2 -- ---------------------------------------------------------------------------
3 -- (c) The University of Glasgow 1997-2003
7 -- Author(s): Simon Marlow, Sven Panne 1997, 1998, 1999
8 -- ---------------------------------------------------------------------------
11 {-# LANGUAGE BangPatterns #-} -- required for versions of Happy before 1.18.6
12 {-# OPTIONS -Wwarn -w -XNoMonomorphismRestriction #-}
13 -- The NoMonomorphismRestriction deals with a Happy infelicity
14 -- With OutsideIn's more conservativ monomorphism restriction
15 -- we aren't generalising
16 -- notHappyAtAll = error "urk"
17 -- which is terrible. Switching off the restriction allows
18 -- the generalisation. Better would be to make Happy generate
19 -- an appropriate signature.
21 -- The above warning supression flag is a temporary kludge.
22 -- While working on this module you are encouraged to remove it and fix
23 -- any warnings in the module. See
24 -- http://hackage.haskell.org/trac/ghc/wiki/Commentary/CodingStyle#Warnings
27 {-# OPTIONS_GHC -O0 -fno-ignore-interface-pragmas #-}
29 Careful optimisation of the parser: we don't want to throw everything
30 at it, because that takes too long and doesn't buy much, but we do want
31 to inline certain key external functions, so we instruct GHC not to
32 throw away inlinings as it would normally do in -O0 mode.
35 module Parser ( parseModule, parseStmt, parseIdentifier, parseType,
40 import HscTypes ( IsBootInterface, WarningTxt(..) )
43 import TysWiredIn ( unitTyCon, unitDataCon, tupleTyCon, tupleCon, nilDataCon,
44 unboxedSingletonTyCon, unboxedSingletonDataCon,
45 listTyCon_RDR, parrTyCon_RDR, consDataCon_RDR )
46 import Type ( funTyCon )
47 import ForeignCall ( Safety(..), CExportSpec(..), CLabelString,
48 CCallConv(..), CCallTarget(..), defaultCCallConv
50 import OccName ( varName, dataName, tcClsName, tvName )
51 import DataCon ( DataCon, dataConName )
52 import SrcLoc ( Located(..), unLoc, getLoc, noLoc, combineSrcSpans,
53 SrcSpan, combineLocs, srcLocFile,
56 import StaticFlags ( opt_SccProfilingOn, opt_Hpc )
57 import Type ( Kind, liftedTypeKind, unliftedTypeKind )
58 import Coercion ( mkArrowKind )
59 import Class ( FunDep )
66 import Maybes ( orElse )
69 import Control.Monad ( unless )
72 import Control.Monad ( mplus )
76 -----------------------------------------------------------------------------
79 Conflicts: 33 shift/reduce
82 The reduce/reduce conflict is weird. It's between tyconsym and consym, and I
83 would think the two should never occur in the same context.
87 -----------------------------------------------------------------------------
90 Conflicts: 34 shift/reduce
93 The reduce/reduce conflict is weird. It's between tyconsym and consym, and I
94 would think the two should never occur in the same context.
98 -----------------------------------------------------------------------------
101 Conflicts: 32 shift/reduce
104 The reduce/reduce conflict is weird. It's between tyconsym and consym, and I
105 would think the two should never occur in the same context.
109 -----------------------------------------------------------------------------
112 Conflicts: 37 shift/reduce
115 The reduce/reduce conflict is weird. It's between tyconsym and consym, and I
116 would think the two should never occur in the same context.
120 -----------------------------------------------------------------------------
121 Conflicts: 38 shift/reduce (1.25)
123 10 for abiguity in 'if x then y else z + 1' [State 178]
124 (shift parses as 'if x then y else (z + 1)', as per longest-parse rule)
125 10 because op might be: : - ! * . `x` VARSYM CONSYM QVARSYM QCONSYM
127 1 for ambiguity in 'if x then y else z :: T' [State 178]
128 (shift parses as 'if x then y else (z :: T)', as per longest-parse rule)
130 4 for ambiguity in 'if x then y else z -< e' [State 178]
131 (shift parses as 'if x then y else (z -< T)', as per longest-parse rule)
132 There are four such operators: -<, >-, -<<, >>-
135 2 for ambiguity in 'case v of { x :: T -> T ... } ' [States 11, 253]
136 Which of these two is intended?
138 (x::T) -> T -- Rhs is T
141 (x::T -> T) -> .. -- Rhs is ...
143 10 for ambiguity in 'e :: a `b` c'. Does this mean [States 11, 253]
146 As well as `b` we can have !, VARSYM, QCONSYM, and CONSYM, hence 5 cases
147 Same duplication between states 11 and 253 as the previous case
149 1 for ambiguity in 'let ?x ...' [State 329]
150 the parser can't tell whether the ?x is the lhs of a normal binding or
151 an implicit binding. Fortunately resolving as shift gives it the only
152 sensible meaning, namely the lhs of an implicit binding.
154 1 for ambiguity in '{-# RULES "name" [ ... #-} [State 382]
155 we don't know whether the '[' starts the activation or not: it
156 might be the start of the declaration with the activation being
157 empty. --SDM 1/4/2002
159 1 for ambiguity in '{-# RULES "name" forall = ... #-}' [State 474]
160 since 'forall' is a valid variable name, we don't know whether
161 to treat a forall on the input as the beginning of a quantifier
162 or the beginning of the rule itself. Resolving to shift means
163 it's always treated as a quantifier, hence the above is disallowed.
164 This saves explicitly defining a grammar for the rule lhs that
165 doesn't include 'forall'.
167 1 for ambiguity when the source file starts with "-- | doc". We need another
168 token of lookahead to determine if a top declaration or the 'module' keyword
169 follows. Shift parses as if the 'module' keyword follows.
171 -- ---------------------------------------------------------------------------
172 -- Adding location info
174 This is done in a stylised way using the three macros below, L0, L1
175 and LL. Each of these macros can be thought of as having type
177 L0, L1, LL :: a -> Located a
179 They each add a SrcSpan to their argument.
181 L0 adds 'noSrcSpan', used for empty productions
182 -- This doesn't seem to work anymore -=chak
184 L1 for a production with a single token on the lhs. Grabs the SrcSpan
187 LL for a production with >1 token on the lhs. Makes up a SrcSpan from
188 the first and last tokens.
190 These suffice for the majority of cases. However, we must be
191 especially careful with empty productions: LL won't work if the first
192 or last token on the lhs can represent an empty span. In these cases,
193 we have to calculate the span using more of the tokens from the lhs, eg.
195 | 'newtype' tycl_hdr '=' newconstr deriving
197 (mkTyData NewType (unLoc $2) [$4] (unLoc $5)) }
199 We provide comb3 and comb4 functions which are useful in such cases.
201 Be careful: there's no checking that you actually got this right, the
202 only symptom will be that the SrcSpans of your syntax will be
206 * We must expand these macros *before* running Happy, which is why this file is
207 * Parser.y.pp rather than just Parser.y - we run the C pre-processor first.
