Git][ghc/ghc][wip/T24464] 16 commits: Refactor fundep solving
by Simon Peyton Jones (＠simonpj) 04 Nov '25
by Simon Peyton Jones (＠simonpj) 04 Nov '25
04 Nov '25
Simon Peyton Jones pushed to branch wip/T24464 at Glasgow Haskell Compiler / GHC
Commits:
fcf6331e by Richard Eisenberg at 2025-11-03T08:33:05+00:00
Refactor fundep solving
This commit is a large-scale refactor of the increasingly-messy code that
handles functional dependencies. It has virtually no effect on what compiles
but improves error messages a bit. And it does the groundwork for #23162.
The big picture is described in
Note [Overview of functional dependencies in type inference]
in GHC.Tc.Solver.FunDeps
* New module GHC.Tc.Solver.FunDeps contains all the fundep-handling
code for the constraint solver.
* Fundep-equalities are solved in a nested scope; they may generate
unifications but otherwise have no other effect.
See GHC.Tc.Solver.FunDeps.solveFunDeps
The nested needs to start from the Givens in the inert set, but
not the Wanteds; hence a new function `resetInertCans`, used in
`nestFunDepsTcS`.
* That in turn means that fundep equalities never show up in error
messages, so the complicated FunDepOrigin tracking can all disappear.
* We need to be careful about tracking unifications, so we kick out
constraints from the inert set after doing unifications. Unification
tracking has been majorly reformed: see Note [WhatUnifications] in
GHC.Tc.Utils.Unify.
A good consequence is that the hard-to-grok `resetUnificationFlag`
has been replaced with a simpler use of
`reportCoarseGrainUnifications`
Smaller things:
* Rename `FunDepEqn` to `FunDepEqns` since it contains multiple
type equalities.
Some compile time improvement
Metrics: compile_time/bytes allocated
Baseline
Test value New value Change
---------------------- --------------------------------------
T5030(normal) 173,839,232 148,115,248 -14.8% GOOD
hard_hole_fits(normal) 286,768,048 284,015,416 -1.0%
geo. mean -0.2%
minimum -14.8%
maximum +0.3%
Metric Decrease:
T5030
- - - - -
231adc30 by Simon Peyton Jones at 2025-11-03T08:33:05+00:00
QuickLook's tcInstFun should make instantiation variables directly
tcInstFun must make "instantiation variables", not regular
unification variables, when instantiating function types. That was
previously implemented by a hack: set the /ambient/ level to QLInstTyVar.
But the hack finally bit me, when I was refactoring WhatUnifications.
And it was always wrong: see the now-expunged (TCAPP2) note.
This commit does it right, by making tcInstFun call its own
instantiation functions. That entails a small bit of duplication,
but the result is much, much cleaner.
- - - - -
39d4a24b by Simon Peyton Jones at 2025-11-03T08:33:05+00:00
Build implication for constraints from (static e)
This commit addresses #26466, by buiding an implication for the
constraints arising from a (static e) form. The implication has
a special ic_info field of StaticFormSkol, which tells the constraint
solver to use an empty set of Givens.
See (SF3) in Note [Grand plan for static forms]
in GHC.Iface.Tidy.StaticPtrTable
This commit also reinstates an `assert` in GHC.Tc.Solver.Equality.
The test `StaticPtrTypeFamily` was failing with an assertion failure,
but it now works.
- - - - -
2e2aec1e by Simon Peyton Jones at 2025-11-03T08:33:05+00:00
Comments about defaulting representation equalities
- - - - -
52a4d1da by Simon Peyton Jones at 2025-11-03T08:33:05+00:00
Improve tracking of rewriter-sets
This refactor substantially improves the treatment of so-called
"rewriter-sets" in the constraint solver.
The story is described in the rewritten
Note [Wanteds rewrite Wanteds: rewriter-sets]
in GHC.Tc.Types.Constraint
Some highlights
* Trace the free coercion holes of a filled CoercionHole,
in CoercionPlusHoles. See Note [Coercion holes] (COH5)
This avoids taking having to take the free coercion variables
of a coercion when zonking a rewrriter-set
* Many knock on changes
* Make fillCoercionHole take CoercionPlusHoles as its argument
rather than to separate arguments.
* Similarly setEqIfWanted, setWantedE, wrapUnifierAndEmit.
* Be more careful about passing the correct CoHoleSet to
`rewriteEqEvidence` and friends
* Make kickOurAfterFillingCoercionHole more clever. See
new Note [Kick out after filling a coercion hole]
Smaller matters
* Rename RewriterSet to CoHoleSet
* Add special-case helper `rewriteEqEvidenceSwapOnly`
- - - - -
3e78e1ba by Simon Peyton Jones at 2025-11-03T08:33:05+00:00
Tidy up constraint solving for foralls
* In `can_eq_nc_forall` make sure to track Givens that are used
in the nested solve step.
* Tiny missing-swap bug-fix in `lookup_eq_in_qcis`
* Fix some leftover mess from
commit 14123ee646f2b9738a917b7cec30f9d3941c13de
Author: Simon Peyton Jones <simon.peytonjones(a)gmail.com>
Date: Wed Aug 20 00:35:48 2025 +0100
Solve forall-constraints via an implication, again
Specifically, trySolveImplication is now dead.
- - - - -
973f2c25 by Simon Peyton Jones at 2025-11-03T08:33:05+00:00
Do not treat CoercionHoles as free variables in coercions
This fixes a long-standing wart in the free-variable finder;
now CoercionHoles are no longer treated as a "free variable"
of a coercion.
I got big and unexpected performance regressions when making
this change. Turned out that CallArity didn't discover that
the free variable finder could be eta-expanded, which gave very
poor code.
So I re-used Note [The one-shot state monad trick] for Endo,
resulting in GHC.Utils.EndoOS. Very simple, big win.
- - - - -
c2b8a0f9 by Simon Peyton Jones at 2025-11-03T08:33:05+00:00
Update debug-tracing in CallArity
No effect on behaviour, and commented out anyway
- - - - -
9aa5ee99 by Simon Peyton Jones at 2025-11-03T08:33:28+00:00
Comments only -- remove dangling Note references
- - - - -
6683f183 by Simon Peyton Jones at 2025-11-03T08:33:28+00:00
Accept error message wibbles
- - - - -
3ba3d9f9 by Luite Stegeman at 2025-11-04T00:59:41-05:00
rts: fix eager black holes: record mutated closure and fix assertion
This fixes two problems with handling eager black holes, introduced
by a1de535f762bc23d4cf23a5b1853591dda12cdc9.
- the closure mutation must be recorded even for eager black holes,
since the mutator has mutated it before calling threadPaused
- The assertion that an unmarked eager black hole must be owned by
the TSO calling threadPaused is incorrect, since multiple threads
can race to claim the black hole.
fixes #26495
- - - - -
cd972168 by Simon Peyton Jones at 2025-11-04T16:07:57+00:00
Start at moving static forms straight to top level
Just for CI. Needs documentation.
- - - - -
ce741189 by Simon Peyton Jones at 2025-11-04T16:07:57+00:00
Wibble
- - - - -
2134952a by Simon Peyton Jones at 2025-11-04T16:07:57+00:00
Fix desugaring
- - - - -
40ab10aa by Simon Peyton Jones at 2025-11-04T16:07:57+00:00
Wibbles
- - - - -
4b67d5ad by Simon Peyton Jones at 2025-11-04T17:42:42+00:00
Work in progress [skip ci]
- - - - -
108 changed files:
- compiler/GHC/Core/Opt/CallArity.hs
- compiler/GHC/Core/Opt/Simplify/Iteration.hs
- compiler/GHC/Core/Opt/Simplify/Utils.hs
- compiler/GHC/Core/TyCo/FVs.hs
- compiler/GHC/Core/TyCo/Rep.hs
- compiler/GHC/Core/TyCon.hs
- compiler/GHC/Hs/Binds.hs
- compiler/GHC/Hs/Decls.hs
- compiler/GHC/Hs/Instances.hs
- compiler/GHC/Hs/Utils.hs
- compiler/GHC/HsToCore.hs
- compiler/GHC/HsToCore/Expr.hs
- compiler/GHC/HsToCore/Monad.hs
- compiler/GHC/HsToCore/Types.hs
- compiler/GHC/Iface/Tidy/StaticPtrTable.hs
- compiler/GHC/Runtime/Eval.hs
- compiler/GHC/Tc/Deriv.hs
- compiler/GHC/Tc/Errors.hs
- compiler/GHC/Tc/Errors/Ppr.hs
- compiler/GHC/Tc/Errors/Types.hs
- compiler/GHC/Tc/Gen/App.hs
- compiler/GHC/Tc/Gen/Bind.hs
- compiler/GHC/Tc/Gen/Expr.hs
- compiler/GHC/Tc/Gen/Sig.hs
- compiler/GHC/Tc/Instance/FunDeps.hs
- compiler/GHC/Tc/Solver.hs
- compiler/GHC/Tc/Solver/Default.hs
- compiler/GHC/Tc/Solver/Dict.hs
- compiler/GHC/Tc/Solver/Equality.hs
- + compiler/GHC/Tc/Solver/FunDeps.hs
- compiler/GHC/Tc/Solver/InertSet.hs
- compiler/GHC/Tc/Solver/Irred.hs
- compiler/GHC/Tc/Solver/Monad.hs
- compiler/GHC/Tc/Solver/Rewrite.hs
- compiler/GHC/Tc/Solver/Solve.hs
- compiler/GHC/Tc/Solver/Solve.hs-boot
- compiler/GHC/Tc/TyCl/Instance.hs
- compiler/GHC/Tc/Types.hs
- compiler/GHC/Tc/Types/BasicTypes.hs
- compiler/GHC/Tc/Types/Constraint.hs
- compiler/GHC/Tc/Types/Evidence.hs
- compiler/GHC/Tc/Types/Origin.hs
- compiler/GHC/Tc/Utils/Concrete.hs
- compiler/GHC/Tc/Utils/Env.hs
- compiler/GHC/Tc/Utils/Instantiate.hs
- compiler/GHC/Tc/Utils/Monad.hs
- compiler/GHC/Tc/Utils/TcMType.hs
- compiler/GHC/Tc/Utils/TcType.hs
- compiler/GHC/Tc/Utils/Unify.hs
- compiler/GHC/Tc/Zonk/TcType.hs
- compiler/GHC/Tc/Zonk/Type.hs
- compiler/GHC/Types/Basic.hs
- compiler/GHC/Types/Unique/DSM.hs
- + compiler/GHC/Utils/EndoOS.hs
- compiler/ghc.cabal.in
- rts/ThreadPaused.c
- testsuite/tests/count-deps/CountDepsAst.stdout
- testsuite/tests/count-deps/CountDepsParser.stdout
- testsuite/tests/default/default-fail05.stderr
- testsuite/tests/dependent/should_fail/T13135_simple.stderr
- testsuite/tests/deriving/should_fail/T3621.stderr
- testsuite/tests/indexed-types/should_fail/T14369.stderr
- testsuite/tests/indexed-types/should_fail/T1897b.stderr
- testsuite/tests/linters/notes.stdout
- testsuite/tests/parser/should_fail/RecordDotSyntaxFail10.stderr
- testsuite/tests/parser/should_fail/RecordDotSyntaxFail13.stderr
- testsuite/tests/parser/should_fail/T20654a.stderr
- testsuite/tests/partial-sigs/should_fail/T14584a.stderr
- testsuite/tests/polykinds/T6068.stdout
- testsuite/tests/quantified-constraints/T15359.hs
- testsuite/tests/rep-poly/RepPolyNPlusK.stderr
- testsuite/tests/rep-poly/RepPolyRightSection.stderr
- testsuite/tests/rep-poly/T13233.stderr
- testsuite/tests/rep-poly/T19709b.stderr
- testsuite/tests/rep-poly/T23903.stderr
- testsuite/tests/typecheck/no_skolem_info/T13499.stderr
- testsuite/tests/typecheck/should_compile/T13651.hs
- − testsuite/tests/typecheck/should_compile/T13651.stderr
- + testsuite/tests/typecheck/should_compile/T14745.hs
- testsuite/tests/typecheck/should_compile/all.T
- testsuite/tests/typecheck/should_compile/hole_constraints_nested.stderr
- testsuite/tests/typecheck/should_compile/tc126.hs
- testsuite/tests/typecheck/should_fail/AmbigFDs.hs
- − testsuite/tests/typecheck/should_fail/AmbigFDs.stderr
- testsuite/tests/typecheck/should_fail/FD3.stderr
- testsuite/tests/typecheck/should_fail/FDsFromGivens2.stderr
- testsuite/tests/typecheck/should_fail/T13506.stderr
- testsuite/tests/typecheck/should_fail/T16512a.stderr
- testsuite/tests/typecheck/should_fail/T18851b.hs
- − testsuite/tests/typecheck/should_fail/T18851b.stderr
- testsuite/tests/typecheck/should_fail/T18851c.hs
- − testsuite/tests/typecheck/should_fail/T18851c.stderr
- testsuite/tests/typecheck/should_fail/T19415.stderr
- testsuite/tests/typecheck/should_fail/T19415b.stderr
- testsuite/tests/typecheck/should_fail/T22684.stderr
- + testsuite/tests/typecheck/should_fail/T23162a.hs
- + testsuite/tests/typecheck/should_fail/T23162a.stderr
- testsuite/tests/typecheck/should_fail/T25325.stderr
- testsuite/tests/typecheck/should_fail/T5246.stderr
- testsuite/tests/typecheck/should_fail/T5978.stderr
- testsuite/tests/typecheck/should_fail/T7368a.stderr
- testsuite/tests/typecheck/should_fail/T7696.stderr
- testsuite/tests/typecheck/should_fail/T8603.stderr
- testsuite/tests/typecheck/should_fail/T9612.stderr
- testsuite/tests/typecheck/should_fail/TcStaticPointersFail03.stderr
- testsuite/tests/typecheck/should_fail/all.T
- testsuite/tests/typecheck/should_fail/tcfail122.stderr
- testsuite/tests/typecheck/should_fail/tcfail143.stderr
The diff was not included because it is too large.
