- ghc-commits - Haskell.org

[Git][ghc/ghc] Deleted branch wip/juhp-hp2ps-c23
by Jens Petersen (＠juhp) 02 May '26

02 May '26

Jens Petersen deleted branch wip/juhp-hp2ps-c23 at Glasgow Haskell Compiler / GHC -- You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc] Deleted branch wip/juhp-hp2ps-fixup2
by Jens Petersen (＠juhp) 02 May '26

02 May '26

Jens Petersen deleted branch wip/juhp-hp2ps-fixup2 at Glasgow Haskell Compiler / GHC -- You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc] Deleted branch wip/juhp-hp2ps-libstdh
by Jens Petersen (＠juhp) 02 May '26

02 May '26

Jens Petersen deleted branch wip/juhp-hp2ps-libstdh at Glasgow Haskell Compiler / GHC -- You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc] Deleted branch juhp-ghc-9.10-patch-56735
by Jens Petersen (＠juhp) 02 May '26

02 May '26

Jens Petersen deleted branch juhp-ghc-9.10-patch-56735 at Glasgow Haskell Compiler / GHC -- You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc] Deleted branch wip/juhp-disable-hadrian-selftest
by Jens Petersen (＠juhp) 02 May '26

02 May '26

Jens Petersen deleted branch wip/juhp-disable-hadrian-selftest at Glasgow Haskell Compiler / GHC -- You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc] Deleted branch juhp-ghc-9.12-genSym-backports
by Jens Petersen (＠juhp) 02 May '26

02 May '26

Jens Petersen deleted branch juhp-ghc-9.12-genSym-backports at Glasgow Haskell Compiler / GHC -- You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][wip/T26989] Wibbles
by Simon Peyton Jones (＠simonpj) 02 May '26

02 May '26

Simon Peyton Jones pushed to branch wip/T26989 at Glasgow Haskell Compiler / GHC Commits: 3925e6fb by Simon Peyton Jones at 2026-05-02T13:15:55+01:00 Wibbles - - - - - 4 changed files: - compiler/GHC/Core/Opt/Simplify/Iteration.hs - compiler/GHC/Core/SimpleOpt.hs - compiler/GHC/Data/List/SetOps.hs - compiler/GHC/Types/Id/Make.hs Changes: ===================================== compiler/GHC/Core/Opt/Simplify/Iteration.hs ===================================== @@ -2052,7 +2052,7 @@ Note [Avoiding simplifying repeatedly] One way in which we can get exponential behaviour is if we simplify a big expression, and then re-simplify it -- and then this happens in a deeply-nested way. So we must be jolly careful about re-simplifying -an expression (#26989). +an expression (#26989). Example: f BIG, where f has a RULE @@ -2359,7 +2359,8 @@ simplInId env var cont , isLazyDataConRep dc -- See Note [Fast path for lazy data constructors] = rebuild env (Var var) cont | otherwise - = case substId env var of + = assertPpr (not (isCoVar var)) (ppr var) $ + case substId env var of ContEx se e mco -> do { cont' <- pushCastMCo env mco cont ; simplExprF (se `setInScopeFromE` env) e cont' } ===================================== compiler/GHC/Core/SimpleOpt.hs ===================================== @@ -326,8 +326,7 @@ simple_opt_expr env expr = go expr (env', b') = subst_opt_bndr env b ---------------------- - go_co co = pprTrace "add_cast" (ppr (optCoercionEnabled $ so_co_opts (soe_opts env)) $$ ppr co) $ - optCoercion (so_co_opts (soe_opts env)) subst co + go_co co = optCoercion (so_co_opts (soe_opts env)) subst co ---------------------- go_alt env (Alt con bndrs rhs) ===================================== compiler/GHC/Data/List/SetOps.hs ===================================== @@ -40,7 +40,7 @@ import Data.List.NonEmpty (NonEmpty(..)) import Data.Ord (comparing) import qualified Data.Set as S -getNth :: Outputable a => [a] -> Int -> a +getNth :: (HasDebugCallStack, Outputable a) => [a] -> Int -> a getNth xs n = assertPpr (xs `lengthExceeds` n) (ppr n $$ ppr xs) $ xs !! n ===================================== compiler/GHC/Types/Id/Make.hs ===================================== @@ -564,7 +564,7 @@ dictSelRule :: Name -> Arity -> Int -> CoreRule -- sel_i t1..tk (D t1..tk op1 ... opm) = opi -- -- See Note [ClassOp/DFun selection] in GHC.Tc.TyCl.Instance -dictSelRule name val_index n_ty_args +dictSelRule name n_ty_args val_index = rule where rule = BuiltinRule { ru_name = fsLit "Class op " `appendFS` @@ -579,14 +579,25 @@ dictSelRule name val_index n_ty_args | (dict_arg : _) <- drop n_ty_args args , Just (_, floats, _, _, con_args) <- exprIsConApp_maybe in_scope_env dict_arg , let meth_e = getNth con_args val_index - meth_id = mkTemplateLocal 1 (exprType meth_e) = Just (RM { rm_floats = floats - , rm_rhs = Lam meth_id (Var meth_id) - , rm_args = [meth_e] - , rm_rule = rule }) + , rm_rhs = mkIdLam (exprType meth_e) + , rm_args = [meth_e] + , rm_rule = rule }) | otherwise = Nothing + +mkIdLam :: Type -> CoreExpr +-- Make an identity lambda (\(x::ty).x), already occ-analysed +mkIdLam ty + = Lam x (varToCoreExpr x) + where + x = mkTemplateLocal 1 ty + `setIdOccInfo` OneOcc { occ_in_lam = NotInsideLam + , occ_n_br = 1 + , occ_int_cxt = NotInteresting + , occ_tail = NoTailCallInfo } + {- ************************************************************************ * * View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/3925e6fb140ca7c3a22c8746b025ea3… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/3925e6fb140ca7c3a22c8746b025ea3… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][wip/T27022] iserv: recover/getQ/putQ should behave same as internal interpreter
by Teo Camarasu (＠teo) 02 May '26