209 #define L0 L noSrcSpan
210 #define L1 sL (getLoc $1)
211 #define LL sL (comb2 $1 $>)
213 -- -----------------------------------------------------------------------------
218 '_' { L _ ITunderscore } -- Haskell keywords
220 'case' { L _ ITcase }
221 'class' { L _ ITclass }
222 'data' { L _ ITdata }
223 'default' { L _ ITdefault }
224 'deriving' { L _ ITderiving }
226 'else' { L _ ITelse }
227 'hiding' { L _ IThiding }
229 'import' { L _ ITimport }
231 'infix' { L _ ITinfix }
232 'infixl' { L _ ITinfixl }
233 'infixr' { L _ ITinfixr }
234 'instance' { L _ ITinstance }
236 'module' { L _ ITmodule }
237 'newtype' { L _ ITnewtype }
239 'qualified' { L _ ITqualified }
240 'then' { L _ ITthen }
241 'type' { L _ ITtype }
242 'where' { L _ ITwhere }
243 '_scc_' { L _ ITscc } -- ToDo: remove
245 'forall' { L _ ITforall } -- GHC extension keywords
246 'foreign' { L _ ITforeign }
247 'export' { L _ ITexport }
248 'label' { L _ ITlabel }
249 'dynamic' { L _ ITdynamic }
250 'safe' { L _ ITsafe }
251 'threadsafe' { L _ ITthreadsafe } -- ToDo: remove deprecated alias
252 'interruptible' { L _ ITinterruptible }
253 'unsafe' { L _ ITunsafe }
255 'family' { L _ ITfamily }
256 'stdcall' { L _ ITstdcallconv }
257 'ccall' { L _ ITccallconv }
258 'prim' { L _ ITprimcallconv }
259 'proc' { L _ ITproc } -- for arrow notation extension
260 'rec' { L _ ITrec } -- for arrow notation extension
261 'group' { L _ ITgroup } -- for list transform extension
262 'by' { L _ ITby } -- for list transform extension
263 'using' { L _ ITusing } -- for list transform extension
265 '{-# INLINE' { L _ (ITinline_prag _ _) }
266 '{-# SPECIALISE' { L _ ITspec_prag }
267 '{-# SPECIALISE_INLINE' { L _ (ITspec_inline_prag _) }
268 '{-# SOURCE' { L _ ITsource_prag }
269 '{-# RULES' { L _ ITrules_prag }
270 '{-# CORE' { L _ ITcore_prag } -- hdaume: annotated core
271 '{-# SCC' { L _ ITscc_prag }
272 '{-# GENERATED' { L _ ITgenerated_prag }
273 '{-# DEPRECATED' { L _ ITdeprecated_prag }
274 '{-# WARNING' { L _ ITwarning_prag }
275 '{-# UNPACK' { L _ ITunpack_prag }
276 '{-# ANN' { L _ ITann_prag }
277 '#-}' { L _ ITclose_prag }
279 '..' { L _ ITdotdot } -- reserved symbols
281 '::' { L _ ITdcolon }
285 '<-' { L _ ITlarrow }
286 '->' { L _ ITrarrow }
289 '=>' { L _ ITdarrow }
293 '-<' { L _ ITlarrowtail } -- for arrow notation
294 '>-' { L _ ITrarrowtail } -- for arrow notation
295 '-<<' { L _ ITLarrowtail } -- for arrow notation
296 '>>-' { L _ ITRarrowtail } -- for arrow notation
299 '{' { L _ ITocurly } -- special symbols
301 '{|' { L _ ITocurlybar }
302 '|}' { L _ ITccurlybar }
303 vocurly { L _ ITvocurly } -- virtual open curly (from layout)
304 vccurly { L _ ITvccurly } -- virtual close curly (from layout)
307 '[:' { L _ ITopabrack }
308 ':]' { L _ ITcpabrack }
311 '(#' { L _ IToubxparen }
312 '#)' { L _ ITcubxparen }
313 '(|' { L _ IToparenbar }
314 '|)' { L _ ITcparenbar }
317 '`' { L _ ITbackquote }
319 VARID { L _ (ITvarid _) } -- identifiers
320 CONID { L _ (ITconid _) }
321 VARSYM { L _ (ITvarsym _) }
322 CONSYM { L _ (ITconsym _) }
323 QVARID { L _ (ITqvarid _) }
324 QCONID { L _ (ITqconid _) }
325 QVARSYM { L _ (ITqvarsym _) }
326 QCONSYM { L _ (ITqconsym _) }
327 PREFIXQVARSYM { L _ (ITprefixqvarsym _) }
328 PREFIXQCONSYM { L _ (ITprefixqconsym _) }
330 IPDUPVARID { L _ (ITdupipvarid _) } -- GHC extension
332 CHAR { L _ (ITchar _) }
333 STRING { L _ (ITstring _) }
334 INTEGER { L _ (ITinteger _) }
335 RATIONAL { L _ (ITrational _) }
337 PRIMCHAR { L _ (ITprimchar _) }
338 PRIMSTRING { L _ (ITprimstring _) }
339 PRIMINTEGER { L _ (ITprimint _) }
340 PRIMWORD { L _ (ITprimword _) }
341 PRIMFLOAT { L _ (ITprimfloat _) }
342 PRIMDOUBLE { L _ (ITprimdouble _) }
344 DOCNEXT { L _ (ITdocCommentNext _) }
345 DOCPREV { L _ (ITdocCommentPrev _) }
346 DOCNAMED { L _ (ITdocCommentNamed _) }
347 DOCSECTION { L _ (ITdocSection _ _) }
350 '[|' { L _ ITopenExpQuote }
351 '[p|' { L _ ITopenPatQuote }
352 '[t|' { L _ ITopenTypQuote }
353 '[d|' { L _ ITopenDecQuote }
354 '|]' { L _ ITcloseQuote }
355 TH_ID_SPLICE { L _ (ITidEscape _) } -- $x
356 '$(' { L _ ITparenEscape } -- $( exp )
357 TH_VAR_QUOTE { L _ ITvarQuote } -- 'x
358 TH_TY_QUOTE { L _ ITtyQuote } -- ''T
359 TH_QUASIQUOTE { L _ (ITquasiQuote _) }
361 %monad { P } { >>= } { return }
362 %lexer { lexer } { L _ ITeof }
363 %name parseModule module
364 %name parseStmt maybe_stmt
365 %name parseIdentifier identifier
366 %name parseType ctype
367 %partial parseHeader header
368 %tokentype { (Located Token) }
371 -----------------------------------------------------------------------------
372 -- Identifiers; one of the entry points
373 identifier :: { Located RdrName }
378 | '(' '->' ')' { LL $ getRdrName funTyCon }
380 -----------------------------------------------------------------------------
383 -- The place for module deprecation is really too restrictive, but if it
384 -- was allowed at its natural place just before 'module', we get an ugly
385 -- s/r conflict with the second alternative. Another solution would be the
386 -- introduction of a new pragma DEPRECATED_MODULE, but this is not very nice,
387 -- either, and DEPRECATED is only expected to be used by people who really
388 -- know what they are doing. :-)
390 module :: { Located (HsModule RdrName) }
391 : maybedocheader 'module' modid maybemodwarning maybeexports 'where' body
392 {% fileSrcSpan >>= \ loc ->
393 return (L loc (HsModule (Just $3) $5 (fst $7) (snd $7) $4 $1