View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/07a053cbfc0b9263888787649503fc…
--
View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/07a053cbfc0b9263888787649503fc…
You're receiving this email because of your account on gitlab.haskell.org.
1
0
[Git][ghc/ghc][wip/spj-apporv-Oct24] rename fun_ctxt to fun_lspan, fun_orig passed in tcInstFun to default to app...
by Apoorv Ingle (＠ani) 04 Nov '25
by Apoorv Ingle (＠ani) 04 Nov '25
04 Nov '25
Apoorv Ingle pushed to branch wip/spj-apporv-Oct24 at Glasgow Haskell Compiler / GHC
Commits:
82766264 by Apoorv Ingle at 2025-11-04T09:57:14-06:00
rename fun_ctxt to fun_lspan, fun_orig passed in tcInstFun to default to app chain head if its user located, fall back to srcCodeOrigin if it's a generated location
- - - - -
2 changed files:
- compiler/GHC/Tc/Gen/App.hs
- compiler/GHC/Tc/Gen/Head.hs
Changes:
=====================================
compiler/GHC/Tc/Gen/App.hs
=====================================
@@ -178,14 +178,17 @@ Note [Instantiation variables are short lived]
-- take in the rn_expr and its location to pass into tcValArgs
tcExprSigma :: Bool -> HsExpr GhcRn -> TcM (HsExpr GhcTc, TcSigmaType)
tcExprSigma inst rn_expr
- = do { (fun@(rn_fun,fun_ctxt), rn_args) <- splitHsApps rn_expr
+ = do { (fun@(rn_fun,fun_lspan), rn_args) <- splitHsApps rn_expr
; do_ql <- wantQuickLook rn_fun
; (tc_fun, fun_sigma) <- tcInferAppHead fun
; code_orig <- getSrcCodeOrigin
- ; let fun_orig = srcCodeOriginCtOrigin rn_expr code_orig
- ; (inst_args, app_res_sigma) <- tcInstFun do_ql inst (fun_orig, rn_fun, fun_ctxt) tc_fun fun_sigma rn_args
- ; tc_args <- tcValArgs do_ql (rn_fun, generatedSrcSpan) inst_args
- ; let tc_expr = rebuildHsApps (tc_fun, fun_ctxt) tc_args
+ ; let fun_orig | isGoodSrcSpan fun_lspan
+ = exprCtOrigin rn_fun
+ | otherwise
+ = srcCodeOriginCtOrigin rn_fun code_orig
+ ; (inst_args, app_res_sigma) <- tcInstFun do_ql inst (fun_orig, rn_fun, fun_lspan) tc_fun fun_sigma rn_args
+ ; tc_args <- tcValArgs do_ql (rn_fun, fun_lspan) inst_args
+ ; let tc_expr = rebuildHsApps (tc_fun, fun_lspan) tc_args
; return (tc_expr, app_res_sigma) }
@@ -2239,7 +2242,7 @@ tcTagToEnum :: (HsExpr GhcTc, SrcSpan) -> [HsExprArg 'TcpTc]
-> TcM (HsExpr GhcTc)
-- tagToEnum# :: forall a. Int# -> a
-- See Note [tagToEnum#] Urgh!
-tcTagToEnum (tc_fun, fun_ctxt) tc_args res_ty
+tcTagToEnum (tc_fun, fun_lspan) tc_args res_ty
| [val_arg] <- dropWhile (not . isHsValArg) tc_args
= do { res_ty <- liftZonkM $ zonkTcType res_ty
@@ -2261,14 +2264,14 @@ tcTagToEnum (tc_fun, fun_ctxt) tc_args res_ty
; let rep_ty = mkTyConApp rep_tc rep_args
tc_fun' = mkHsWrap (WpTyApp rep_ty) tc_fun
df_wrap = mkWpCastR (mkSymCo coi)
- tc_expr = rebuildHsApps (tc_fun', fun_ctxt) [val_arg]
+ tc_expr = rebuildHsApps (tc_fun', fun_lspan) [val_arg]
; return (mkHsWrap df_wrap tc_expr) }}}}}
| otherwise
= failWithTc TcRnTagToEnumMissingValArg
where
- vanilla_result = return (rebuildHsApps (tc_fun, fun_ctxt) tc_args)
+ vanilla_result = return (rebuildHsApps (tc_fun, fun_lspan) tc_args)
check_enumeration ty' tc
| -- isTypeDataTyCon: see wrinkle (W1) in
=====================================
compiler/GHC/Tc/Gen/Head.hs
=====================================
@@ -457,7 +457,7 @@ tcInferAppHead :: (HsExpr GhcRn, SrcSpan)
--
-- See Note [tcApp: typechecking applications] in GHC.Tc.Gen.App
tcInferAppHead (fun,fun_lspan)
- = addLExprCtxt fun_lspan fun $
+ = setSrcSpan fun_lspan $
do { mb_tc_fun <- tcInferAppHead_maybe fun
; case mb_tc_fun of
Just (fun', fun_sigma) -> return (fun', fun_sigma)
@@ -471,10 +471,10 @@ tcInferAppHead_maybe fun =
case fun of
HsVar _ nm -> Just <$> tcInferId nm
XExpr (HsRecSelRn f) -> Just <$> tcInferRecSelId f
- XExpr (ExpandedThingRn _ e) -> Just <$> -- (addExpansionErrCtxt o (srcCodeOriginErrCtxMsg o) $
+ XExpr (ExpandedThingRn o e) -> Just <$> (addExpansionErrCtxt o (srcCodeOriginErrCtxMsg o) $
-- We do not want to instantiate c.f. T19167
tcExprSigma False e
- -- )
+ )
ExprWithTySig _ e hs_ty -> Just <$> tcExprWithSig e hs_ty
HsOverLit _ lit -> Just <$> tcInferOverLit lit
_ -> return Nothing
View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/82766264d33b174dd966d6efa7d3ec4…
--
View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/82766264d33b174dd966d6efa7d3ec4…
You're receiving this email because of your account on gitlab.haskell.org.
1
0
[Git][ghc/ghc][wip/backports-9.14] level imports: Fix infinite loop with cyclic module imports
by Zubin (＠wz1000) 04 Nov '25
by Zubin (＠wz1000) 04 Nov '25
04 Nov '25
Zubin pushed to branch wip/backports-9.14 at Glasgow Haskell Compiler / GHC
Commits:
0e08b9ae by Matthew Pickering at 2025-11-04T20:56:40+05:30
level imports: Fix infinite loop with cyclic module imports
I didn't anticipate that downsweep would run before we checked for
cyclic imports. Therefore we need to use the reachability function which
handles cyclic graphs.
Fixes #26087
(cherry picked from commit 8b731e3c900291655a767123bcda55eddd63920c)
- - - - -
5 changed files:
- compiler/GHC/Unit/Module/Graph.hs
- + testsuite/tests/splice-imports/T26087.stderr
- + testsuite/tests/splice-imports/T26087A.hs
- + testsuite/tests/splice-imports/T26087B.hs
- testsuite/tests/splice-imports/all.T
Changes:
=====================================
compiler/GHC/Unit/Module/Graph.hs
=====================================
@@ -866,7 +866,7 @@ mkTransZeroDeps = first graphReachability {- module graph is acyclic -} . module
-- | Transitive dependencies, but with the stage that each module is required at.
mkStageDeps :: [ModuleGraphNode] -> (ReachabilityIndex StageSummaryNode, (NodeKey, ModuleStage) -> Maybe StageSummaryNode)
-mkStageDeps = first graphReachability . moduleGraphNodesStages
+mkStageDeps = first cyclicGraphReachability . moduleGraphNodesStages
type ZeroSummaryNode = Node Int ZeroScopeKey
=====================================
testsuite/tests/splice-imports/T26087.stderr
=====================================
@@ -0,0 +1,6 @@
+./T26087B.hs: error: [GHC-92213]
+ Module graph contains a cycle:
+ module ‘main:T26087B’ (./T26087B.hs)
+ imports module ‘main:T26087A’ (T26087A.hs)
+ which imports module ‘main:T26087B’ (./T26087B.hs)
+
=====================================
testsuite/tests/splice-imports/T26087A.hs
=====================================
@@ -0,0 +1,4 @@
+{-# LANGUAGE ExplicitLevelImports #-}
+module T26087A where
+
+import quote T26087B
=====================================
testsuite/tests/splice-imports/T26087B.hs
=====================================
@@ -0,0 +1,4 @@
+{-# LANGUAGE ExplicitLevelImports, TemplateHaskell #-}
+module T26087B where
+
+import T26087A
=====================================
testsuite/tests/splice-imports/all.T
=====================================
@@ -46,7 +46,4 @@ test('SI35',
compile_and_run,
['-package ghc'])
test('SI36', [extra_files(["SI36_A.hs", "SI36_B1.hs", "SI36_B2.hs", "SI36_B3.hs", "SI36_C1.hs", "SI36_C2.hs", "SI36_C3.hs"])], multimod_compile_fail, ['SI36', '-v0'])
-test('T26090', [], multimod_compile_fail, ['T26090', '-v0'])
-test('ModuleExport', [], multimod_compile_fail, ['ModuleExport', '-v0'])
-test('LevelImportExports', [], makefile_test, [])
-test('DodgyLevelExport', [], multimod_compile, ['DodgyLevelExport', '-v0 -Wdodgy-exports'])
+test('T26087', [], multimod_compile_fail, ['T26087A', ''])
View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/0e08b9aee6ce58fe7099f14ac1d7dea…
--
View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/0e08b9aee6ce58fe7099f14ac1d7dea…
You're receiving this email because of your account on gitlab.haskell.org.