02 May '26

Teo Camarasu pushed to branch wip/T27022 at Glasgow Haskell Compiler / GHC Commits: ab0e7b30 by Teo Camarasu at 2026-05-02T12:24:44+01:00 iserv: recover/getQ/putQ should behave same as internal interpreter The internal and external interpreter should behave the same when handling `recover`, the exeception recovery method of Q. In practice, they diverge. In case of failure, the internal interpreter only restores error message state to before the computation, wheras the external interperter restores error message state *and* the state of putQ/getQ. As far as I can tell this is a simple mistake in the implementation. Note [TH recover with -fexternal-interpreter] describes the correct behaviour but the implementation doesn't mirror this. This change restores the correct behaviour by keeping the effects of putQ in the erroring computation. This is a breaking change since it modifies the behaviour of programs that rely on recover ignoring putQ from failling computations when used with the external interpreter. Although I highly doubt anyone relies on this behaviour. This divergence was first introduced in d00c308633fe7d216d31a1087e00e63532d87d6d. As far as I can tell this was unintentional and tha commit was trying to solve a different bug. Resolves #27022 - - - - - 4 changed files: - + changelog.d/T27022 - libraries/ghci/GHCi/TH.hs - + testsuite/tests/th/T27022.hs - testsuite/tests/th/all.T Changes: ===================================== changelog.d/T27022 ===================================== @@ -0,0 +1,11 @@ +section: compiler +synopsis: Fix a divergence in the interaction between ``recover`` and ``putQ`` between the internal and external interpreter +description: The ``recover`` method in TemplateHaskell now behaves the same + with the internal and external interpreter. + In the past, when an error was encountered in a computation in a ``recover`` block, + the external interpreter would discard any state changes from ``putQ``, + whereas the internal interpreter would not. + This was a long-standing error in the implementation of the external interpreter. + Both now keep state changes from ``putQ`` in ``recover`` blocks. +mr: !15994 +issues: #27022 ===================================== libraries/ghci/GHCi/TH.hs ===================================== @@ -175,9 +175,7 @@ instance TH.Quasi GHCiQ where e <- try $ unmask $ runGHCiQ (a <* ghcCmd FailIfErrs) sRef remoteTHCall (qsPipe s) (EndRecover (isLeft e)) case e of - Left GHCiQException{} -> - -- in case of error, restore the state to the start of the `recover` block. - newIORef s >>= h + Left GHCiQException{} -> h sRef Right r -> return r qLookupName isType occ = ghcCmd (LookupName isType occ) qReify name = ghcCmd (Reify name) ===================================== testsuite/tests/th/T27022.hs ===================================== @@ -0,0 +1,8 @@ +{-# LANGUAGE TemplateHaskell #-} +-- | This tests the behaviour of TH's recover method. +-- It should behave the same in the internal and external interperter. +-- In the past, they have diverged, and the external interpreter would roll back the state of putQ/getQ whereas the internal interpreter would not. +module T27022 where + +import Language.Haskell.TH.Syntax +main = print $(putQ "0" >> recover (pure ()) (putQ "42" >> fail "oops") >> getQ @String >>= lift ) ===================================== testsuite/tests/th/all.T ===================================== @@ -650,3 +650,4 @@ test('GadtConSigs_th_dump1', normal, compile, ['-v0 -ddump-splices -dsuppress-un test('T26099', normal, compile_fail, ['']) test('T8306_th', only_ways(['ghci']), ghci_script, ['T8306_th.script']) test('T26862_th', only_ways(['ghci']), ghci_script, ['T26862_th.script']) +test('T27022', normal, compile, ['']) View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/ab0e7b306f6b009845c8f6a7f05a24f… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/ab0e7b306f6b009845c8f6a7f05a24f… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][wip/T27022] iserv: recover/getQ/putQ should behave same as internal interpreter
by Teo Camarasu (＠teo) 02 May '26