396 {% fileSrcSpan >>= \ loc ->
397 return (L loc (HsModule Nothing Nothing
398 (fst $1) (snd $1) Nothing Nothing
401 maybedocheader :: { Maybe LHsDocString }
402 : moduleheader { $1 }
403 | {- empty -} { Nothing }
405 missing_module_keyword :: { () }
406 : {- empty -} {% pushCurrentContext }
408 maybemodwarning :: { Maybe WarningTxt }
409 : '{-# DEPRECATED' strings '#-}' { Just (DeprecatedTxt $ unLoc $2) }
410 | '{-# WARNING' strings '#-}' { Just (WarningTxt $ unLoc $2) }
411 | {- empty -} { Nothing }
413 body :: { ([LImportDecl RdrName], [LHsDecl RdrName]) }
415 | vocurly top close { $2 }
417 body2 :: { ([LImportDecl RdrName], [LHsDecl RdrName]) }
419 | missing_module_keyword top close { $2 }
421 top :: { ([LImportDecl RdrName], [LHsDecl RdrName]) }
422 : importdecls { (reverse $1,[]) }
423 | importdecls ';' cvtopdecls { (reverse $1,$3) }
424 | cvtopdecls { ([],$1) }
426 cvtopdecls :: { [LHsDecl RdrName] }
427 : topdecls { cvTopDecls $1 }
429 -----------------------------------------------------------------------------
430 -- Module declaration & imports only
432 header :: { Located (HsModule RdrName) }
433 : maybedocheader 'module' modid maybemodwarning maybeexports 'where' header_body
434 {% fileSrcSpan >>= \ loc ->
435 return (L loc (HsModule (Just $3) $5 $7 [] $4 $1
437 | missing_module_keyword importdecls
438 {% fileSrcSpan >>= \ loc ->
439 return (L loc (HsModule Nothing Nothing $2 [] Nothing
442 header_body :: { [LImportDecl RdrName] }
443 : '{' importdecls { $2 }
444 | vocurly importdecls { $2 }
446 -----------------------------------------------------------------------------
449 maybeexports :: { Maybe [LIE RdrName] }
450 : '(' exportlist ')' { Just $2 }
451 | {- empty -} { Nothing }
453 exportlist :: { [LIE RdrName] }
454 : expdoclist ',' expdoclist { $1 ++ $3 }
457 exportlist1 :: { [LIE RdrName] }
458 : expdoclist export expdoclist ',' exportlist { $1 ++ ($2 : $3) ++ $5 }
459 | expdoclist export expdoclist { $1 ++ ($2 : $3) }
462 expdoclist :: { [LIE RdrName] }
463 : exp_doc expdoclist { $1 : $2 }
466 exp_doc :: { LIE RdrName }
467 : docsection { L1 (case (unLoc $1) of (n, doc) -> IEGroup n doc) }
468 | docnamed { L1 (IEDocNamed ((fst . unLoc) $1)) }
469 | docnext { L1 (IEDoc (unLoc $1)) }
471 -- No longer allow things like [] and (,,,) to be exported
472 -- They are built in syntax, always available
473 export :: { LIE RdrName }
474 : qvar { L1 (IEVar (unLoc $1)) }
475 | oqtycon { L1 (IEThingAbs (unLoc $1)) }
476 | oqtycon '(' '..' ')' { LL (IEThingAll (unLoc $1)) }
477 | oqtycon '(' ')' { LL (IEThingWith (unLoc $1) []) }
478 | oqtycon '(' qcnames ')' { LL (IEThingWith (unLoc $1) (reverse $3)) }
479 | 'module' modid { LL (IEModuleContents (unLoc $2)) }
481 qcnames :: { [RdrName] }
482 : qcnames ',' qcname_ext { unLoc $3 : $1 }
483 | qcname_ext { [unLoc $1] }
485 qcname_ext :: { Located RdrName } -- Variable or data constructor
486 -- or tagged type constructor
488 | 'type' qcon { sL (comb2 $1 $2)
489 (setRdrNameSpace (unLoc $2)
492 -- Cannot pull into qcname_ext, as qcname is also used in expression.
493 qcname :: { Located RdrName } -- Variable or data constructor
497 -----------------------------------------------------------------------------
498 -- Import Declarations
500 -- import decls can be *empty*, or even just a string of semicolons
501 -- whereas topdecls must contain at least one topdecl.
503 importdecls :: { [LImportDecl RdrName] }
504 : importdecls ';' importdecl { $3 : $1 }
505 | importdecls ';' { $1 }
506 | importdecl { [ $1 ] }
509 importdecl :: { LImportDecl RdrName }
510 : 'import' maybe_src optqualified maybe_pkg modid maybeas maybeimpspec
511 { L (comb4 $1 $5 $6 $7) (ImportDecl $5 $4 $2 $3 (unLoc $6) (unLoc $7)) }
513 maybe_src :: { IsBootInterface }
514 : '{-# SOURCE' '#-}' { True }
515 | {- empty -} { False }
517 maybe_pkg :: { Maybe FastString }
518 : STRING { Just (getSTRING $1) }
519 | {- empty -} { Nothing }
521 optqualified :: { Bool }
522 : 'qualified' { True }
523 | {- empty -} { False }
525 maybeas :: { Located (Maybe ModuleName) }
526 : 'as' modid { LL (Just (unLoc $2)) }
527 | {- empty -} { noLoc Nothing }
529 maybeimpspec :: { Located (Maybe (Bool, [LIE RdrName])) }
530 : impspec { L1 (Just (unLoc $1)) }
531 | {- empty -} { noLoc Nothing }
533 impspec :: { Located (Bool, [LIE RdrName]) }
534 : '(' exportlist ')' { LL (False, $2) }
535 | 'hiding' '(' exportlist ')' { LL (True, $3) }
537 -----------------------------------------------------------------------------
538 -- Fixity Declarations
542 | INTEGER {% checkPrecP (L1 (fromInteger (getINTEGER $1))) }
544 infix :: { Located FixityDirection }
545 : 'infix' { L1 InfixN }
546 | 'infixl' { L1 InfixL }
547 | 'infixr' { L1 InfixR }
549 ops :: { Located [Located RdrName] }
550 : ops ',' op { LL ($3 : unLoc $1) }
553 -----------------------------------------------------------------------------
554 -- Top-Level Declarations
556 topdecls :: { OrdList (LHsDecl RdrName) }
557 : topdecls ';' topdecl { $1 `appOL` $3 }
558 | topdecls ';' { $1 }
561 topdecl :: { OrdList (LHsDecl RdrName) }
562 : cl_decl { unitOL (L1 (TyClD (unLoc $1))) }
563 | ty_decl { unitOL (L1 (TyClD (unLoc $1))) }
564 | 'instance' inst_type where_inst
565 { let (binds, sigs, ats, _) = cvBindsAndSigs (unLoc $3)
567 unitOL (L (comb3 $1 $2 $3) (InstD (InstDecl $2 binds sigs ats)))}
568 | stand_alone_deriving { unitOL (LL (DerivD (unLoc $1))) }