1
0
[Git][ghc/ghc][wip/romes/26202] 4 commits: build: Relax ghc/ghc-boot Cabal bound to 3.16
by Rodrigo Mesquita (＠alt-romes) 04 Nov '25
by Rodrigo Mesquita (＠alt-romes) 04 Nov '25
04 Nov '25
Rodrigo Mesquita pushed to branch wip/romes/26202 at Glasgow Haskell Compiler / GHC
Commits:
b5508f2c by Rodrigo Mesquita at 2025-11-04T14:10:56+00:00
build: Relax ghc/ghc-boot Cabal bound to 3.16
Fixes #26202
- - - - -
c5b3541f by Rodrigo Mesquita at 2025-11-04T14:10:56+00:00
cabal-reinstall: Use haddock-api +in-tree-ghc
Fixes #26202
- - - - -
c6d4b945 by Rodrigo Mesquita at 2025-11-04T14:10:56+00:00
cabal-reinstall: Pass --strict to Happy
This is necessary to make the generated Parser build successfully
This mimics Hadrian, which always passes --strict to happy.
Fixes #26202
- - - - -
79df1e0e by Rodrigo Mesquita at 2025-11-04T14:10:56+00:00
genprimopcode: Require higher happy version
I've bumped the happy version to forbid deprecated Happy versions which
don't successfully compile.
- - - - -
6 changed files:
- cabal.project-reinstall
- compiler/Setup.hs
- compiler/ghc.cabal.in
- libraries/ghc-boot/Setup.hs
- libraries/ghc-boot/ghc-boot.cabal.in
- utils/genprimopcode/genprimopcode.cabal
Changes:
=====================================
cabal.project-reinstall
=====================================
@@ -59,6 +59,7 @@ constraints: ghc +internal-interpreter +dynamic-system-linke,
ghc-bin +internal-interpreter +threaded,
ghci +internal-interpreter,
haddock +in-ghc-tree,
+ haddock-api +in-ghc-tree,
any.array installed,
any.base installed,
any.deepseq installed,
@@ -68,6 +69,8 @@ constraints: ghc +internal-interpreter +dynamic-system-linke,
any.pretty installed,
any.template-haskell installed
+package *
+ happy-options: --strict
benchmarks: False
tests: False
=====================================
compiler/Setup.hs
=====================================
@@ -1,4 +1,5 @@
{-# LANGUAGE NamedFieldPuns #-}
+{-# LANGUAGE CPP #-}
module Main where
import Distribution.Simple
@@ -12,6 +13,8 @@ import Distribution.Simple.Program
import Distribution.Simple.Utils
import Distribution.Simple.Setup
import Distribution.Simple.PackageIndex
+import qualified Distribution.Simple.LocalBuildInfo as LBI
+
import System.IO
import System.Process
@@ -59,8 +62,9 @@ primopIncls =
ghcAutogen :: Verbosity -> LocalBuildInfo -> IO ()
ghcAutogen verbosity lbi@LocalBuildInfo{pkgDescrFile,withPrograms,componentNameMap,installedPkgs}
= do
+ let i = LBI.interpretSymbolicPathLBI lbi
-- Get compiler/ root directory from the cabal file
- let Just compilerRoot = takeDirectory <$> pkgDescrFile
+ let Just compilerRoot = takeDirectory . i <$> pkgDescrFile
-- Require the necessary programs
(gcc ,withPrograms) <- requireProgram normal gccProgram withPrograms
@@ -80,15 +84,19 @@ ghcAutogen verbosity lbi@LocalBuildInfo{pkgDescrFile,withPrograms,componentNameM
-- Call genprimopcode to generate *.hs-incl
forM_ primopIncls $ \(file,command) -> do
contents <- readProcess "genprimopcode" [command] primopsStr
- rewriteFileEx verbosity (buildDir lbi </> file) contents
+ rewriteFileEx verbosity (i (buildDir lbi) </> file) contents
-- Write GHC.Platform.Constants
- let platformConstantsPath = autogenPackageModulesDir lbi </> "GHC/Platform/Constants.hs"
+ let platformConstantsPath = i (autogenPackageModulesDir lbi) </> "GHC/Platform/Constants.hs"
targetOS = case lookup "target os" settings of
Nothing -> error "no target os in settings"
Just os -> os
createDirectoryIfMissingVerbose verbosity True (takeDirectory platformConstantsPath)
+#if MIN_VERSION_Cabal(3,14,0)
+ withTempFile "Constants_tmp.hs" $ \tmp h -> do
+#else
withTempFile (takeDirectory platformConstantsPath) "Constants_tmp.hs" $ \tmp h -> do
+#endif
hClose h
callProcess "deriveConstants" ["--gen-haskell-type","-o",tmp,"--target-os",targetOS]
renameFile tmp platformConstantsPath
@@ -103,7 +111,7 @@ ghcAutogen verbosity lbi@LocalBuildInfo{pkgDescrFile,withPrograms,componentNameM
_ -> error "Couldn't find unique ghc-internal library when building ghc"
-- Write GHC.Settings.Config
- configHsPath = autogenPackageModulesDir lbi </> "GHC/Settings/Config.hs"
+ configHsPath = i (autogenPackageModulesDir lbi) </> "GHC/Settings/Config.hs"
configHs = generateConfigHs cProjectUnitId cGhcInternalUnitId settings
createDirectoryIfMissingVerbose verbosity True (takeDirectory configHsPath)
rewriteFileEx verbosity configHsPath configHs
=====================================
compiler/ghc.cabal.in
=====================================
@@ -50,7 +50,7 @@ extra-source-files:
custom-setup
- setup-depends: base >= 3 && < 5, Cabal >= 1.6 && <3.14, directory, process, filepath, containers
+ setup-depends: base >= 3 && < 5, Cabal >= 1.6 && <3.16, directory, process, filepath, containers
Flag internal-interpreter
Description: Build with internal interpreter support.
=====================================
libraries/ghc-boot/Setup.hs
=====================================
@@ -10,6 +10,7 @@ import Distribution.Verbosity
import Distribution.Simple.Program
import Distribution.Simple.Utils
import Distribution.Simple.Setup
+import qualified Distribution.Simple.LocalBuildInfo as LBI
import System.IO
import System.Directory
@@ -32,12 +33,13 @@ main = defaultMainWithHooks ghcHooks
ghcAutogen :: Verbosity -> LocalBuildInfo -> IO ()
ghcAutogen verbosity lbi@LocalBuildInfo{..} = do
-- Get compiler/ root directory from the cabal file
- let Just compilerRoot = takeDirectory <$> pkgDescrFile
+ let Just compilerRoot = takeDirectory . i <$> pkgDescrFile
- let platformHostFile = "GHC/Platform/Host.hs"
- platformHostPath = autogenPackageModulesDir lbi </> platformHostFile
+ i = LBI.interpretSymbolicPathLBI lbi
+ platformHostFile = "GHC/Platform/Host.hs"
+ platformHostPath = i (autogenPackageModulesDir lbi) </> platformHostFile
ghcVersionFile = "GHC/Version.hs"
- ghcVersionPath = autogenPackageModulesDir lbi </> ghcVersionFile
+ ghcVersionPath = i (autogenPackageModulesDir lbi) </> ghcVersionFile
-- Get compiler settings
settings <- lookupEnv "HADRIAN_SETTINGS" >>= \case
=====================================
libraries/ghc-boot/ghc-boot.cabal.in
=====================================
@@ -28,7 +28,7 @@ build-type: Custom
extra-source-files: changelog.md
custom-setup
- setup-depends: base >= 3 && < 5, Cabal >= 1.6 && <3.14, directory, filepath
+ setup-depends: base >= 3 && < 5, Cabal >= 1.6 && <3.16, directory, filepath
source-repository head
type: git
=====================================
utils/genprimopcode/genprimopcode.cabal
=====================================
@@ -32,4 +32,4 @@ Executable genprimopcode
Build-Depends: base >= 4 && < 5,
array
if flag(build-tool-depends)
- build-tool-depends: alex:alex >= 3.2.6, happy:happy >= 1.20.0
+ build-tool-depends: alex:alex >= 3.2.6, happy:happy >= 2.1.5 || == 1.20.0 || == 1.20.1.1
View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/6a683716d04ee2668a2572ba6235bd…
--
View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/6a683716d04ee2668a2572ba6235bd…
You're receiving this email because of your account on gitlab.haskell.org.
1
0
[Git][ghc/ghc][wip/romes/26202] 74 commits: T22033 is only relevant if the word size is 64-bit
by Rodrigo Mesquita (＠alt-romes) 04 Nov '25
by Rodrigo Mesquita (＠alt-romes) 04 Nov '25
04 Nov '25
Rodrigo Mesquita pushed to branch wip/romes/26202 at Glasgow Haskell Compiler / GHC
Commits:
f7adfed2 by ARATA Mizuki at 2025-10-08T08:37:24-04:00
T22033 is only relevant if the word size is 64-bit
Fixes #25497
- - - - -
ff1650c9 by Ben Gamari at 2025-10-08T08:38:07-04:00
rts/posix: Enforce iteration limit on heap reservation logic
Previously we could loop indefinitely when attempting to get an address
space reservation for our heap. Limit the logic to 8 iterations to
ensure we instead issue a reasonable error message.
Addresses #26151.
- - - - -
01844557 by Ben Gamari at 2025-10-08T08:38:07-04:00
rts/posix: Hold on to low reservations when reserving heap
Previously when the OS gave us an address space reservation in low
memory we would immediately release it and try again. However, on some
platforms this meant that we would get the same allocation again in the
next iteration (since mmap's `hint` argument is just that, a hint).
Instead we now hold on to low reservations until we have found a
suitable heap reservation.
Fixes #26151.
- - - - -
b2c8d052 by Sven Tennie at 2025-10-08T08:38:47-04:00
Build terminfo only in upper stages in cross-builds (#26288)
Currently, there's no way to provide library paths for [n]curses for
both - build and target - in cross-builds. As stage0 is only used to
build upper stages, it should be fine to build terminfo only for them.
This re-enables building cross-compilers with terminfo.
- - - - -
c58f9a61 by Julian Ospald at 2025-10-08T08:39:36-04:00
ghc-toolchain: Drop `ld.gold` from merge object command
It's deprecated.
Also see #25716
- - - - -
2b8baada by sheaf at 2025-10-08T18:23:37-04:00
Improvements to 'mayLookIdentical'
This commit makes significant improvements to the machinery that decides
when we should pretty-print the "invisible bits" of a type, such as:
- kind applications, e.g. '@k' in 'Proxy @k ty'
- RuntimeReps, e.g. 'TYPE r'
- multiplicities and linear arrows 'a %1 -> b'
To do this, this commit refactors 'mayLookIdentical' to return **which**
of the invisible bits don't match up, e.g. in
(a %1 -> b) ~ (a %Many -> b)
we find that the invisible bit that doesn't match up is a multiplicity,
so we should set 'sdocLinearTypes = True' when pretty-printing, and with
e.g.
Proxy @k1 ~ Proxy @k2
we find that the invisible bit that doesn't match up is an invisible
TyCon argument, so we set 'sdocPrintExplicitKinds = True'.
We leverage these changes to remove the ad-hoc treatment of linearity
of data constructors with 'dataConDisplayType' and 'dataConNonLinearType'.
This is now handled by the machinery of 'pprWithInvisibleBits'.
Fixes #26335 #26340
- - - - -
129ce32d by sheaf at 2025-10-08T18:23:37-04:00
Store SDoc context in SourceError
This commits modifies the SourceError datatype which is used for
throwing and then reporting exceptions by storing all the info we need
to be able to print the SDoc, including whether we should print with
explicit kinds, explicit runtime-reps, etc.
This is done using the new datatype:
data SourceErrorContext
= SEC
!DiagOpts
!(DiagnosticOpts GhcMessage)
Now, when we come to report an error (by handling the exception), we
have access to the full context we need.
Fixes #26387
- - - - -
f9790ca8 by Ben Gamari at 2025-10-08T18:24:19-04:00
gitlab-ci: Make RELEASE_JOB an input
Rather than an undocumented variable.
