|
1
|
|
-.. _release-9-14-1:
|
|
2
|
|
-
|
|
3
|
|
-Version 9.14.1
|
|
4
|
|
-==============
|
|
5
|
|
-
|
|
6
|
|
-The significant changes to the various parts of the compiler are listed in the
|
|
7
|
|
-following sections. See the `migration guide
|
|
8
|
|
-<https://gitlab.haskell.org/ghc/ghc/-/wikis/migration/9.14>`_ on the GHC Wiki
|
|
9
|
|
-for specific guidance on migrating programs to this release.
|
|
10
|
|
-
|
|
11
|
|
-Language
|
|
12
|
|
-~~~~~~~~
|
|
13
|
|
-
|
|
14
|
|
-* `GHC proposal 493: allow expressions in SPECIALISE pragmas <https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0493-specialise-expressions.rst>`_
|
|
15
|
|
- has been implemented. SPECIALISE pragmas now allow arbitrary expressions such as: ::
|
|
16
|
|
-
|
|
17
|
|
- {-# SPECIALISE f @Int False :: Int -> Char #-}
|
|
18
|
|
-
|
|
19
|
|
- The ability to specify multiple specialisations in a single SPECIALISE pragma,
|
|
20
|
|
- with syntax of the form (note the comma between the type signatures): ::
|
|
21
|
|
-
|
|
22
|
|
- {-# SPECIALISE g : Int -> Int, Float -> Float #-}
|
|
23
|
|
-
|
|
24
|
|
- has been deprecated, and is scheduled to be removed in GHC 9.18.
|
|
25
|
|
- This deprecation is controlled by the newly introduced ``-Wdeprecated-pragmas``
|
|
26
|
|
- flag in ``-Wdefault``.
|
|
27
|
|
-
|
|
28
|
|
-* ``-Wincomplete-record-selectors`` is now part of `-Wall`, as specified
|
|
29
|
|
- by `GHC Proposal 516: add warning for incomplete record selectors <https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0516-incomplete-record-selectors.rst>`_.
|
|
30
|
|
- Hence, if a library is compiled with ``-Werror``, compilation may now fail. Solution: fix the library.
|
|
31
|
|
- Workaround: add ``-Werror=no-incomplete-record-selectors``.
|
|
32
|
|
-
|
|
33
|
|
- Note that this warning is at least
|
|
34
|
|
- as serious as a warning about missing patterns from a function definition, perhaps even
|
|
35
|
|
- more so, since it is invisible in the source program.
|
|
36
|
|
-
|
|
37
|
|
-* The combination of :extension:`ScopedTypeVariables` and :extension:`TypeApplications`
|
|
38
|
|
- no longer enables type applications in patterns, which now always requires
|
|
39
|
|
- :extension:`TypeAbstractions`. The warning flag``deprecated-type-abstractions``
|
|
40
|
|
- has also been removed from the compiler.
|
|
41
|
|
-
|
|
42
|
|
-* :extension:`OverloadedRecordUpdate` now passes the arguments to a ``setField`` function
|
|
43
|
|
- in the flipped order, as specified by `GHC Proposal 583: HasField redesign <https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0583-hasfield-redesign.rst>`_.
|
|
44
|
|
-
|
|
45
|
|
- Previously GHC expected ``setField`` to have this type: ::
|
|
46
|
|
-
|
|
47
|
|
- setField :: forall (fld :: Symbol) a r. r -> a -> r
|
|
48
|
|
-
|
|
49
|
|
- And that's what GHC expects now: ::
|
|
50
|
|
-
|
|
51
|
|
- setField :: forall (fld :: Symbol) a r. a -> r -> r
|
|
52
|
|
-
|
|
53
|
|
- That will break the combination of :extension:`OverloadedRecordUpdate` with :extension:`RebindableSyntax`.
|
|
54
|
|
-
|
|
55
|
|
-* Multiline strings are now accepted in foreign imports. (#25157)
|
|
56
|
|
-
|
|
57
|
|
-* GHC now does a better job at inferring types in calls to ``coerce``: instead of
|
|
58
|
|
- complaining about ambiguous type variables, GHC will consider that such type
|
|
59
|
|
- variables are determined by the ``Coercible`` constraints they appear in.
|
|
60
|
|
-
|
|
61
|
|
-* With :extension:`LinearTypes` record fields can now be non-linear. This means that
|
|
62
|
|
- the following record declaration is now valid:
|
|
63
|
|
-
|
|
64
|
|
- ::
|
|
65
|
|
-
|
|
66
|
|
- data Record = Rec { x %'Many :: Int, y :: Char }
|
|
67
|
|
-
|
|
68
|
|
- This causes the constructor to have type ``Rec :: Int %'Many -> Char %1 -> Record``.