569 | 'default' '(' comma_types0 ')' { unitOL (LL $ DefD (DefaultDecl $3)) }
570 | 'foreign' fdecl { unitOL (LL (unLoc $2)) }
571 | '{-# DEPRECATED' deprecations '#-}' { $2 }
572 | '{-# WARNING' warnings '#-}' { $2 }
573 | '{-# RULES' rules '#-}' { $2 }
574 | annotation { unitOL $1 }
577 -- Template Haskell Extension
578 -- The $(..) form is one possible form of infixexp
579 -- but we treat an arbitrary expression just as if
580 -- it had a $(..) wrapped around it
581 | infixexp { unitOL (LL $ mkTopSpliceDecl $1) }
585 cl_decl :: { LTyClDecl RdrName }
586 : 'class' tycl_hdr fds where_cls {% mkClassDecl (comb4 $1 $2 $3 $4) $2 $3 $4 }
588 -- Type declarations (toplevel)
590 ty_decl :: { LTyClDecl RdrName }
591 -- ordinary type synonyms
592 : 'type' type '=' ctypedoc
593 -- Note ctype, not sigtype, on the right of '='
594 -- We allow an explicit for-all but we don't insert one
595 -- in type Foo a = (b,b)
596 -- Instead we just say b is out of scope
598 -- Note the use of type for the head; this allows
599 -- infix type constructors to be declared
600 {% mkTySynonym (comb2 $1 $4) False $2 $4 }
602 -- type family declarations
603 | 'type' 'family' type opt_kind_sig
604 -- Note the use of type for the head; this allows
605 -- infix type constructors to be declared
606 {% mkTyFamily (comb3 $1 $3 $4) TypeFamily $3 (unLoc $4) }
608 -- type instance declarations
609 | 'type' 'instance' type '=' ctype
610 -- Note the use of type for the head; this allows
611 -- infix type constructors and type patterns
612 {% mkTySynonym (comb2 $1 $5) True $3 $5 }
614 -- ordinary data type or newtype declaration
615 | data_or_newtype tycl_hdr constrs deriving
616 {% mkTyData (comb4 $1 $2 $3 $4) (unLoc $1) False $2
617 Nothing (reverse (unLoc $3)) (unLoc $4) }
618 -- We need the location on tycl_hdr in case
619 -- constrs and deriving are both empty
621 -- ordinary GADT declaration
622 | data_or_newtype tycl_hdr opt_kind_sig
625 {% mkTyData (comb4 $1 $2 $4 $5) (unLoc $1) False $2
626 (unLoc $3) (unLoc $4) (unLoc $5) }
627 -- We need the location on tycl_hdr in case
628 -- constrs and deriving are both empty
630 -- data/newtype family
631 | 'data' 'family' type opt_kind_sig
632 {% mkTyFamily (comb3 $1 $2 $4) DataFamily $3 (unLoc $4) }
634 -- data/newtype instance declaration
635 | data_or_newtype 'instance' tycl_hdr constrs deriving
636 {% mkTyData (comb4 $1 $3 $4 $5) (unLoc $1) True $3
637 Nothing (reverse (unLoc $4)) (unLoc $5) }
639 -- GADT instance declaration
640 | data_or_newtype 'instance' tycl_hdr opt_kind_sig
643 {% mkTyData (comb4 $1 $3 $5 $6) (unLoc $1) True $3
644 (unLoc $4) (unLoc $5) (unLoc $6) }
646 -- Associated type family declarations
648 -- * They have a different syntax than on the toplevel (no family special
651 -- * They also need to be separate from instances; otherwise, data family
652 -- declarations without a kind signature cause parsing conflicts with empty
653 -- data declarations.
655 at_decl_cls :: { LTyClDecl RdrName }
656 -- type family declarations
657 : 'type' type opt_kind_sig
658 -- Note the use of type for the head; this allows
659 -- infix type constructors to be declared
660 {% mkTyFamily (comb3 $1 $2 $3) TypeFamily $2 (unLoc $3) }
662 -- default type instance
663 | 'type' type '=' ctype
664 -- Note the use of type for the head; this allows
665 -- infix type constructors and type patterns
666 {% mkTySynonym (comb2 $1 $4) True $2 $4 }
668 -- data/newtype family declaration
669 | 'data' type opt_kind_sig
670 {% mkTyFamily (comb3 $1 $2 $3) DataFamily $2 (unLoc $3) }
672 -- Associated type instances
674 at_decl_inst :: { LTyClDecl RdrName }
675 -- type instance declarations
676 : 'type' type '=' ctype
677 -- Note the use of type for the head; this allows
678 -- infix type constructors and type patterns
679 {% mkTySynonym (comb2 $1 $4) True $2 $4 }
681 -- data/newtype instance declaration
682 | data_or_newtype tycl_hdr constrs deriving
683 {% mkTyData (comb4 $1 $2 $3 $4) (unLoc $1) True $2
684 Nothing (reverse (unLoc $3)) (unLoc $4) }
686 -- GADT instance declaration
687 | data_or_newtype tycl_hdr opt_kind_sig
690 {% mkTyData (comb4 $1 $2 $4 $5) (unLoc $1) True $2
691 (unLoc $3) (unLoc $4) (unLoc $5) }
693 data_or_newtype :: { Located NewOrData }
694 : 'data' { L1 DataType }
695 | 'newtype' { L1 NewType }
697 opt_kind_sig :: { Located (Maybe Kind) }
699 | '::' kind { LL (Just (unLoc $2)) }
701 -- tycl_hdr parses the header of a class or data type decl,
702 -- which takes the form
705 -- (Eq a, Ord b) => T a b
706 -- T Int [a] -- for associated types
707 -- Rather a lot of inlining here, else we get reduce/reduce errors
708 tycl_hdr :: { Located (Maybe (LHsContext RdrName), LHsType RdrName) }
709 : context '=>' type { LL (Just $1, $3) }
710 | type { L1 (Nothing, $1) }
712 -----------------------------------------------------------------------------
713 -- Stand-alone deriving
715 -- Glasgow extension: stand-alone deriving declarations
716 stand_alone_deriving :: { LDerivDecl RdrName }
717 : 'deriving' 'instance' inst_type { LL (DerivDecl $3) }
719 -----------------------------------------------------------------------------
720 -- Nested declarations
722 -- Declaration in class bodies
724 decl_cls :: { Located (OrdList (LHsDecl RdrName)) }
725 decl_cls : at_decl_cls { LL (unitOL (L1 (TyClD (unLoc $1)))) }
728 decls_cls :: { Located (OrdList (LHsDecl RdrName)) } -- Reversed