- - - - -
14281a22 by Ben Gamari at 2025-10-11T14:06:47-04:00
rts/nonmoving: Fix comment spelling
- - - - -
bedd38b0 by Ben Gamari at 2025-10-11T14:06:47-04:00
rts/nonmoving: Use atomic operations to update bd->flags
- - - - -
215d6841 by Ben Gamari at 2025-10-11T14:06:47-04:00
nonmoving: Use get_itbl instead of explicit loads
This is cleaner and also fixes unnecessary (and unsound) use of
`volatile`.
- - - - -
2c94aa3a by Ben Gamari at 2025-10-11T14:06:47-04:00
rts/Scav: Handle WHITEHOLEs in scavenge_one
`scavenge_one`, used to scavenge mutable list entries, may encounter
`WHITEHOLE`s when the non-moving GC is in use via two paths:
1. when an MVAR is being marked concurrently
2. when the object belongs to a chain of selectors being short-cutted.
Fixes #26204.
- - - - -
6bd8155c by Matthew Pickering at 2025-10-11T14:07:29-04:00
Add support for generating bytecode objects
This commit adds the `-fwrite-byte-code` option which makes GHC emit a
`.gbc` file which contains a serialised representation of bytecode.
The bytecode can be loaded by the compiler to avoid having to
reinterpret a module when using the bytecode interpreter (for example,
in GHCi).
There are also the new options:
* -gbcdir=<DIR>: Specify the directory to place the gbc files
* -gbcsuf=<suffix>: Specify the suffix for gbc files
The option `-fbyte-code-and-object-code` now implies
`-fwrite-byte-code`.
These performance tests fail due to https://github.com/haskell/directory/issues/204
-------------------------
Metric Increase:
MultiComponentModules
MultiLayerModules
MultiComponentModulesRecomp
MultiLayerModulesRecomp
MultiLayerModulesTH_Make
MultiLayerModulesTH_OneShot
T13701
-------------------------
The bytecode serialisation part was implemented by Cheng Shao
Co-authored-by: Cheng Shao <terrorjack(a)type.dance>
- - - - -
dc8f9599 by Matthew Pickering at 2025-10-11T14:07:30-04:00
Revert "Add a perf test for #26425"
This test has a large memory spike currently, which makes the test
sensitive, since if you allocate a little more or less, the precise
location where GC happens shifts and you observe a different part of the
spike.
Andreas told me to revert the patch for now, and he will add it back
when he fixes the memory spike.
This reverts commit 41bdb16fd083110a06507248f648c507a2feb4af.
- - - - -
e10dcd65 by Sven Tennie at 2025-10-12T10:24:56+00:00
T22859: Increase threadDelay for small machines
The previously used thread delay led to failures on my RISC-V test
setups.
- - - - -
d59ef6b6 by Hai / @BestYeen at 2025-10-14T21:51:14-04:00
Change Alex and Happy m4 scripts to display which version was found in the system, adapt small formatting details in Happy script to be more like the Alex script again.
- - - - -
c98abb6a by Hai / @BestYeen at 2025-10-14T21:52:08-04:00
Update occurrences of return to pure and add a sample for redefining :m to mean :main
- - - - -
70ee825a by Cheng Shao at 2025-10-14T21:52:50-04:00
testsuite: fix T3586 for non-SSE3 platforms
`T3586.hs` contains `-fvia-C -optc-msse3` which I think is a
best-effort basis to harvest the C compiler's auto vectorization
optimizations via the C backend back when the test was added. The
`-fvia-C` part is now a deprecated no-op because GHC can't fall back
to the C backend on a non-unregisterised build, and `-optc-msse3`
might actually cause the test to fail on non x86/x64 platforms, e.g.
recent builds of wasi-sdk would report `wasm32-wasi-clang: error:
unsupported option '-msse3' for target 'wasm32-unknown-wasi'`.
So this patch cleans up this historical cruft. `-fvia-C` is removed,
and `-optc-msse3` is only passed when cpuid contains `pni` (which
indicates support of SSE3).
- - - - -
4be32153 by Teo Camarasu at 2025-10-15T08:06:09-04:00
Add submodules for template-haskell-lift and template-haskell-quasiquoter
These two new boot libraries expose stable subsets of the
template-haskell interface.
This is an implemenation of the GHC proposal https://github.com/ghc-proposals/ghc-proposals/pull/696
Work towards #25262
- - - - -
0c00c9c3 by Ben Gamari at 2025-10-15T08:06:51-04:00
rts: Eliminate uses of implicit constant arrays
Folding of `const`-sized variable-length arrays to a constant-length
array is a gnu extension which clang complains about.
Closes #26502.
- - - - -
bf902a1d by Fendor at 2025-10-15T16:00:59-04:00
Refactor distinct constructor tables map construction
Adds `GHC.Types.Unique.FM.alterUFM_L`, `GHC.Types.Unique.DFM.alterUDFM_L`
`GHC.Data.Word64Map.alterLookup` to support fusion of distinct
constructor data insertion and lookup during the construction of the `DataCon`
map in `GHC.Stg.Debug.numberDataCon`.
Co-authored-by: Fendor <fendor(a)posteo.de>
Co-authored-by: Finley McIlwaine <finleymcilwaine(a)gmail.com>
- - - - -
b3585ba1 by Fendor at 2025-10-15T16:00:59-04:00
Allow per constructor refinement of distinct-constructor-tables
Introduce `-fno-distinct-constructor-tables`. A distinct constructor table
configuration is built from the combination of flags given, in order. For
example, to only generate distinct constructor tables for a few specific
constructors and no others, just pass
`-fdistinct-constructor-tables-only=C1,...,CN`.
This flag can be supplied multiple times to extend the set of
constructors to generate a distinct info table for.
You can disable generation of distinct constructor tables for all
configurations by passing `-fno-distinct-constructor-tables`.
The various configurations of these flags is included in the `DynFlags`
fingerprints, which should result in the expected recompilation logic.
Adds a test that checks for distinct tables for various given or omitted
constructors.
Updates CountDepsAst and CountDepsParser tests to account for new dependencies.
Fixes #23703
Co-authored-by: Fendor <fendor(a)posteo.de>
Co-authored-by: Finley McIlwaine <finleymcilwaine(a)gmail.com>
- - - - -
e17dc695 by fendor at 2025-10-15T16:01:41-04:00
Fix typos in haddock documentation for stack annotation API
- - - - -
f85058d3 by Zubin Duggal at 2025-10-17T13:50:52+05:30
compiler: Attempt to systematize Unique tags by introducing an ADT for each different tag
Fixes #26264
Metric Decrease:
T9233
- - - - -
c85c845d by sheaf at 2025-10-17T22:35:32-04:00
Don't prematurely final-zonk PatSyn declarations
This commit makes GHC hold off on the final zonk for pattern synonym
declarations, in 'GHC.Tc.TyCl.PatSyn.tc_patsyn_finish'.
This accommodates the fact that pattern synonym declarations without a
type signature can contain unfilled metavariables, e.g. if the RHS of
the pattern synonym involves view-patterns whose type mentions promoted
(level 0) metavariables. Just like we do for ordinary function bindings,
we should allow these metavariables to be settled later, instead of
eagerly performing a final zonk-to-type.
Now, the final zonking-to-type for pattern synonyms is performed in
GHC.Tc.Module.zonkTcGblEnv.
Fixes #26465
- - - - -
ba3e5bdd by Rodrigo Mesquita at 2025-10-18T16:57:18-04:00
Move code-gen aux symbols from ghc-internal to rts
These symbols were all previously defined in ghc-internal and made the
dependency structure awkward, where the rts may refer to some of these
symbols and had to work around that circular dependency the way
described in #26166.
Moreover, the code generator will produce code that uses these symbols!
Therefore, they should be available in the rts:
PRINCIPLE: If the code generator may produce code which uses this
symbol, then it should be defined in the rts rather than, say,
ghc-internal.
That said, the main motivation is towards fixing #26166.
Towards #26166. Pre-requisite of !14892
- - - - -
f31de2a9 by Ben Gamari at 2025-10-18T16:57:18-04:00
rts: Avoid static symbol references to ghc-internal
This resolves #26166, a bug due to new constraints placed by Apple's
linker on undefined references.
One source of such references in the RTS is the many symbols referenced
in ghc-internal. To mitigate #26166, we make these references dynamic,
as described in Note [RTS/ghc-internal interface].
Fixes #26166
Co-authored-by: Rodrigo Mesquita <rodrigo.m.mesquita(a)gmail.com>
Co-authored-by: Cheng Shao <terrorjack(a)type.dance>
- - - - -
43fdfddc by Ben Gamari at 2025-10-18T16:57:18-04:00
compiler: Rename isMathFun -> isLibcFun
This set includes more than just math functions.
- - - - -
4ed5138f by Ben Gamari at 2025-10-18T16:57:18-04:00
compiler: Add libc allocator functions to libc_funs
Prototypes for these are now visible from `Prim.h`, resulting in
multiple-declaration warnings in the unregisterised job.
- - - - -
9a0a076b by Ben Gamari at 2025-10-18T16:57:18-04:00
rts: Minimize header dependencies of Prim.h
Otherwise we will end up with redundant and incompatible declarations
resulting in warnings during the unregisterised build.
- - - - -
26b8a414 by Diego Antonio Rosario Palomino at 2025-10-18T16:58:10-04:00
Cmm Parser: Fix incorrect example in comment
The Parser.y file contains a comment with an incorrect example of textual
Cmm (used in .cmm files). This commit updates the comment to ensure it
reflects valid textual Cmm syntax.
Fixes #26313
- - - - -
d4a9d6d6 by ARATA Mizuki at 2025-10-19T18:43:47+09:00
Handle implications between x86 feature flags
This includes:
* Multiple -msse* options can be specified
* -mavx implies -msse4.2
* -mavx2 implies -mavx
* -mfma implies -mavx
* -mavx512f implies -mavx2 and -mfma
* -mavx512{cd,er,pf} imply -mavx512f
Closes #24989
Co-authored-by: sheaf <sam.derbyshire(a)gmail.com>
- - - - -
c9b8465c by Cheng Shao at 2025-10-20T10:16:00-04:00
wasm: workaround WebKit bug in dyld
This patch works around a WebKit bug and allows dyld to run on WebKit
based platforms as well. See added note for detailed explanation.
Co-authored-by: Codex <codex(a)openai.com>
- - - - -
91b6be10 by Julian Ospald at 2025-10-20T18:21:03-04:00
Improve error handling in 'getPackageArchives'
When the library dirs in the package conf files are not set up correctly,
the JS linker will happily ignore such packages and not link against them,
although they're part of the link plan.
Fixes #26383
- - - - -
6c5269da by Sven Tennie at 2025-10-20T18:21:44-04:00
Align coding style
Improve readability by using the same style for all constructor calls in
this function.
- - - - -
3d305889 by Sven Tennie at 2025-10-20T18:21:44-04:00
Reduce complexity by removing joins with mempty
ldArgs, cArgs and cppArgs are all `mempty`. Thus concatenating them adds
nothing but some complexity while reading the code.
- - - - -
38d65187 by Matthew Pickering at 2025-10-21T13:12:20+01:00
Fix stack decoding when using profiled runtime
There are three fixes in this commit.
* We need to replicate the `InfoTable` and `InfoTableProf`
approach for the other stack constants (see the new Stack.ConstantsProf
file).
* Then we need to appropiately import the profiled or non-profiled
versions.
* Finally, there was an incorrect addition in `stackFrameSize`. We need
to cast after performing addition on words.
Fixes #26507
- - - - -
17231bfb by fendor at 2025-10-21T13:12:20+01:00
Add regression test for #26507
- - - - -
4f5bf93b by Simon Peyton Jones at 2025-10-25T04:05:34-04:00
Postscript to fix for #26255
This MR has comments only
- - - - -
6ef22fa0 by IC Rainbow at 2025-10-26T18:23:01-04:00
Add SIMD primops for bitwise logical operations
This adds 128-bit wide and/or/xor instructions for X86 NCG,
with both SSE and AVX encodings.