02 May '26

Teo Camarasu pushed to branch wip/T27022 at Glasgow Haskell Compiler / GHC Commits: 9369b573 by Teo Camarasu at 2026-05-02T12:17:42+01:00 iserv: recover/getQ/putQ should behave same as internal interpreter The internal and external interpreter should behave the same when handling `recover`, the exeception recovery method of Q. In practice, they diverge. In case of failure, the internal interpreter only restores error message state to before the computation, wheras the external interperter restores error message state *and* the state of putQ/getQ. As far as I can tell this is a simple mistake in the implementation. Note [TH recover with -fexternal-interpreter] describes the correct behaviour but the implementation doesn't mirror this. This change restores the correct behaviour by keeping the effects of putQ in the erroring computation. This is a breaking change since it modifies the behaviour of programs that rely on recover ignoring putQ from failling computations when used with the external interpreter. Although I highly doubt anyone relies on this behaviour. Resolves #27022 - - - - - 4 changed files: - + changelog.d/T27022 - libraries/ghci/GHCi/TH.hs - + testsuite/tests/th/T27022.hs - testsuite/tests/th/all.T Changes: ===================================== changelog.d/T27022 ===================================== @@ -0,0 +1,11 @@ +section: compiler +synopsis: Fix a divergence in the interaction between ``recover`` and ``putQ`` between the internal and external interpreter +description: The ``recover`` method in TemplateHaskell now behaves the same + with the internal and external interpreter. + In the past, when an error was encountered in a computation in a ``recover`` block, + the external interpreter would discard any state changes from ``putQ``, + whereas the internal interpreter would not. + This was a long-standing error in the implementation of the external interpreter. + Both now keep state changes from ``putQ`` in ``recover`` blocks. +mr: !15994 +issues: #27022 ===================================== libraries/ghci/GHCi/TH.hs ===================================== @@ -175,9 +175,7 @@ instance TH.Quasi GHCiQ where e <- try $ unmask $ runGHCiQ (a <* ghcCmd FailIfErrs) sRef remoteTHCall (qsPipe s) (EndRecover (isLeft e)) case e of - Left GHCiQException{} -> - -- in case of error, restore the state to the start of the `recover` block. - newIORef s >>= h + Left GHCiQException{} -> h sRef Right r -> return r qLookupName isType occ = ghcCmd (LookupName isType occ) qReify name = ghcCmd (Reify name) ===================================== testsuite/tests/th/T27022.hs ===================================== @@ -0,0 +1,8 @@ +{-# LANGUAGE TemplateHaskell #-} +-- | This tests the behaviour of TH's recover method. +-- It should behave the same in the internal and external interperter. +-- In the past, they have diverged, and the external interpreter would roll back the state of putQ/getQ whereas the internal interpreter would not. +module T27022 where + +import Language.Haskell.TH.Syntax +main = print $(putQ "0" >> recover (pure ()) (putQ "42" >> fail "oops") >> getQ @String >>= lift ) ===================================== testsuite/tests/th/all.T ===================================== @@ -650,3 +650,4 @@ test('GadtConSigs_th_dump1', normal, compile, ['-v0 -ddump-splices -dsuppress-un test('T26099', normal, compile_fail, ['']) test('T8306_th', only_ways(['ghci']), ghci_script, ['T8306_th.script']) test('T26862_th', only_ways(['ghci']), ghci_script, ['T26862_th.script']) +test('T27022', normal, compile, ['']) View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/9369b5739826150b0fd067ac8bc1b00… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/9369b5739826150b0fd067ac8bc1b00… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc] Pushed new branch wip/T27022
by Teo Camarasu (＠teo) 02 May '26

02 May '26

Teo Camarasu pushed new branch wip/T27022 at Glasgow Haskell Compiler / GHC -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/tree/wip/T27022 You're receiving this email because of your account on gitlab.haskell.org.

1 0