729 : decls_cls ';' decl_cls { LL (unLoc $1 `appOL` unLoc $3) }
730 | decls_cls ';' { LL (unLoc $1) }
732 | {- empty -} { noLoc nilOL }
736 :: { Located (OrdList (LHsDecl RdrName)) } -- Reversed
737 : '{' decls_cls '}' { LL (unLoc $2) }
738 | vocurly decls_cls close { $2 }
742 where_cls :: { Located (OrdList (LHsDecl RdrName)) } -- Reversed
743 -- No implicit parameters
744 -- May have type declarations
745 : 'where' decllist_cls { LL (unLoc $2) }
746 | {- empty -} { noLoc nilOL }
748 -- Declarations in instance bodies
750 decl_inst :: { Located (OrdList (LHsDecl RdrName)) }
751 decl_inst : at_decl_inst { LL (unitOL (L1 (TyClD (unLoc $1)))) }
754 decls_inst :: { Located (OrdList (LHsDecl RdrName)) } -- Reversed
755 : decls_inst ';' decl_inst { LL (unLoc $1 `appOL` unLoc $3) }
756 | decls_inst ';' { LL (unLoc $1) }
758 | {- empty -} { noLoc nilOL }
761 :: { Located (OrdList (LHsDecl RdrName)) } -- Reversed
762 : '{' decls_inst '}' { LL (unLoc $2) }
763 | vocurly decls_inst close { $2 }
767 where_inst :: { Located (OrdList (LHsDecl RdrName)) } -- Reversed
768 -- No implicit parameters
769 -- May have type declarations
770 : 'where' decllist_inst { LL (unLoc $2) }
771 | {- empty -} { noLoc nilOL }
773 -- Declarations in binding groups other than classes and instances
775 decls :: { Located (OrdList (LHsDecl RdrName)) }
776 : decls ';' decl { let { this = unLoc $3;
778 these = rest `appOL` this }
779 in rest `seq` this `seq` these `seq`
781 | decls ';' { LL (unLoc $1) }
783 | {- empty -} { noLoc nilOL }
785 decllist :: { Located (OrdList (LHsDecl RdrName)) }
786 : '{' decls '}' { LL (unLoc $2) }
787 | vocurly decls close { $2 }
789 -- Binding groups other than those of class and instance declarations
791 binds :: { Located (HsLocalBinds RdrName) } -- May have implicit parameters
792 -- No type declarations
793 : decllist { L1 (HsValBinds (cvBindGroup (unLoc $1))) }
794 | '{' dbinds '}' { LL (HsIPBinds (IPBinds (unLoc $2) emptyTcEvBinds)) }
795 | vocurly dbinds close { L (getLoc $2) (HsIPBinds (IPBinds (unLoc $2) emptyTcEvBinds)) }
797 wherebinds :: { Located (HsLocalBinds RdrName) } -- May have implicit parameters
798 -- No type declarations
799 : 'where' binds { LL (unLoc $2) }
800 | {- empty -} { noLoc emptyLocalBinds }
803 -----------------------------------------------------------------------------
804 -- Transformation Rules
806 rules :: { OrdList (LHsDecl RdrName) }
807 : rules ';' rule { $1 `snocOL` $3 }
810 | {- empty -} { nilOL }
812 rule :: { LHsDecl RdrName }
813 : STRING activation rule_forall infixexp '=' exp
814 { LL $ RuleD (HsRule (getSTRING $1)
815 ($2 `orElse` AlwaysActive)
816 $3 $4 placeHolderNames $6 placeHolderNames) }
818 activation :: { Maybe Activation }
819 : {- empty -} { Nothing }
820 | explicit_activation { Just $1 }
822 explicit_activation :: { Activation } -- In brackets
823 : '[' INTEGER ']' { ActiveAfter (fromInteger (getINTEGER $2)) }
824 | '[' '~' INTEGER ']' { ActiveBefore (fromInteger (getINTEGER $3)) }
826 rule_forall :: { [RuleBndr RdrName] }
827 : 'forall' rule_var_list '.' { $2 }
830 rule_var_list :: { [RuleBndr RdrName] }
832 | rule_var rule_var_list { $1 : $2 }
834 rule_var :: { RuleBndr RdrName }
835 : varid { RuleBndr $1 }
836 | '(' varid '::' ctype ')' { RuleBndrSig $2 $4 }
838 -----------------------------------------------------------------------------
839 -- Warnings and deprecations (c.f. rules)
841 warnings :: { OrdList (LHsDecl RdrName) }
842 : warnings ';' warning { $1 `appOL` $3 }
843 | warnings ';' { $1 }
845 | {- empty -} { nilOL }
847 -- SUP: TEMPORARY HACK, not checking for `module Foo'
848 warning :: { OrdList (LHsDecl RdrName) }
850 { toOL [ LL $ WarningD (Warning n (WarningTxt $ unLoc $2))
853 deprecations :: { OrdList (LHsDecl RdrName) }
854 : deprecations ';' deprecation { $1 `appOL` $3 }
855 | deprecations ';' { $1 }
857 | {- empty -} { nilOL }
859 -- SUP: TEMPORARY HACK, not checking for `module Foo'
860 deprecation :: { OrdList (LHsDecl RdrName) }
862 { toOL [ LL $ WarningD (Warning n (DeprecatedTxt $ unLoc $2))
865 strings :: { Located [FastString] }
866 : STRING { L1 [getSTRING $1] }
867 | '[' stringlist ']' { LL $ fromOL (unLoc $2) }
869 stringlist :: { Located (OrdList FastString) }
870 : stringlist ',' STRING { LL (unLoc $1 `snocOL` getSTRING $3) }
871 | STRING { LL (unitOL (getSTRING $1)) }
873 -----------------------------------------------------------------------------
875 annotation :: { LHsDecl RdrName }
876 : '{-# ANN' name_var aexp '#-}' { LL (AnnD $ HsAnnotation (ValueAnnProvenance (unLoc $2)) $3) }
877 | '{-# ANN' 'type' tycon aexp '#-}' { LL (AnnD $ HsAnnotation (TypeAnnProvenance (unLoc $3)) $4) }
878 | '{-# ANN' 'module' aexp '#-}' { LL (AnnD $ HsAnnotation ModuleAnnProvenance $3) }
881 -----------------------------------------------------------------------------
882 -- Foreign import and export declarations
884 fdecl :: { LHsDecl RdrName }
885 fdecl : 'import' callconv safety fspec
886 {% mkImport $2 $3 (unLoc $4) >>= return.LL }
887 | 'import' callconv fspec
888 {% do { d <- mkImport $2 (PlaySafe False) (unLoc $3);
890 | 'export' callconv fspec
891 {% mkExport $2 (unLoc $3) >>= return.LL }
893 callconv :: { CCallConv }
894 : 'stdcall' { StdCallConv }
895 | 'ccall' { CCallConv }
896 | 'prim' { PrimCallConv}
899 : 'unsafe' { PlayRisky }
900 | 'safe' { PlaySafe False }
901 | 'interruptible' { PlayInterruptible }
902 | 'threadsafe' { PlaySafe True } -- deprecated alias