```
andFloatX4# :: FloatX4# -> FloatX4# -> FloatX4# -- andps / vandps
andDoubleX2# :: DoubleX2# -> DoubleX2# -> DoubleX2# -- andpd / vandpd
andInt8X16# :: Int8X16# -> Int8X16# -> Int8X16# -- pand / vpand
```
The new primops are available on ARM when using LLVM backend.
Tests added:
- simd015 (floats and doubles)
- simd016 (integers)
- simd017 (words)
Fixes #26417
- - - - -
fbdc623a by sheaf at 2025-10-26T18:23:52-04:00
Add hints for unsolved HasField constraints
This commit adds hints and explanations for unsolved 'HasField'
constraints.
GHC will now provide additional explanations for an unsolved constraint
of the form 'HasField fld_name rec_ty fld_ty'; the details are laid out in
Note [Error messages for unsolved HasField constraints], but briefly:
1. Provide similar name suggestions (e.g. mis-spelled field name)
and import suggestions (record field not in scope).
These result in actionable 'GhcHints', which is helpful to provide
code actions in HLS.
2. Explain why GHC did not solve the constraint, e.g.:
- 'fld_name' is not a string literal (e.g. a type variable)
- 'rec_ty' is a TyCon without any fields, e.g. 'Int' or 'Bool'.
- 'fld_ty' contains existentials variables or foralls.
- The record field is a pattern synonym field (GHC does not generate
HasField instances for those).
- 'HasField' is a custom 'TyCon', not actually the built-in
'HasField' typeclass from 'GHC.Records'.
On the way, we slightly refactor the mechanisms for import suggestions
in GHC.Rename.Unbound. This is to account for the fact that, for
'HasField', we don't care whether the field is imported qualified or
unqualified. 'importSuggestions' was refactored, we now have
'sameQualImportSuggestions' and 'anyQualImportSuggestions'.
Fixes #18776 #22382 #26480
- - - - -
99d5707f by sheaf at 2025-10-26T18:23:52-04:00
Rename PatSyn MatchContext to PatSynCtx to avoid punning
- - - - -
5dc2e9ea by Julian Ospald at 2025-10-27T18:17:23-04:00
Skip uniques test if sources are not available
- - - - -
544b9ec9 by Vladislav Zavialov at 2025-10-27T18:18:06-04:00
Re-export GHC.Hs.Basic from GHC.Hs
Clean up some import sections in GHC by re-exporting GHC.Hs.Basic
from GHC.Hs.
- - - - -
643ce801 by Julian Ospald at 2025-10-28T18:18:55-04:00
rts: remove unneccesary cabal flags
We perform those checks via proper autoconf macros
instead that do the right thing and then add those
libs to the rts buildinfo.
- - - - -
d69ea8fe by Vladislav Zavialov at 2025-10-28T18:19:37-04:00
Test case for #17705
Starting with GHC 9.12 (the first release to include 5745dbd3),
all examples in this ticket are handled as expected.
- - - - -
4038a28b by Andreas Klebinger at 2025-10-30T12:38:52-04:00
Add a perf test for #26425
- - - - -
f997618e by Andreas Klebinger at 2025-10-30T12:38:52-04:00
OccAnal: Be stricter for better compiler perf.
In particular we are now stricter:
* When combining usageDetails.
* When computing binder info.
In combineUsageDetails when combining the underlying adds we compute a
new `LocalOcc` for each entry by combining the two existing ones.
Rather than wait for those entries to be forced down the road we now
force them immediately. Speeding up T26425 by about 10% with little
effect on the common case.
We also force binders we put into the Core AST everywhere now.
Failure to do so risks leaking the occ env used to set the binders
OccInfo.
For T26425 compiler residency went down by a factor of ~10x.
Compile time also improved by a factor of ~1.6.
-------------------------
Metric Decrease:
T18698a
T26425
T9233
-------------------------
- - - - -
5618645b by Vladislav Zavialov at 2025-10-30T12:39:33-04:00
Fix namespace specifiers in subordinate exports (#12488)
This patch fixes an oversight in the `lookupChildrenExport` function that
caused explicit namespace specifiers of subordinate export items to be
ignored:
module M (T (type A)) where -- should be rejected
data T = A
Based on the `IEWrappedName` data type, there are 5 cases to consider:
1. Unadorned name: P(X)
2. Named default: P(default X)
3. Pattern synonym: P(pattern X)
4. Type name: P(type X)
5. Data name: P(data X)
Case 1 is already handled correctly; cases 2 and 3 are parse errors; and
it is cases 4 and 5 that we are concerned with in this patch.
Following the precedent established in `LookupExactName`, we introduce
a boolean flag in `LookupChildren` to control whether to look up in all
namespaces or in a specific one. If an export item is accompanied by an
explicit namespace specifier `type` or `data`, we restrict the lookup in
`lookupGRE` to a specific namespace.
The newly introduced diagnostic `TcRnExportedSubordinateNotFound`
provides error messages and suggestions more tailored to this context
than the previously used `reportUnboundName`.
- - - - -
f75ab223 by Peter Trommler at 2025-10-31T18:43:13-04:00
ghc-toolchain: detect PowerPC 64 bit ABI
Check preprocessor macro defined for ABI v2 and assume v1 otherwise.
Fixes #26521
- - - - -
d086c474 by Peter Trommler at 2025-10-31T18:43:13-04:00
ghc-toolchain: refactor, move lastLine to Utils
- - - - -
995dfe0d by Vladislav Zavialov at 2025-10-31T18:43:54-04:00
Tests for -Wduplicate-exports, -Wdodgy-exports
Add test cases for the previously untested diagnostics:
[GHC-51876] TcRnDupeModuleExport
[GHC-64649] TcRnNullExportedModule
This also revealed a typo (incorrect capitalization of "module") in the
warning text for TcRnDupeModuleExport, which is now fixed.
- - - - -
f6961b02 by Cheng Shao at 2025-11-01T00:08:01+01:00
wasm: reformat dyld source code
This commit reformats dyld source code with prettier, to avoid
introducing unnecessary diffs in subsequent patches when they're
formatted before committing.
- - - - -
0c9032a0 by Cheng Shao at 2025-11-01T00:08:01+01:00
wasm: simplify _initialize logic in dyld
This commit simplifies how we _initialize a wasm shared library in
dyld and removes special treatment for libc.so, see added comment for
detailed explanation.
- - - - -
ec1b40bd by Cheng Shao at 2025-11-01T00:08:01+01:00
wasm: support running dyld fully client side in the browser
This commit refactors the wasm dyld script so that it can be used to
load and run wasm shared libraries fully client-side in the browser
without needing a wasm32-wasi-ghci backend:
- A new `DyLDBrowserHost` class is exported, which runs in the browser
and uses the in-memory vfs without any RPC calls. This meant to be
used to create a `rpc` object for the fully client side use cases.
- The exported `main` function now can be used to load user-specified
shared libraries, and the user can use the returned `DyLD` instance
to run their own exported Haskell functions.
- The in-browser wasi implementation is switched to
https://github.com/haskell-wasm/browser_wasi_shim for bugfixes and
major performance improvements not landed upstream yet.
- When being run by deno, it now correctly switches to non-nodejs code
paths, so it's more convenient to test dyld logic with deno.
See added comments for details, as well as the added `playground001`
test case for an example of using it to build an in-browser Haskell
playground.
- - - - -
8f3e481f by Cheng Shao at 2025-11-01T00:08:01+01:00
testsuite: add playground001 to test haskell playground
This commit adds the playground001 test case to test the haskell
playground in browser, see comments for details.
- - - - -
af40606a by Cheng Shao at 2025-11-01T00:08:04+01:00
Revert "testsuite: add T26431 test case"
This reverts commit 695036686f8c6d78611edf3ed627608d94def6b7. T26431
is now retired, wasm ghc internal-interpreter logic is tested by
playground001.
- - - - -
86c82745 by Vladislav Zavialov at 2025-11-01T07:24:29-04:00
Supplant TcRnExportHiddenComponents with TcRnDodgyExports (#26534)
Remove a bogus special case in lookup_ie_kids_all,
making TcRnExportHiddenComponents obsolete.
- - - - -
fcf6331e by Richard Eisenberg at 2025-11-03T08:33:05+00:00
Refactor fundep solving
This commit is a large-scale refactor of the increasingly-messy code that
handles functional dependencies. It has virtually no effect on what compiles
but improves error messages a bit. And it does the groundwork for #23162.
The big picture is described in
Note [Overview of functional dependencies in type inference]
in GHC.Tc.Solver.FunDeps
* New module GHC.Tc.Solver.FunDeps contains all the fundep-handling
code for the constraint solver.
* Fundep-equalities are solved in a nested scope; they may generate
unifications but otherwise have no other effect.
See GHC.Tc.Solver.FunDeps.solveFunDeps
The nested needs to start from the Givens in the inert set, but
not the Wanteds; hence a new function `resetInertCans`, used in
`nestFunDepsTcS`.
* That in turn means that fundep equalities never show up in error
messages, so the complicated FunDepOrigin tracking can all disappear.
* We need to be careful about tracking unifications, so we kick out
constraints from the inert set after doing unifications. Unification
tracking has been majorly reformed: see Note [WhatUnifications] in
GHC.Tc.Utils.Unify.
A good consequence is that the hard-to-grok `resetUnificationFlag`
has been replaced with a simpler use of
`reportCoarseGrainUnifications`
Smaller things:
* Rename `FunDepEqn` to `FunDepEqns` since it contains multiple
type equalities.
Some compile time improvement
Metrics: compile_time/bytes allocated
Baseline
Test value New value Change
---------------------- --------------------------------------
T5030(normal) 173,839,232 148,115,248 -14.8% GOOD
hard_hole_fits(normal) 286,768,048 284,015,416 -1.0%
geo. mean -0.2%
minimum -14.8%
maximum +0.3%
Metric Decrease:
T5030
- - - - -
231adc30 by Simon Peyton Jones at 2025-11-03T08:33:05+00:00
QuickLook's tcInstFun should make instantiation variables directly
tcInstFun must make "instantiation variables", not regular
unification variables, when instantiating function types. That was
previously implemented by a hack: set the /ambient/ level to QLInstTyVar.
But the hack finally bit me, when I was refactoring WhatUnifications.
And it was always wrong: see the now-expunged (TCAPP2) note.
This commit does it right, by making tcInstFun call its own
instantiation functions. That entails a small bit of duplication,
but the result is much, much cleaner.
- - - - -
39d4a24b by Simon Peyton Jones at 2025-11-03T08:33:05+00:00
Build implication for constraints from (static e)
This commit addresses #26466, by buiding an implication for the
constraints arising from a (static e) form. The implication has
a special ic_info field of StaticFormSkol, which tells the constraint
solver to use an empty set of Givens.
See (SF3) in Note [Grand plan for static forms]
in GHC.Iface.Tidy.StaticPtrTable
This commit also reinstates an `assert` in GHC.Tc.Solver.Equality.
The test `StaticPtrTypeFamily` was failing with an assertion failure,
but it now works.
- - - - -
2e2aec1e by Simon Peyton Jones at 2025-11-03T08:33:05+00:00
Comments about defaulting representation equalities
- - - - -
52a4d1da by Simon Peyton Jones at 2025-11-03T08:33:05+00:00
Improve tracking of rewriter-sets
This refactor substantially improves the treatment of so-called
"rewriter-sets" in the constraint solver.
The story is described in the rewritten
Note [Wanteds rewrite Wanteds: rewriter-sets]
in GHC.Tc.Types.Constraint
Some highlights
* Trace the free coercion holes of a filled CoercionHole,
in CoercionPlusHoles. See Note [Coercion holes] (COH5)
This avoids taking having to take the free coercion variables
of a coercion when zonking a rewrriter-set
* Many knock on changes
* Make fillCoercionHole take CoercionPlusHoles as its argument
rather than to separate arguments.