|
|
69
|
|
-
|
|
70
|
|
-* The :extension:`ExplicitNamespaces` extension now allows the ``data``
|
|
71
|
|
- namespace specifier in import and export lists.
|
|
72
|
|
-
|
|
73
|
|
-* The ``-Wdata-kinds-tc`` warning has been deprecated, and the use of promoted
|
|
74
|
|
- data types in kinds is now an error (rather than a warning) unless the
|
|
75
|
|
- :extension:`DataKinds` extension is enabled. For example, the following code
|
|
76
|
|
- will be rejected unless :extension:`DataKinds` is on:
|
|
77
|
|
-
|
|
78
|
|
- import Data.Kind (Type)
|
|
79
|
|
- import GHC.TypeNats (Nat)
|
|
80
|
|
-
|
|
81
|
|
- -- Nat shouldn't be allowed here without DataKinds
|
|
82
|
|
- data Vec :: Nat -> Type -> Type
|
|
83
|
|
-
|
|
84
|
|
- (The ``-Wdata-kinds-tc`` warning was introduced in GHC 9.10 as part of a fix
|
|
85
|
|
- for an accidental oversight in which programs like the one above were
|
|
86
|
|
- mistakenly accepted without the use of :extension:`DataKinds`.)
|
|
87
|
|
-
|
|
88
|
|
-* The :extension:`MonadComprehensions` extension now implies :extension:`ParallelListComp` as was originally intended (see `Monad Comprehensions <https://ghc.gitlab.haskell.org/ghc/doc/users_guide/exts/monad_comprehensions.html>`_).
|
|
89
|
|
-
|
|
90
|
|
-* In accordance with `GHC Proposal #281 <https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0281-visible-forall.rst>`_,
|
|
91
|
|
- section 4.7 "Data constructors", the :extension:`RequiredTypeArguments`
|
|
92
|
|
- extension now allows visible forall in types of data constructors
|
|
93
|
|
- (:ghc-ticket:`25127`). The following declaration is now accepted by GHC:
|
|
94
|
|
-
|
|
95
|
|
- ::
|
|
96
|
|
-
|
|
97
|
|
- data T a where
|
|
98
|
|
- Typed :: forall a -> a -> T a
|
|
99
|
|
-
|
|
100
|
|
- See :ref:`visible-forall-in-gadts` for details.
|
|
101
|
|
-
|
|
102
|
|
-Compiler
|
|
103
|
|
-~~~~~~~~
|
|
104
|
|
-
|
|
105
|
|
-- An improved error message is introduced to refer users to the heap-controlling flags of the RTS when there is a heap overflow during compilation. (#25198)
|
|
106
|
|
-
|
|
107
|
|
-- The kind checker now does a better job of finding type family instances for
|
|
108
|
|
- use in the kinds of other declarations in the same module. This fixes a number
|
|
109
|
|
- of tickets:
|
|
110
|
|
- :ghc-ticket:`12088`, :ghc-ticket:`12239`, :ghc-ticket:`14668`, :ghc-ticket:`15561`,
|
|
111
|
|
- :ghc-ticket:`16410`, :ghc-ticket:`16448`, :ghc-ticket:`16693`, :ghc-ticket:`19611`,
|
|
112
|
|
- :ghc-ticket:`20875`, :ghc-ticket:`21172`, :ghc-ticket:`22257`, :ghc-ticket:`25238`,
|
|
113
|
|
- :ghc-ticket:`25834`.
|
|
114
|
|
-
|
|
115
|
|
-- The compiler no longer accepts invalid ``type`` namespace specifiers in
|
|
116
|
|
- subordinate import lists (:ghc-ticket:`22581`).
|
|
117
|
|
-
|
|
118
|
|
-- A new flag, :ghc-flag:`-Wuseless-specialisations`, controls warnings emitted when GHC
|
|
119
|
|
- determines that a SPECIALISE pragma would have no effect.
|
|
120
|
|
-
|
|
121
|
|
-- A new flag, :ghc-flag:`-Wrule-lhs-equalities`, controls warnings emitted for RULES
|
|
122
|
|
- whose left-hand side attempts to quantify over equality constraints that
|
|
123
|
|
- previous GHC versions accepted quantifying over. GHC will now drop such RULES,
|
|
124
|
|
- emitting a warning message controlled by this flag.
|
|
125
|
|
-
|
|
126
|
|
- This warning is intended to give visibility to the fact that the RULES that
|
|
127
|
|
- previous GHC versions generated in such circumstances could never fire.