904 fspec :: { Located (Located FastString, Located RdrName, LHsType RdrName) }
905 : STRING var '::' sigtypedoc { LL (L (getLoc $1) (getSTRING $1), $2, $4) }
906 | var '::' sigtypedoc { LL (noLoc nilFS, $1, $3) }
907 -- if the entity string is missing, it defaults to the empty string;
908 -- the meaning of an empty entity string depends on the calling
911 -----------------------------------------------------------------------------
914 opt_sig :: { Maybe (LHsType RdrName) }
915 : {- empty -} { Nothing }
916 | '::' sigtype { Just $2 }
918 opt_asig :: { Maybe (LHsType RdrName) }
919 : {- empty -} { Nothing }
920 | '::' atype { Just $2 }
922 sigtype :: { LHsType RdrName } -- Always a HsForAllTy,
923 -- to tell the renamer where to generalise
924 : ctype { L1 (mkImplicitHsForAllTy (noLoc []) $1) }
925 -- Wrap an Implicit forall if there isn't one there already
927 sigtypedoc :: { LHsType RdrName } -- Always a HsForAllTy
928 : ctypedoc { L1 (mkImplicitHsForAllTy (noLoc []) $1) }
929 -- Wrap an Implicit forall if there isn't one there already
931 sig_vars :: { Located [Located RdrName] }
932 : sig_vars ',' var { LL ($3 : unLoc $1) }
935 sigtypes1 :: { [LHsType RdrName] } -- Always HsForAllTys
937 | sigtype ',' sigtypes1 { $1 : $3 }
939 -----------------------------------------------------------------------------
942 infixtype :: { LHsType RdrName }
943 : btype qtyconop type { LL $ HsOpTy $1 $2 $3 }
944 | btype tyvarop type { LL $ HsOpTy $1 $2 $3 }
946 strict_mark :: { Located HsBang }
947 : '!' { L1 HsStrict }
948 | '{-# UNPACK' '#-}' '!' { LL HsUnpack }
950 -- A ctype is a for-all type
951 ctype :: { LHsType RdrName }
952 : 'forall' tv_bndrs '.' ctype { LL $ mkExplicitHsForAllTy $2 (noLoc []) $4 }
953 | context '=>' ctype { LL $ mkImplicitHsForAllTy $1 $3 }
954 -- A type of form (context => type) is an *implicit* HsForAllTy
955 | ipvar '::' type { LL (HsPredTy (HsIParam (unLoc $1) $3)) }
958 ----------------------
959 -- Notes for 'ctypedoc'
960 -- It would have been nice to simplify the grammar by unifying `ctype` and
961 -- ctypedoc` into one production, allowing comments on types everywhere (and
962 -- rejecting them after parsing, where necessary). This is however not possible
963 -- since it leads to ambiguity. The reason is the support for comments on record
965 -- data R = R { field :: Int -- ^ comment on the field }
966 -- If we allow comments on types here, it's not clear if the comment applies
967 -- to 'field' or to 'Int'. So we must use `ctype` to describe the type.
969 ctypedoc :: { LHsType RdrName }
970 : 'forall' tv_bndrs '.' ctypedoc { LL $ mkExplicitHsForAllTy $2 (noLoc []) $4 }
971 | context '=>' ctypedoc { LL $ mkImplicitHsForAllTy $1 $3 }
972 -- A type of form (context => type) is an *implicit* HsForAllTy
973 | ipvar '::' type { LL (HsPredTy (HsIParam (unLoc $1) $3)) }
976 ----------------------
977 -- Notes for 'context'
978 -- We parse a context as a btype so that we don't get reduce/reduce
979 -- errors in ctype. The basic problem is that
981 -- looks so much like a tuple type. We can't tell until we find the =>
983 -- We have the t1 ~ t2 form both in 'context' and in type,
984 -- to permit an individual equational constraint without parenthesis.
985 -- Thus for some reason we allow f :: a~b => blah
986 -- but not f :: ?x::Int => blah
987 context :: { LHsContext RdrName }
988 : btype '~' btype {% checkContext
989 (LL $ HsPredTy (HsEqualP $1 $3)) }
990 | btype {% checkContext $1 }
992 type :: { LHsType RdrName }
994 | btype qtyconop type { LL $ HsOpTy $1 $2 $3 }
995 | btype tyvarop type { LL $ HsOpTy $1 $2 $3 }
996 | btype '->' ctype { LL $ HsFunTy $1 $3 }
997 | btype '~' btype { LL $ HsPredTy (HsEqualP $1 $3) }
999 typedoc :: { LHsType RdrName }
1001 | btype docprev { LL $ HsDocTy $1 $2 }
1002 | btype qtyconop type { LL $ HsOpTy $1 $2 $3 }
1003 | btype qtyconop type docprev { LL $ HsDocTy (L (comb3 $1 $2 $3) (HsOpTy $1 $2 $3)) $4 }
1004 | btype tyvarop type { LL $ HsOpTy $1 $2 $3 }
1005 | btype tyvarop type docprev { LL $ HsDocTy (L (comb3 $1 $2 $3) (HsOpTy $1 $2 $3)) $4 }
1006 | btype '->' ctypedoc { LL $ HsFunTy $1 $3 }
1007 | btype docprev '->' ctypedoc { LL $ HsFunTy (L (comb2 $1 $2) (HsDocTy $1 $2)) $4 }
1008 | btype '~' btype { LL $ HsPredTy (HsEqualP $1 $3) }
1010 btype :: { LHsType RdrName }
1011 : btype atype { LL $ HsAppTy $1 $2 }
1014 atype :: { LHsType RdrName }
1015 : gtycon { L1 (HsTyVar (unLoc $1)) }
1016 | tyvar { L1 (HsTyVar (unLoc $1)) }
1017 | strict_mark atype { LL (HsBangTy (unLoc $1) $2) } -- Constructor sigs only
1018 | '{' fielddecls '}' { LL $ HsRecTy $2 } -- Constructor sigs only
1019 | '(' ctype ',' comma_types1 ')' { LL $ HsTupleTy Boxed ($2:$4) }
1020 | '(#' comma_types1 '#)' { LL $ HsTupleTy Unboxed $2 }
1021 | '[' ctype ']' { LL $ HsListTy $2 }
1022 | '[:' ctype ':]' { LL $ HsPArrTy $2 }
1023 | '(' ctype ')' { LL $ HsParTy $2 }
1024 | '(' ctype '::' kind ')' { LL $ HsKindSig $2 (unLoc $4) }
1025 | quasiquote { L1 (HsQuasiQuoteTy (unLoc $1)) }
1026 | '$(' exp ')' { LL $ mkHsSpliceTy $2 }
1027 | TH_ID_SPLICE { LL $ mkHsSpliceTy $ L1 $ HsVar $
1028 mkUnqual varName (getTH_ID_SPLICE $1) }
1030 | INTEGER { L1 (HsNumTy (getINTEGER $1)) }
1032 -- An inst_type is what occurs in the head of an instance decl
1033 -- e.g. (Foo a, Gaz b) => Wibble a b
1034 -- It's kept as a single type, with a MonoDictTy at the right
1035 -- hand corner, for convenience.