* Similarly setEqIfWanted, setWantedE, wrapUnifierAndEmit.
* Be more careful about passing the correct CoHoleSet to
`rewriteEqEvidence` and friends
* Make kickOurAfterFillingCoercionHole more clever. See
new Note [Kick out after filling a coercion hole]
Smaller matters
* Rename RewriterSet to CoHoleSet
* Add special-case helper `rewriteEqEvidenceSwapOnly`
- - - - -
3e78e1ba by Simon Peyton Jones at 2025-11-03T08:33:05+00:00
Tidy up constraint solving for foralls
* In `can_eq_nc_forall` make sure to track Givens that are used
in the nested solve step.
* Tiny missing-swap bug-fix in `lookup_eq_in_qcis`
* Fix some leftover mess from
commit 14123ee646f2b9738a917b7cec30f9d3941c13de
Author: Simon Peyton Jones <simon.peytonjones(a)gmail.com>
Date: Wed Aug 20 00:35:48 2025 +0100
Solve forall-constraints via an implication, again
Specifically, trySolveImplication is now dead.
- - - - -
973f2c25 by Simon Peyton Jones at 2025-11-03T08:33:05+00:00
Do not treat CoercionHoles as free variables in coercions
This fixes a long-standing wart in the free-variable finder;
now CoercionHoles are no longer treated as a "free variable"
of a coercion.
I got big and unexpected performance regressions when making
this change. Turned out that CallArity didn't discover that
the free variable finder could be eta-expanded, which gave very
poor code.
So I re-used Note [The one-shot state monad trick] for Endo,
resulting in GHC.Utils.EndoOS. Very simple, big win.
- - - - -
c2b8a0f9 by Simon Peyton Jones at 2025-11-03T08:33:05+00:00
Update debug-tracing in CallArity
No effect on behaviour, and commented out anyway
- - - - -
9aa5ee99 by Simon Peyton Jones at 2025-11-03T08:33:28+00:00
Comments only -- remove dangling Note references
- - - - -
6683f183 by Simon Peyton Jones at 2025-11-03T08:33:28+00:00
Accept error message wibbles
- - - - -
3ba3d9f9 by Luite Stegeman at 2025-11-04T00:59:41-05:00
rts: fix eager black holes: record mutated closure and fix assertion
This fixes two problems with handling eager black holes, introduced
by a1de535f762bc23d4cf23a5b1853591dda12cdc9.
- the closure mutation must be recorded even for eager black holes,
since the mutator has mutated it before calling threadPaused
- The assertion that an unmarked eager black hole must be owned by
the TSO calling threadPaused is incorrect, since multiple threads
can race to claim the black hole.
fixes #26495
- - - - -
e2de3220 by Rodrigo Mesquita at 2025-11-04T14:05:15+00:00
build: Relax ghc/ghc-boot Cabal bound to 3.16
Fixes #26202
- - - - -
b473a9e3 by Rodrigo Mesquita at 2025-11-04T14:05:15+00:00
cabal-reinstall: Use haddock-api +in-tree-ghc
Fixes #26202
- - - - -
12e1f76d by Rodrigo Mesquita at 2025-11-04T14:05:15+00:00
cabal-reinstall: Pass --strict to Happy
This is necessary to make the generated Parser build successfully
This mimics Hadrian, which always passes --strict to happy.
Fixes #26202
- - - - -
3694ba79 by Rodrigo Mesquita at 2025-11-04T14:05:15+00:00
genprimopcode: Require higher happy version
I've bumped the happy version to forbid deprecated Happy versions which
don't successfully compile.
- - - - -
6a683716 by Rodrigo Mesquita at 2025-11-04T14:05:15+00:00
Bump CI index-state
Necessary to fix cabal.project-reinstall
- - - - -
521 changed files:
- .gitlab-ci.yml
- .gitlab/ci.sh
- .gitlab/generate-ci/gen_ci.hs
- .gitlab/jobs.yaml
- .gitmodules
- cabal.project-reinstall
- compiler/GHC.hs
- compiler/GHC/Builtin/PrimOps.hs
- compiler/GHC/Builtin/Uniques.hs
- compiler/GHC/Builtin/primops.txt.pp
- compiler/GHC/ByteCode/Breakpoints.hs
- + compiler/GHC/ByteCode/Serialize.hs
- compiler/GHC/ByteCode/Types.hs
- compiler/GHC/Cmm/CLabel.hs
- compiler/GHC/Cmm/Info.hs
- compiler/GHC/Cmm/Info/Build.hs
- compiler/GHC/Cmm/MachOp.hs
- compiler/GHC/Cmm/Parser.y
- compiler/GHC/Cmm/Pipeline.hs
- compiler/GHC/Cmm/UniqueRenamer.hs
- compiler/GHC/CmmToAsm/AArch64/CodeGen.hs
- compiler/GHC/CmmToAsm/Config.hs
- compiler/GHC/CmmToAsm/Wasm/FromCmm.hs
- compiler/GHC/CmmToAsm/X86/CodeGen.hs
- compiler/GHC/CmmToAsm/X86/Instr.hs
- compiler/GHC/CmmToAsm/X86/Ppr.hs
- compiler/GHC/CmmToC.hs
- compiler/GHC/CmmToLlvm/Base.hs
- compiler/GHC/CmmToLlvm/CodeGen.hs
- compiler/GHC/Core/ConLike.hs
- compiler/GHC/Core/Make.hs
- compiler/GHC/Core/Multiplicity.hs
- compiler/GHC/Core/Opt/CallArity.hs
- compiler/GHC/Core/Opt/Monad.hs
- compiler/GHC/Core/Opt/OccurAnal.hs
- compiler/GHC/Core/Opt/Pipeline.hs
- compiler/GHC/Core/Opt/Simplify/Iteration.hs
- compiler/GHC/Core/Opt/Simplify/Monad.hs
- compiler/GHC/Core/Opt/Simplify/Utils.hs
- compiler/GHC/Core/PatSyn.hs
- compiler/GHC/Core/TyCo/Compare.hs
- compiler/GHC/Core/TyCo/FVs.hs
- compiler/GHC/Core/TyCo/Ppr.hs
- compiler/GHC/Core/TyCo/Rep.hs
- compiler/GHC/Core/TyCon.hs
- compiler/GHC/CoreToStg/Prep.hs
- compiler/GHC/Data/FlatBag.hs
- compiler/GHC/Data/SmallArray.hs
- compiler/GHC/Data/Word64Map/Internal.hs
- compiler/GHC/Data/Word64Map/Lazy.hs
- compiler/GHC/Driver/Backend.hs
- compiler/GHC/Driver/Backpack.hs
- + compiler/GHC/Driver/ByteCode.hs
- compiler/GHC/Driver/CodeOutput.hs
- compiler/GHC/Driver/Config/CmmToAsm.hs
- compiler/GHC/Driver/Config/Finder.hs
- compiler/GHC/Driver/Config/Stg/Debug.hs
- compiler/GHC/Driver/Downsweep.hs
- compiler/GHC/Driver/DynFlags.hs
- compiler/GHC/Driver/Env/Types.hs
- compiler/GHC/Driver/Errors.hs
- compiler/GHC/Driver/Flags.hs
- compiler/GHC/Driver/Main.hs
- compiler/GHC/Driver/Make.hs
- compiler/GHC/Driver/MakeFile.hs
- compiler/GHC/Driver/Messager.hs
- compiler/GHC/Driver/Monad.hs
- compiler/GHC/Driver/Pipeline/Execute.hs
- compiler/GHC/Driver/Session.hs
- compiler/GHC/Hs.hs
- compiler/GHC/Hs/Decls.hs
- compiler/GHC/Hs/Expr.hs
- compiler/GHC/Hs/Type.hs
- compiler/GHC/HsToCore/Breakpoints.hs
- compiler/GHC/HsToCore/Foreign/C.hs
- compiler/GHC/HsToCore/Foreign/JavaScript.hs
- compiler/GHC/HsToCore/Foreign/Wasm.hs
- compiler/GHC/HsToCore/Monad.hs
- compiler/GHC/HsToCore/Pmc/Utils.hs
- compiler/GHC/HsToCore/Quote.hs
- compiler/GHC/HsToCore/Usage.hs
- compiler/GHC/HsToCore/Utils.hs
- compiler/GHC/Iface/Binary.hs
- compiler/GHC/Iface/Decl.hs
- compiler/GHC/Iface/Ext/Ast.hs
- compiler/GHC/Iface/Flags.hs
- compiler/GHC/Iface/Load.hs
- compiler/GHC/Iface/Make.hs
- compiler/GHC/Iface/Recomp/Flags.hs
- compiler/GHC/Iface/Rename.hs
- compiler/GHC/Iface/Tidy/StaticPtrTable.hs
- compiler/GHC/Iface/Type.hs
- compiler/GHC/JS/JStg/Monad.hs
- compiler/GHC/Linker/Deps.hs
- compiler/GHC/Linker/Dynamic.hs
- compiler/GHC/Linker/Loader.hs
- compiler/GHC/Linker/Static.hs
- compiler/GHC/Linker/Types.hs
- compiler/GHC/Parser/Errors/Types.hs
- compiler/GHC/Parser/Header.hs
- compiler/GHC/Parser/PostProcess.hs
- compiler/GHC/Platform.hs
- compiler/GHC/Platform/Reg.hs
- compiler/GHC/Rename/Bind.hs
- compiler/GHC/Rename/Env.hs
- compiler/GHC/Rename/Expr.hs
- compiler/GHC/Rename/HsType.hs
- compiler/GHC/Rename/Names.hs
- compiler/GHC/Rename/Pat.hs
- compiler/GHC/Rename/Unbound.hs
- compiler/GHC/Rename/Utils.hs
- compiler/GHC/Runtime/Eval.hs
- compiler/GHC/Stg/Debug.hs
- + compiler/GHC/Stg/Debug/Types.hs
- compiler/GHC/Stg/EnforceEpt.hs
- compiler/GHC/Stg/Pipeline.hs
- compiler/GHC/StgToCmm/ExtCode.hs
- compiler/GHC/StgToCmm/Monad.hs
- compiler/GHC/StgToCmm/Prim.hs
- compiler/GHC/StgToJS/CodeGen.hs
- compiler/GHC/StgToJS/Ids.hs
- compiler/GHC/StgToJS/Linker/Linker.hs
- compiler/GHC/StgToJS/Prim.hs
- compiler/GHC/Tc/Deriv/Generate.hs
- compiler/GHC/Tc/Deriv/Generics.hs
- compiler/GHC/Tc/Errors.hs
- compiler/GHC/Tc/Errors/Ppr.hs
- compiler/GHC/Tc/Errors/Types.hs
- compiler/GHC/Tc/Gen/App.hs
- compiler/GHC/Tc/Gen/Export.hs
- compiler/GHC/Tc/Gen/Expr.hs
- compiler/GHC/Tc/Gen/HsType.hs
- compiler/GHC/Tc/Gen/Pat.hs
- compiler/GHC/Tc/Gen/Sig.hs
- compiler/GHC/Tc/Gen/Splice.hs
- compiler/GHC/Tc/Instance/Class.hs
- compiler/GHC/Tc/Instance/FunDeps.hs
- compiler/GHC/Tc/Module.hs
- compiler/GHC/Tc/Solver.hs
- compiler/GHC/Tc/Solver/Default.hs
- compiler/GHC/Tc/Solver/Dict.hs
- compiler/GHC/Tc/Solver/Equality.hs
- + compiler/GHC/Tc/Solver/FunDeps.hs
- compiler/GHC/Tc/Solver/InertSet.hs
- compiler/GHC/Tc/Solver/Irred.hs
- compiler/GHC/Tc/Solver/Monad.hs
- compiler/GHC/Tc/Solver/Rewrite.hs
- compiler/GHC/Tc/Solver/Solve.hs
- compiler/GHC/Tc/Solver/Solve.hs-boot
- compiler/GHC/Tc/TyCl.hs
- compiler/GHC/Tc/TyCl/Instance.hs
- compiler/GHC/Tc/TyCl/PatSyn.hs
- compiler/GHC/Tc/TyCl/Utils.hs
- compiler/GHC/Tc/Types.hs
- compiler/GHC/Tc/Types/Constraint.hs
- compiler/GHC/Tc/Types/CtLoc.hs
- compiler/GHC/Tc/Types/Evidence.hs
- compiler/GHC/Tc/Types/Origin.hs
- compiler/GHC/Tc/Utils/Concrete.hs
- compiler/GHC/Tc/Utils/Instantiate.hs
- compiler/GHC/Tc/Utils/Monad.hs
- compiler/GHC/Tc/Utils/TcMType.hs
- compiler/GHC/Tc/Utils/TcType.hs
- compiler/GHC/Tc/Utils/Unify.hs
- compiler/GHC/Tc/Utils/Unify.hs-boot
- compiler/GHC/Tc/Validity.hs
- compiler/GHC/Tc/Zonk/TcType.hs