|
|
128
|
|
-
|
|
129
|
|
-- A new flag, :ghc-flag:`-Wunusable-unpack-pragmas`, controls warnings emitted
|
|
130
|
|
- when GHC is unable to unpack a data constructor field annotated by the
|
|
131
|
|
- ``{-# UNPACK #-}`` pragma.
|
|
132
|
|
-
|
|
133
|
|
- Previous GHC versions issued this warning unconditionally. Now it is possible
|
|
134
|
|
- to disable it with ``-Wno-unusable-unpack-pragmas`` or turn it into an error
|
|
135
|
|
- with ``-Werror=unusable-unpack-pragmas``.
|
|
136
|
|
-
|
|
137
|
|
-- Introduce a new warning :ghc-flag:`-Wpattern-namespace-specifier` to detect
|
|
138
|
|
- uses of the now deprecated ``pattern`` namespace specifier in import/export
|
|
139
|
|
- lists. See `GHC Proposal #581, section 2.3 <https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0581-namespace-specified-imports.rst#deprecate-use-of-pattern-in-import-export-lists>`_.
|
|
140
|
|
-
|
|
141
|
|
-- Code coverage (:ghc-flag:`-fhpc`) now treats uses of record fields
|
|
142
|
|
- (including via :extension:`RecordWildCards` or :extension:`NamedFieldPuns`)
|
|
143
|
|
- as if the fields were accessed using the generated record selector functions,
|
|
144
|
|
- marking the fields as covered in coverage reports (:ghc-ticket:`17834`,
|
|
145
|
|
- :ghc-ticket:`26191`). Note that this currently only works when record fields
|
|
146
|
|
- (or values contained within them) are bound to variables, and usage of those
|
|
147
|
|
- variables marks the record selectors as covered. That is, a pattern like
|
|
148
|
|
- ``Foo{bar = Bar{baz = b}}`` will mark ``bar`` and ``baz`` as covered if ``b``
|
|
149
|
|
- is used, but the similar pattern ``Foo{bar = Bar{baz = 42}}`` will mark
|
|
150
|
|
- neither as covered.
|
|
151
|
|
-
|
|
152
|
|
-- SIMD support in the X86 native code generator has been extended with 128-bit
|
|
153
|
|
- integer operations. Also, ``shuffleFloatX4#`` and ``shuffleDoubleX2#`` no longer
|
|
154
|
|
- require ``-mavx``.
|
|
155
|
|
-
|
|
156
|
|
-- JSON diagnostics produced with (:ghc-flag:`-fdiagnostics-as-json`) now
|
|
157
|
|
- include the `rendered` diagnostics message, in the exact same format as what
|
|
158
|
|
- GHC would have produced without -fdiagnostics-as-json (including ANSI escape
|
|
159
|
|
- sequences).
|
|
160
|
|
-
|
|
161
|
|
-GHCi
|
|
162
|
|
-~~~~
|
|
163
|
|
-
|
|
164
|
|
-- :ghci-cmd:`:info` now outputs type declarations with @-binders that are
|
|
165
|
|
- considered semantically significant. See the documentation for :ghci-cmd:`:info`
|
|
166
|
|
- itself for a more detailed explanation.
|
|
167
|
|
-
|
|
168
|
|
-- GHCi errors and warnings now have their own numeric error codes that are
|
|
169
|
|
- displayed alongside the error.
|
|
170
|
|
-
|
|
171
|
|
-Runtime system
|
|
172
|
|
-~~~~~~~~~~~~~~
|
|
173
|
|
-
|
|
174
|
|
-- Add new runtime flag :rts-flag:`--optimistic-linking` which instructs the
|
|
175
|
|
- runtime linker to continue in the presence of unknown symbols. By default this
|
|
176
|
|
- flag is not passed, preserving previous behavior.
|
|
177
|
|
-
|
|
178
|
|
-Cmm
|
|
179
|
|
-~~~
|
|
180
|
|
-
|
|
181
|
|
-``base`` library
|
|
182
|
|
-~~~~~~~~~~~~~~~~
|
|
183
|
|
-
|
|
184
|
|
-``ghc-prim`` library
|
|
185
|
|
-~~~~~~~~~~~~~~~~~~~~
|
|
186
|
|
-
|
|
187
|
|
-``ghc`` library
|
|
188
|
|
-~~~~~~~~~~~~~~~
|
|
189
|
|
-
|
|
190
|
|
-* The `UnknownDiagnostic` constructor now takes an additional type argument
|
|
191
|
|
- for the type of hints corresponding to the diagnostic, and an additional
|
|
192
|
|
- value-level argument used for existential wrapping of the hints of the inner
|
|
193
|
|
- diagnostic.