1036 inst_type :: { LHsType RdrName }
1037 : sigtype {% checkInstType $1 }
1039 inst_types1 :: { [LHsType RdrName] }
1040 : inst_type { [$1] }
1041 | inst_type ',' inst_types1 { $1 : $3 }
1043 comma_types0 :: { [LHsType RdrName] }
1044 : comma_types1 { $1 }
1045 | {- empty -} { [] }
1047 comma_types1 :: { [LHsType RdrName] }
1049 | ctype ',' comma_types1 { $1 : $3 }
1051 tv_bndrs :: { [LHsTyVarBndr RdrName] }
1052 : tv_bndr tv_bndrs { $1 : $2 }
1053 | {- empty -} { [] }
1055 tv_bndr :: { LHsTyVarBndr RdrName }
1056 : tyvar { L1 (UserTyVar (unLoc $1) placeHolderKind) }
1057 | '(' tyvar '::' kind ')' { LL (KindedTyVar (unLoc $2)
1060 fds :: { Located [Located (FunDep RdrName)] }
1061 : {- empty -} { noLoc [] }
1062 | '|' fds1 { LL (reverse (unLoc $2)) }
1064 fds1 :: { Located [Located (FunDep RdrName)] }
1065 : fds1 ',' fd { LL ($3 : unLoc $1) }
1068 fd :: { Located (FunDep RdrName) }
1069 : varids0 '->' varids0 { L (comb3 $1 $2 $3)
1070 (reverse (unLoc $1), reverse (unLoc $3)) }
1072 varids0 :: { Located [RdrName] }
1073 : {- empty -} { noLoc [] }
1074 | varids0 tyvar { LL (unLoc $2 : unLoc $1) }
1076 -----------------------------------------------------------------------------
1079 kind :: { Located Kind }
1081 | akind '->' kind { LL (mkArrowKind (unLoc $1) (unLoc $3)) }
1083 akind :: { Located Kind }
1084 : '*' { L1 liftedTypeKind }
1085 | '!' { L1 unliftedTypeKind }
1086 | '(' kind ')' { LL (unLoc $2) }
1089 -----------------------------------------------------------------------------
1090 -- Datatype declarations
1092 gadt_constrlist :: { Located [LConDecl RdrName] } -- Returned in order
1093 : 'where' '{' gadt_constrs '}' { L (comb2 $1 $3) (unLoc $3) }
1094 | 'where' vocurly gadt_constrs close { L (comb2 $1 $3) (unLoc $3) }
1095 | {- empty -} { noLoc [] }
1097 gadt_constrs :: { Located [LConDecl RdrName] }
1098 : gadt_constr ';' gadt_constrs { L (comb2 (head $1) $3) ($1 ++ unLoc $3) }
1099 | gadt_constr { L (getLoc (head $1)) $1 }
1100 | {- empty -} { noLoc [] }
1102 -- We allow the following forms:
1103 -- C :: Eq a => a -> T a
1104 -- C :: forall a. Eq a => !a -> T a
1105 -- D { x,y :: a } :: T a
1106 -- forall a. Eq a => D { x,y :: a } :: T a
1108 gadt_constr :: { [LConDecl RdrName] } -- Returns a list because of: C,D :: ty
1109 : con_list '::' sigtype
1110 { map (sL (comb2 $1 $3)) (mkGadtDecl (unLoc $1) $3) }
1112 -- Deprecated syntax for GADT record declarations
1113 | oqtycon '{' fielddecls '}' '::' sigtype
1114 {% do { cd <- mkDeprecatedGadtRecordDecl (comb2 $1 $6) $1 $3 $6
1117 constrs :: { Located [LConDecl RdrName] }
1118 : maybe_docnext '=' constrs1 { L (comb2 $2 $3) (addConDocs (unLoc $3) $1) }
1120 constrs1 :: { Located [LConDecl RdrName] }
1121 : constrs1 maybe_docnext '|' maybe_docprev constr { LL (addConDoc $5 $2 : addConDocFirst (unLoc $1) $4) }
1122 | constr { L1 [$1] }
1124 constr :: { LConDecl RdrName }
1125 : maybe_docnext forall context '=>' constr_stuff maybe_docprev
1126 { let (con,details) = unLoc $5 in
1127 addConDoc (L (comb4 $2 $3 $4 $5) (mkSimpleConDecl con (unLoc $2) $3 details))
1129 | maybe_docnext forall constr_stuff maybe_docprev
1130 { let (con,details) = unLoc $3 in
1131 addConDoc (L (comb2 $2 $3) (mkSimpleConDecl con (unLoc $2) (noLoc []) details))
1134 forall :: { Located [LHsTyVarBndr RdrName] }
1135 : 'forall' tv_bndrs '.' { LL $2 }
1136 | {- empty -} { noLoc [] }
1138 constr_stuff :: { Located (Located RdrName, HsConDeclDetails RdrName) }
1139 -- We parse the constructor declaration
1141 -- as a btype (treating C as a type constructor) and then convert C to be
1142 -- a data constructor. Reason: it might continue like this:
1144 -- in which case C really would be a type constructor. We can't resolve this
1145 -- ambiguity till we come across the constructor oprerator :% (or not, more usually)
1146 : btype {% splitCon $1 >>= return.LL }
1147 | btype conop btype { LL ($2, InfixCon $1 $3) }
1149 fielddecls :: { [ConDeclField RdrName] }
1150 : {- empty -} { [] }
1151 | fielddecls1 { $1 }
1153 fielddecls1 :: { [ConDeclField RdrName] }
1154 : fielddecl maybe_docnext ',' maybe_docprev fielddecls1
1155 { [ addFieldDoc f $4 | f <- $1 ] ++ addFieldDocs $5 $2 }
1156 -- This adds the doc $4 to each field separately
1159 fielddecl :: { [ConDeclField RdrName] } -- A list because of f,g :: Int
1160 : maybe_docnext sig_vars '::' ctype maybe_docprev { [ ConDeclField fld $4 ($1 `mplus` $5)
1161 | fld <- reverse (unLoc $2) ] }
1163 -- We allow the odd-looking 'inst_type' in a deriving clause, so that
1164 -- we can do deriving( forall a. C [a] ) in a newtype (GHC extension).
1165 -- The 'C [a]' part is converted to an HsPredTy by checkInstType
1166 -- We don't allow a context, but that's sorted out by the type checker.