- compiler/GHC/Tc/Zonk/Type.hs
- compiler/GHC/ThToHs.hs
- compiler/GHC/Types/Basic.hs
- compiler/GHC/Types/Error/Codes.hs
- compiler/GHC/Types/Hint.hs
- compiler/GHC/Types/Hint/Ppr.hs
- compiler/GHC/Types/Name/Cache.hs
- compiler/GHC/Types/Name/Reader.hs
- compiler/GHC/Types/SourceError.hs
- compiler/GHC/Types/SptEntry.hs
- compiler/GHC/Types/Tickish.hs
- compiler/GHC/Types/TyThing/Ppr.hs
- compiler/GHC/Types/Unique.hs
- compiler/GHC/Types/Unique/DFM.hs
- compiler/GHC/Types/Unique/DSM.hs
- compiler/GHC/Types/Unique/FM.hs
- compiler/GHC/Types/Unique/Supply.hs
- compiler/GHC/Types/Var/Env.hs
- compiler/GHC/Unit/Finder.hs
- compiler/GHC/Unit/Finder/Types.hs
- compiler/GHC/Unit/Module/Graph.hs
- compiler/GHC/Unit/Module/Location.hs
- compiler/GHC/Unit/Module/ModSummary.hs
- compiler/GHC/Unit/Module/WholeCoreBindings.hs
- compiler/GHC/Utils/Binary.hs
- + compiler/GHC/Utils/EndoOS.hs
- compiler/GHC/Utils/Error.hs
- compiler/GHC/Utils/Outputable.hs
- compiler/Language/Haskell/Syntax/Binds.hs
- compiler/Language/Haskell/Syntax/Expr.hs
- compiler/Setup.hs
- compiler/ghc.cabal.in
- docs/users_guide/9.16.1-notes.rst
- docs/users_guide/debug-info.rst
- docs/users_guide/ghci.rst
- docs/users_guide/phases.rst
- docs/users_guide/separate_compilation.rst
- docs/users_guide/using.rst
- ghc/GHCi/UI.hs
- hadrian/cabal.project
- hadrian/src/Packages.hs
- hadrian/src/Rules/Gmp.hs
- hadrian/src/Rules/Libffi.hs
- hadrian/src/Settings/Builders/Cabal.hs
- hadrian/src/Settings/Builders/Common.hs
- hadrian/src/Settings/Builders/DeriveConstants.hs
- hadrian/src/Settings/Builders/Hsc2Hs.hs
- hadrian/src/Settings/Default.hs
- hadrian/src/Settings/Packages.hs
- libraries/base/src/GHC/Base.hs
- libraries/base/src/GHC/Exts.hs
- libraries/ghc-boot/Setup.hs
- libraries/ghc-boot/ghc-boot.cabal.in
- libraries/ghc-experimental/CHANGELOG.md
- libraries/ghc-experimental/src/GHC/Stack/Annotation/Experimental.hs
- + libraries/ghc-internal/cbits/RtsIface.c
- libraries/ghc-internal/cbits/Stack_c.c
- libraries/ghc-internal/ghc-internal.cabal.in
- + libraries/ghc-internal/include/RtsIfaceSymbols.h
- + libraries/ghc-internal/src/GHC/Internal/Stack/ConstantsProf.hsc
- libraries/ghc-internal/src/GHC/Internal/Stack/Decode.hs
- libraries/ghc-internal/src/GHC/Internal/TH/Lift.hs
- libraries/ghc-internal/src/GHC/Internal/TH/Quote.hs
- + libraries/ghc-internal/tests/backtraces/T26507.hs
- + libraries/ghc-internal/tests/backtraces/T26507.stderr
- libraries/ghc-internal/tests/backtraces/all.T
- libraries/ghc-internal/tests/stack-annotation/all.T
- libraries/ghc-prim/changelog.md
- + libraries/template-haskell-lift
- + libraries/template-haskell-quasiquoter
- m4/fp_check_pthreads.m4
- m4/fptools_alex.m4
- m4/fptools_happy.m4
- rts/BuiltinClosures.c
- rts/CloneStack.h
- rts/Compact.cmm
- rts/ContinuationOps.cmm
- rts/Exception.cmm
- rts/Prelude.h
- rts/PrimOps.cmm
- rts/Printer.c
- rts/RtsAPI.c
- rts/RtsStartup.c
- rts/RtsSymbols.c
- + rts/RtsToHsIface.c
- rts/StgStdThunks.cmm
- rts/ThreadPaused.c
- rts/configure.ac
- − rts/external-symbols.list.in
- rts/include/Rts.h
- rts/include/RtsAPI.h
- rts/include/Stg.h
- + rts/include/rts/RtsToHsIface.h
- rts/include/rts/Types.h
- rts/include/rts/storage/Block.h
- rts/include/stg/Prim.h
- rts/posix/OSMem.c
- rts/posix/Signals.c
- libraries/ghc-internal/cbits/atomic.c → rts/prim/atomic.c
- libraries/ghc-internal/cbits/bitrev.c → rts/prim/bitrev.c
- libraries/ghc-internal/cbits/bswap.c → rts/prim/bswap.c
- libraries/ghc-internal/cbits/clz.c → rts/prim/clz.c
- libraries/ghc-internal/cbits/ctz.c → rts/prim/ctz.c
- libraries/ghc-internal/cbits/int64x2minmax.c → rts/prim/int64x2minmax.c
- libraries/ghc-internal/cbits/longlong.c → rts/prim/longlong.c
- libraries/ghc-internal/cbits/mulIntMayOflo.c → rts/prim/mulIntMayOflo.c
- libraries/ghc-internal/cbits/pdep.c → rts/prim/pdep.c
- libraries/ghc-internal/cbits/pext.c → rts/prim/pext.c
- libraries/ghc-internal/cbits/popcnt.c → rts/prim/popcnt.c
- libraries/ghc-internal/cbits/vectorQuotRem.c → rts/prim/vectorQuotRem.c
- libraries/ghc-internal/cbits/word2float.c → rts/prim/word2float.c
- rts/rts.buildinfo.in
- rts/rts.cabal
- rts/sm/NonMoving.c
- rts/sm/NonMovingMark.c
- rts/sm/Scav.c
- rts/wasm/JSFFI.c
- rts/wasm/scheduler.cmm
- rts/win32/libHSghc-internal.def
- testsuite/driver/cpu_features.py
- testsuite/tests/bytecode/T24634/T24634a.stdout
- testsuite/tests/bytecode/T24634/T24634b.stdout
- testsuite/tests/codeGen/should_gen_asm/all.T
- + testsuite/tests/codeGen/should_gen_asm/mavx-should-enable-popcnt.asm
- + testsuite/tests/codeGen/should_gen_asm/mavx-should-enable-popcnt.hs
- + testsuite/tests/codeGen/should_gen_asm/msse-option-order.asm
- + testsuite/tests/codeGen/should_gen_asm/msse-option-order.hs
- testsuite/tests/count-deps/CountDepsAst.stdout
- testsuite/tests/count-deps/CountDepsParser.stdout
- testsuite/tests/cpranal/should_compile/T18174.stderr
- testsuite/tests/default/default-fail05.stderr
- testsuite/tests/dependent/should_fail/T13135_simple.stderr
- testsuite/tests/deriving/should_fail/T3621.stderr
- testsuite/tests/diagnostic-codes/codes.stdout
- testsuite/tests/driver/T11429c.stderr
- testsuite/tests/driver/T21682.stderr
- + testsuite/tests/driver/bytecode-object/A.hs
- + testsuite/tests/driver/bytecode-object/BytecodeForeign.c
- + testsuite/tests/driver/bytecode-object/BytecodeForeign.hs
- + testsuite/tests/driver/bytecode-object/BytecodeMain.hs
- + testsuite/tests/driver/bytecode-object/BytecodeTest.hs
- + testsuite/tests/driver/bytecode-object/Makefile
- + testsuite/tests/driver/bytecode-object/all.T
- + testsuite/tests/driver/bytecode-object/bytecode_object12.stderr
- + testsuite/tests/driver/bytecode-object/bytecode_object13.stdout
- + testsuite/tests/driver/bytecode-object/bytecode_object14.stdout
- + testsuite/tests/driver/bytecode-object/bytecode_object15.stdout
- + testsuite/tests/driver/bytecode-object/bytecode_object16.stdout
- + testsuite/tests/driver/bytecode-object/bytecode_object17.stdout
- + testsuite/tests/driver/bytecode-object/bytecode_object18.stdout
- + testsuite/tests/driver/bytecode-object/bytecode_object19.script
- + testsuite/tests/driver/bytecode-object/bytecode_object19.stdout
- + testsuite/tests/driver/bytecode-object/bytecode_object25.stdout
- + testsuite/tests/driver/bytecode-object/bytecode_object4.stdout
- + testsuite/tests/driver/bytecode-object/bytecode_object5.stdout
- + testsuite/tests/driver/bytecode-object/bytecode_object6.stdout
- testsuite/tests/driver/fat-iface/T22405/T22405.stdout
- testsuite/tests/driver/fat-iface/T22405/T22405b.stdout
- testsuite/tests/driver/fat-iface/fat011.stderr
- + testsuite/tests/ghc-api-browser/README.md
- + testsuite/tests/ghc-api-browser/all.T
- + testsuite/tests/ghc-api-browser/index.html
- + testsuite/tests/ghc-api-browser/playground001.hs
- + testsuite/tests/ghc-api-browser/playground001.js
- + testsuite/tests/ghc-api-browser/playground001.sh
- testsuite/tests/ghci-wasm/T26431.stdout → testsuite/tests/ghc-api-browser/playground001.stdout
- testsuite/tests/ghc-api/T10942.hs
- + testsuite/tests/ghc-api/T26264.hs
- + testsuite/tests/ghc-api/T26264.stdout
- testsuite/tests/ghc-api/all.T
- testsuite/tests/ghc-api/annotations-literals/literals.hs
- − testsuite/tests/ghci-wasm/T26431.hs
- testsuite/tests/ghci-wasm/all.T
- testsuite/tests/indexed-types/should_fail/T14369.stderr
- testsuite/tests/indexed-types/should_fail/T14887.stderr
- testsuite/tests/indexed-types/should_fail/T1897b.stderr
- testsuite/tests/interface-stability/ghc-experimental-exports.stdout
- testsuite/tests/interface-stability/ghc-experimental-exports.stdout-mingw32
- testsuite/tests/interface-stability/ghc-prim-exports.stdout
- testsuite/tests/interface-stability/ghc-prim-exports.stdout-mingw32
- testsuite/tests/linear/should_fail/T19361.stderr
- testsuite/tests/linters/all.T
- testsuite/tests/linters/notes.stdout
- testsuite/tests/llvm/should_run/all.T
- testsuite/tests/module/mod4.stderr
- testsuite/tests/overloadedrecflds/should_compile/BootFldReexport.stderr
- testsuite/tests/overloadedrecflds/should_fail/T16745.stderr
- testsuite/tests/overloadedrecflds/should_fail/T18999_NoDisambiguateRecordFields.stderr
- + testsuite/tests/overloadedrecflds/should_fail/T26480.hs
- + testsuite/tests/overloadedrecflds/should_fail/T26480.stderr
- + testsuite/tests/overloadedrecflds/should_fail/T26480_aux1.hs
- + testsuite/tests/overloadedrecflds/should_fail/T26480_aux2.hs
- + testsuite/tests/overloadedrecflds/should_fail/T26480b.hs
- + testsuite/tests/overloadedrecflds/should_fail/T26480b.stderr
- testsuite/tests/overloadedrecflds/should_fail/all.T
- testsuite/tests/overloadedrecflds/should_fail/hasfieldfail01.stderr
- testsuite/tests/overloadedrecflds/should_fail/hasfieldfail02.stderr