|
|
194
|
|
-
|
|
195
|
|
-* Changes to the HPT and HUG interface:
|
|
196
|
|
-
|
|
197
|
|
- - `addToHpt` and `addListToHPT` were moved from `GHC.Unit.Home.ModInfo` to `GHC.Unit.Home.PackageTable` and deprecated in favour of `addHomeModInfoToHpt` and `addHomeModInfosToHpt`.
|
|
198
|
|
- - `UnitEnvGraph` and operations `unitEnv_lookup_maybe`, `unitEnv_foldWithKey, `unitEnv_singleton`, `unitEnv_adjust`, `unitEnv_insert`, `unitEnv_new` were moved from `GHC.Unit.Env` to `GHC.Unit.Home.Graph`.
|
|
199
|
|
- - The HomePackageTable (HPT) is now exported from `GHC.Unit.Home.PackageTable`,
|
|
200
|
|
- and is now backed by an IORef to avoid by construction very bad memory leaks.
|
|
201
|
|
- This means the API to the HPT now is for the most part in IO. For instance,
|
|
202
|
|
- `emptyHomePackageTable` and `addHomeModInfoToHpt` are now in IO.
|
|
203
|
|
- - `mkHomeUnitEnv` was moved to `GHC.Unit.Home.PackageTable`, and now takes two
|
|
204
|
|
- extra explicit arguments. To restore previous behaviour, pass `emptyUnitState`
|
|
205
|
|
- and `Nothing` as the first two arguments additionally.
|
|
206
|
|
- - `hugElts` was removed. Users should prefer `allUnits` to get the keys of the
|
|
207
|
|
- HUG (the typical use case), or `traverse` or `unitEnv_foldWithKey` in other
|
|
208
|
|
- cases.
|
|
209
|
|
-
|
|
210
|
|
-* Changes to `Language.Haskell.Syntax.Expr`
|
|
211
|
|
-
|
|
212
|
|
- - The `ParStmtBlock` list argument of the `ParStmt` constructor of `StmtLR` is now `NonEmpty`.
|
|
213
|
|
-
|
|
214
|
|
-* As part of the implementation of `GHC proposal 493 <https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0493-specialise-expressions.rst>`_,
|
|
215
|
|
- the `SpecSig` constructor of `Sig` has been deprecated. It is replaced by
|
|
216
|
|
- the constructor `SpecSigE` which supports expressions at the head, rather than
|
|
217
|
|
- a lone variable.
|
|
218
|
|
-
|
|
219
|
|
-``ghc-heap`` library
|
|
220
|
|
-~~~~~~~~~~~~~~~~~~~~
|
|
221
|
|
-
|
|
222
|
|
-* The functions `getClosureInfoTbl_maybe`, `getClosureInfoTbl`,
|
|
223
|
|
- `getClosurePtrArgs` and `getClosurePtrArgs_maybe` have been added to allow
|
|
224
|
|
- reading of the relevant Closure attributes without reliance on incomplete
|
|
225
|
|
- selectors.
|
|
226
|
|
-
|
|
227
|
|
-``ghc-experimental`` library
|
|
228
|
|
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~
|
|
229
|
|
-
|
|
230
|
|
-- ``ghc-experimental`` now exposes ``GHC.RTS.Flags`` and ``GHC.Stats`` as
|
|
231
|
|
- ``GHC.RTS.Flags.Experimental`` and ``GHC.Stats.Experimental``. These are
|
|
232
|
|
- *also* exposed in ``base``, however the ``base`` versions will be deprecated as
|
|
233
|
|
- part of the split base project. See `CLC proposal 289
|
|
234
|
|
- <https://github.com/haskell/core-libraries-committee/issues/289>`__.
|
|
235
|
|
- Downstream consumers of these flags are encouraged to migrate to the
|
|
236
|
|
- ``ghc-experimental`` versions.
|
|
237
|
|
-
|
|
238
|
|
-
|
|
239
|
|
-
|
|
240
|
|
-``template-haskell`` library
|
|
241
|
|
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~
|
|
242
|
|
-
|
|
243
|
|
-- As part of the implementation of `GHC proposal 493 <https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0493-specialise-expressions.rst>`_,
|
|
244
|
|
- the ``SpecialiseP`` constructor of the Template Haskell ``Pragma`` type, as
|
|
245
|
|
- well as the helpers ``pragSpecD`` and ``pragSpecInlD``, have been deprecated.