- testsuite/tests/parser/should_fail/RecordDotSyntaxFail10.stderr
- testsuite/tests/parser/should_fail/RecordDotSyntaxFail11.stderr
- testsuite/tests/parser/should_fail/RecordDotSyntaxFail13.stderr
- testsuite/tests/parser/should_fail/RecordDotSyntaxFail8.hs
- testsuite/tests/parser/should_fail/RecordDotSyntaxFail8.stderr
- + testsuite/tests/parser/should_fail/T12488c.hs
- + testsuite/tests/parser/should_fail/T12488c.stderr
- + testsuite/tests/parser/should_fail/T12488d.hs
- + testsuite/tests/parser/should_fail/T12488d.stderr
- testsuite/tests/parser/should_fail/T20654a.stderr
- testsuite/tests/parser/should_fail/all.T
- testsuite/tests/partial-sigs/should_fail/T14584a.stderr
- + testsuite/tests/patsyn/should_compile/T26465b.hs
- + testsuite/tests/patsyn/should_compile/T26465c.hs
- + testsuite/tests/patsyn/should_compile/T26465d.hs
- + testsuite/tests/patsyn/should_compile/T26465d.stderr
- testsuite/tests/patsyn/should_compile/all.T
- + testsuite/tests/patsyn/should_fail/T26465.hs
- + testsuite/tests/patsyn/should_fail/T26465.stderr
- testsuite/tests/patsyn/should_fail/all.T
- testsuite/tests/perf/compiler/Makefile
- + testsuite/tests/perf/compiler/MultiLayerModulesDefsGhciWithBytecodeFiles.script
- testsuite/tests/perf/compiler/T26425.hs
- testsuite/tests/perf/compiler/all.T
- testsuite/tests/perf/should_run/T3586.hs
- testsuite/tests/perf/should_run/UniqLoop.hs
- testsuite/tests/perf/should_run/all.T
- testsuite/tests/polykinds/T6068.stdout
- testsuite/tests/quantified-constraints/T15359.hs
- testsuite/tests/regalloc/regalloc_unit_tests.hs
- + testsuite/tests/rename/should_compile/T12488b.hs
- + testsuite/tests/rename/should_compile/T12488f.hs
- testsuite/tests/rename/should_compile/all.T
- + testsuite/tests/rename/should_fail/T12488a.hs
- + testsuite/tests/rename/should_fail/T12488a.stderr
- + testsuite/tests/rename/should_fail/T12488a_foo.hs
- + testsuite/tests/rename/should_fail/T12488a_foo.stderr
- + testsuite/tests/rename/should_fail/T12488e.hs
- + testsuite/tests/rename/should_fail/T12488e.stderr
- + testsuite/tests/rename/should_fail/T12488g.hs
- + testsuite/tests/rename/should_fail/T12488g.stderr
- testsuite/tests/rename/should_fail/T19843h.stderr
- testsuite/tests/rename/should_fail/T25899e2.stderr
- testsuite/tests/rename/should_fail/all.T
- testsuite/tests/rep-poly/RepPolyNPlusK.stderr
- testsuite/tests/rep-poly/RepPolyRightSection.stderr
- testsuite/tests/rep-poly/T13233.stderr
- testsuite/tests/rep-poly/T19709b.stderr
- testsuite/tests/rep-poly/T23903.stderr
- testsuite/tests/roles/should_compile/Roles13.stderr
- testsuite/tests/rts/T22859.hs
- + testsuite/tests/rts/ipe/distinct-tables/Main.hs
- + testsuite/tests/rts/ipe/distinct-tables/Makefile
- + testsuite/tests/rts/ipe/distinct-tables/X.hs
- + testsuite/tests/rts/ipe/distinct-tables/all.T
- + testsuite/tests/rts/ipe/distinct-tables/distinct_tables01.stdout
- + testsuite/tests/rts/ipe/distinct-tables/distinct_tables02.stdout
- + testsuite/tests/rts/ipe/distinct-tables/distinct_tables03.stdout
- + testsuite/tests/rts/ipe/distinct-tables/distinct_tables04.stdout
- + testsuite/tests/rts/ipe/distinct-tables/distinct_tables05.stdout
- + testsuite/tests/rts/ipe/distinct-tables/distinct_tables06.stdout
- + testsuite/tests/rts/ipe/distinct-tables/distinct_tables07.stdout
- + testsuite/tests/rts/ipe/distinct-tables/distinct_tables08.stdout
- + testsuite/tests/rts/ipe/distinct-tables/distinct_tables09.stdout
- + testsuite/tests/rts/ipe/distinct-tables/distinct_tables10.stdout
- + testsuite/tests/rts/ipe/distinct-tables/distinct_tables11.stdout
- + testsuite/tests/rts/ipe/distinct-tables/distinct_tables12.stdout
- + testsuite/tests/rts/ipe/distinct-tables/distinct_tables13.stdout
- testsuite/tests/simd/should_run/all.T
- + testsuite/tests/simd/should_run/simd015.hs
- + testsuite/tests/simd/should_run/simd015.stdout
- + testsuite/tests/simd/should_run/simd016.hs
- + testsuite/tests/simd/should_run/simd016.stdout
- + testsuite/tests/simd/should_run/simd017.hs
- + testsuite/tests/simd/should_run/simd017.stdout
- testsuite/tests/simplCore/should_compile/OpaqueNoCastWW.stderr
- testsuite/tests/simplCore/should_compile/T17673.stderr
- testsuite/tests/simplCore/should_compile/T18078.stderr
- testsuite/tests/simplCore/should_compile/T18995.stderr
- testsuite/tests/simplCore/should_compile/T19890.stderr
- testsuite/tests/simplCore/should_compile/T21948.stderr
- testsuite/tests/simplCore/should_compile/T21960.stderr
- testsuite/tests/simplCore/should_compile/T24808.stderr
- − testsuite/tests/simplCore/should_compile/T25713.stderr
- testsuite/tests/simplCore/should_compile/T4201.stdout
- testsuite/tests/simplCore/should_compile/T8331.stderr
- testsuite/tests/simplCore/should_compile/rule2.stderr
- testsuite/tests/simplStg/should_compile/T22840.stderr
- testsuite/tests/th/T8761.stderr
- testsuite/tests/typecheck/no_skolem_info/T13499.stderr
- testsuite/tests/typecheck/no_skolem_info/T20232.stderr
- testsuite/tests/typecheck/should_compile/T13651.hs
- − testsuite/tests/typecheck/should_compile/T13651.stderr
- + testsuite/tests/typecheck/should_compile/T14745.hs
- + testsuite/tests/typecheck/should_compile/T17705.hs
- testsuite/tests/typecheck/should_compile/all.T
- testsuite/tests/typecheck/should_compile/hole_constraints_nested.stderr
- testsuite/tests/typecheck/should_compile/tc126.hs
- testsuite/tests/typecheck/should_fail/AmbigFDs.hs
- − testsuite/tests/typecheck/should_fail/AmbigFDs.stderr
- testsuite/tests/typecheck/should_fail/FD3.stderr
- testsuite/tests/typecheck/should_fail/FDsFromGivens2.stderr
- testsuite/tests/typecheck/should_fail/T11672.stderr
- testsuite/tests/typecheck/should_fail/T12373.stderr
- testsuite/tests/typecheck/should_fail/T13506.stderr
- testsuite/tests/typecheck/should_fail/T15807.stderr
- testsuite/tests/typecheck/should_fail/T16074.stderr
- testsuite/tests/typecheck/should_fail/T16512a.stderr
- testsuite/tests/typecheck/should_fail/T18357a.stderr
- testsuite/tests/typecheck/should_fail/T18851b.hs
- − testsuite/tests/typecheck/should_fail/T18851b.stderr
- testsuite/tests/typecheck/should_fail/T18851c.hs
- − testsuite/tests/typecheck/should_fail/T18851c.stderr
- testsuite/tests/typecheck/should_fail/T19415.stderr
- testsuite/tests/typecheck/should_fail/T19415b.stderr
- testsuite/tests/typecheck/should_fail/T19627.stderr
- testsuite/tests/typecheck/should_fail/T21530a.stderr
- testsuite/tests/typecheck/should_fail/T22684.stderr
- + testsuite/tests/typecheck/should_fail/T23162a.hs
- + testsuite/tests/typecheck/should_fail/T23162a.stderr
- testsuite/tests/typecheck/should_fail/T25325.stderr
- testsuite/tests/typecheck/should_fail/T5246.stderr
- testsuite/tests/typecheck/should_fail/T5978.stderr
- testsuite/tests/typecheck/should_fail/T7368a.stderr
- testsuite/tests/typecheck/should_fail/T7696.stderr
- testsuite/tests/typecheck/should_fail/T8603.stderr
- testsuite/tests/typecheck/should_fail/T9612.stderr
- testsuite/tests/typecheck/should_fail/TcStaticPointersFail03.stderr
- testsuite/tests/typecheck/should_fail/VisFlag1.stderr
- testsuite/tests/typecheck/should_fail/all.T
- testsuite/tests/typecheck/should_fail/tcfail122.stderr
- testsuite/tests/typecheck/should_fail/tcfail143.stderr
- + testsuite/tests/warnings/should_compile/DodgyExports02.hs
- + testsuite/tests/warnings/should_compile/DodgyExports02.stderr
- + testsuite/tests/warnings/should_compile/DodgyExports03.hs
- + testsuite/tests/warnings/should_compile/DodgyExports03.stderr
- + testsuite/tests/warnings/should_compile/DuplicateModExport.hs
- + testsuite/tests/warnings/should_compile/DuplicateModExport.stderr
- + testsuite/tests/warnings/should_compile/EmptyModExport.hs
- + testsuite/tests/warnings/should_compile/EmptyModExport.stderr
- testsuite/tests/warnings/should_compile/all.T
- utils/check-exact/ExactPrint.hs
- utils/check-exact/Parsers.hs
- utils/check-exact/Preprocess.hs
- utils/deriveConstants/Main.hs
- utils/genprimopcode/genprimopcode.cabal
- utils/ghc-toolchain/ghc-toolchain.cabal
- utils/ghc-toolchain/src/GHC/Toolchain/CheckArm.hs
- + utils/ghc-toolchain/src/GHC/Toolchain/CheckPower.hs
- utils/ghc-toolchain/src/GHC/Toolchain/ParseTriple.hs
- utils/ghc-toolchain/src/GHC/Toolchain/Tools/MergeObjs.hs
- utils/ghc-toolchain/src/GHC/Toolchain/Utils.hs
- utils/haddock/cabal.project
- utils/haddock/haddock-api/src/Haddock/Convert.hs
- utils/haddock/haddock-api/src/Haddock/Interface/Rename.hs
- utils/haddock/haddock-api/src/Haddock/Types.hs
- utils/jsffi/dyld.mjs
The diff was not included because it is too large.
View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/41a76da820158a058f541bf726e11e…
--
View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/41a76da820158a058f541bf726e11e…
You're receiving this email because of your account on gitlab.haskell.org.
1
0