|
|
246
|
|
-
|
|
247
|
|
- They are replaced, respectively, by ``SpecialiseEP``, ``pragSpecED`` and
|
|
248
|
|
- ``pragSpecInlED``.
|
|
249
|
|
-
|
|
250
|
|
-Included libraries
|
|
251
|
|
-~~~~~~~~~~~~~~~~~~
|
|
252
|
|
-
|
|
253
|
|
-The package database provided with this distribution also contains a number of
|
|
254
|
|
-packages other than GHC itself. See the changelogs provided with these packages
|
|
255
|
|
-for further change information.
|
|
256
|
|
-
|
|
257
|
|
-.. ghc-package-list::
|
|
258
|
|
-
|
|
259
|
|
- libraries/array/array.cabal: Dependency of ``ghc`` library
|
|
260
|
|
- libraries/base/base.cabal: Core library
|
|
261
|
|
- libraries/binary/binary.cabal: Dependency of ``ghc`` library
|
|
262
|
|
- libraries/bytestring/bytestring.cabal: Dependency of ``ghc`` library
|
|
263
|
|
- libraries/Cabal/Cabal/Cabal.cabal: Dependency of ``ghc-pkg`` utility
|
|
264
|
|
- libraries/Cabal/Cabal-syntax/Cabal-syntax.cabal: Dependency of ``ghc-pkg`` utility
|
|
265
|
|
- libraries/containers/containers/containers.cabal: Dependency of ``ghc`` library
|
|
266
|
|
- libraries/deepseq/deepseq.cabal: Dependency of ``ghc`` library
|
|
267
|
|
- libraries/directory/directory.cabal: Dependency of ``ghc`` library
|
|
268
|
|
- libraries/exceptions/exceptions.cabal: Dependency of ``ghc`` and ``haskeline`` library
|
|
269
|
|
- libraries/filepath/filepath.cabal: Dependency of ``ghc`` library
|
|
270
|
|
- compiler/ghc.cabal: The compiler itself
|
|
271
|
|
- libraries/ghci/ghci.cabal: The REPL interface
|
|
272
|
|
- libraries/ghc-boot/ghc-boot.cabal: Internal compiler library
|
|
273
|
|
- libraries/ghc-boot-th/ghc-boot-th.cabal: Internal compiler library
|
|
274
|
|
- libraries/ghc-compact/ghc-compact.cabal: Core library
|
|
275
|
|
- libraries/ghc-heap/ghc-heap.cabal: GHC heap-walking library
|
|
276
|
|
- libraries/ghc-prim/ghc-prim.cabal: Core library
|
|
277
|
|
- utils/haddock/haddock-api/haddock-api.cabal: Dependency of ``haddock`` executable
|
|
278
|
|
- utils/haddock/haddock-library/haddock-library.cabal: Dependency of ``haddock`` executable
|
|
279
|
|
- libraries/haskeline/haskeline.cabal: Dependency of ``ghci`` executable
|
|
280
|
|
- libraries/hpc/hpc.cabal: Dependency of ``hpc`` executable
|
|
281
|
|
- libraries/integer-gmp/integer-gmp.cabal: Core library
|
|
282
|
|
- libraries/mtl/mtl.cabal: Dependency of ``Cabal`` library
|
|
283
|
|
- libraries/parsec/parsec.cabal: Dependency of ``Cabal`` library
|
|
284
|
|
- libraries/pretty/pretty.cabal: Dependency of ``ghc`` library
|
|
285
|
|
- libraries/process/process.cabal: Dependency of ``ghc`` library
|
|
286
|
|
- libraries/stm/stm.cabal: Dependency of ``haskeline`` library
|
|
287
|
|
- libraries/template-haskell/template-haskell.cabal: Core library
|
|
288
|
|
- libraries/terminfo/terminfo.cabal: Dependency of ``haskeline`` library
|
|
289
|
|
- libraries/text/text.cabal: Dependency of ``Cabal`` library
|
|
290
|
|
- libraries/time/time.cabal: Dependency of ``ghc`` library
|
|
291
|
|
- libraries/transformers/transformers.cabal: Dependency of ``ghc`` library
|
|
292
|
|
- libraries/unix/unix.cabal: Dependency of ``ghc`` library
|
|
293
|
|
- libraries/Win32/Win32.cabal: Dependency of ``ghc`` library
|
|
294
|
|
- libraries/xhtml/xhtml.cabal: Dependency of ``haddock`` executable
|
|
295
|
|
- libraries/os-string/os-string.cabal: Dependency of ``filepath`` library
|
|
296
|
|
- libraries/file-io/file-io.cabal: Dependency of ``directory`` library |