1167 deriving :: { Located (Maybe [LHsType RdrName]) }
1168 : {- empty -} { noLoc Nothing }
1169 | 'deriving' qtycon {% do { let { L loc tv = $2 }
1170 ; p <- checkInstType (L loc (HsTyVar tv))
1171 ; return (LL (Just [p])) } }
1172 | 'deriving' '(' ')' { LL (Just []) }
1173 | 'deriving' '(' inst_types1 ')' { LL (Just $3) }
1174 -- Glasgow extension: allow partial
1175 -- applications in derivings
1177 -----------------------------------------------------------------------------
1178 -- Value definitions
1180 {- Note [Declaration/signature overlap]
1181 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1182 There's an awkward overlap with a type signature. Consider
1183 f :: Int -> Int = ...rhs...
1184 Then we can't tell whether it's a type signature or a value
1185 definition with a result signature until we see the '='.
1186 So we have to inline enough to postpone reductions until we know.
1190 ATTENTION: Dirty Hackery Ahead! If the second alternative of vars is var
1191 instead of qvar, we get another shift/reduce-conflict. Consider the
1194 { (^^) :: Int->Int ; } Type signature; only var allowed
1196 { (^^) :: Int->Int = ... ; } Value defn with result signature;
1197 qvar allowed (because of instance decls)
1199 We can't tell whether to reduce var to qvar until after we've read the signatures.
1202 docdecl :: { LHsDecl RdrName }
1203 : docdecld { L1 (DocD (unLoc $1)) }
1205 docdecld :: { LDocDecl }
1206 : docnext { L1 (DocCommentNext (unLoc $1)) }
1207 | docprev { L1 (DocCommentPrev (unLoc $1)) }
1208 | docnamed { L1 (case (unLoc $1) of (n, doc) -> DocCommentNamed n doc) }
1209 | docsection { L1 (case (unLoc $1) of (n, doc) -> DocGroup n doc) }
1211 decl :: { Located (OrdList (LHsDecl RdrName)) }
1214 | '!' aexp rhs {% do { let { e = LL (SectionR (LL (HsVar bang_RDR)) $2) };
1215 pat <- checkPattern e;
1216 return $ LL $ unitOL $ LL $ ValD $
1217 PatBind pat (unLoc $3)
1218 placeHolderType placeHolderNames } }
1219 -- Turn it all into an expression so that
1220 -- checkPattern can check that bangs are enabled
1222 | infixexp opt_sig rhs {% do { r <- checkValDef $1 $2 $3;
1223 let { l = comb2 $1 $> };
1224 return $! (sL l (unitOL $! (sL l $ ValD r))) } }
1225 | docdecl { LL $ unitOL $1 }
1227 rhs :: { Located (GRHSs RdrName) }
1228 : '=' exp wherebinds { sL (comb3 $1 $2 $3) $ GRHSs (unguardedRHS $2) (unLoc $3) }
1229 | gdrhs wherebinds { LL $ GRHSs (reverse (unLoc $1)) (unLoc $2) }
1231 gdrhs :: { Located [LGRHS RdrName] }
1232 : gdrhs gdrh { LL ($2 : unLoc $1) }
1235 gdrh :: { LGRHS RdrName }
1236 : '|' guardquals '=' exp { sL (comb2 $1 $>) $ GRHS (unLoc $2) $4 }
1238 sigdecl :: { Located (OrdList (LHsDecl RdrName)) }
1239 : infixexp '::' sigtypedoc {% do s <- checkValSig $1 $3
1240 ; return (LL $ unitOL (LL $ SigD s)) }
1241 -- See Note [Declaration/signature overlap] for why we need infixexp here
1242 | var ',' sig_vars '::' sigtypedoc
1243 { LL $ toOL [ LL $ SigD (TypeSig n $5) | n <- $1 : unLoc $3 ] }
1244 | infix prec ops { LL $ toOL [ LL $ SigD (FixSig (FixitySig n (Fixity $2 (unLoc $1))))
1246 | '{-# INLINE' activation qvar '#-}'
1247 { LL $ unitOL (LL $ SigD (InlineSig $3 (mkInlinePragma (getINLINE $1) $2))) }
1248 | '{-# SPECIALISE' qvar '::' sigtypes1 '#-}'
1249 { LL $ toOL [ LL $ SigD (SpecSig $2 t defaultInlinePragma)
1251 | '{-# SPECIALISE_INLINE' activation qvar '::' sigtypes1 '#-}'
1252 { LL $ toOL [ LL $ SigD (SpecSig $3 t (mkInlinePragma (getSPEC_INLINE $1) $2))
1254 | '{-# SPECIALISE' 'instance' inst_type '#-}'
1255 { LL $ unitOL (LL $ SigD (SpecInstSig $3)) }
1257 -----------------------------------------------------------------------------
1260 quasiquote :: { Located (HsQuasiQuote RdrName) }
1261 : TH_QUASIQUOTE { let { loc = getLoc $1
1262 ; ITquasiQuote (quoter, quote, quoteSpan) = unLoc $1
1263 ; quoterId = mkUnqual varName quoter }
1264 in L1 (mkHsQuasiQuote quoterId quoteSpan quote) }
1266 exp :: { LHsExpr RdrName }
1267 : infixexp '::' sigtype { LL $ ExprWithTySig $1 $3 }
1268 | infixexp '-<' exp { LL $ HsArrApp $1 $3 placeHolderType HsFirstOrderApp True }
1269 | infixexp '>-' exp { LL $ HsArrApp $3 $1 placeHolderType HsFirstOrderApp False }
1270 | infixexp '-<<' exp { LL $ HsArrApp $1 $3 placeHolderType HsHigherOrderApp True }
1271 | infixexp '>>-' exp { LL $ HsArrApp $3 $1 placeHolderType HsHigherOrderApp False}
1274 infixexp :: { LHsExpr RdrName }
1276 | infixexp qop exp10 { LL (OpApp $1 $2 (panic "fixity") $3) }
1278 exp10 :: { LHsExpr RdrName }
1279 : '\\' apat apats opt_asig '->' exp
1280 { LL $ HsLam (mkMatchGroup [LL $ Match ($2:$3) $4
1283 | 'let' binds 'in' exp { LL $ HsLet (unLoc $2) $4 }
1284 | 'if' exp optSemi 'then' exp optSemi 'else' exp
1285 {% checkDoAndIfThenElse $2 $3 $5 $6 $8 >>
1286 return (LL $ mkHsIf $2 $5 $8) }
1287 | 'case' exp 'of' altslist { LL $ HsCase $2 (mkMatchGroup (unLoc $4)) }
1288 | '-' fexp { LL $ NegApp $2 noSyntaxExpr }
1290 | 'do' stmtlist {% let loc = comb2 $1 $2 in
1291 checkDo loc (unLoc $2) >>= \ (stmts,body) ->
1292 return (L loc (mkHsDo DoExpr stmts body)) }
1293 | 'mdo' stmtlist {% let loc = comb2 $1 $2 in
1294 checkDo loc (unLoc $2) >>= \ (stmts,body) ->
1295 return (L loc (mkHsDo (MDoExpr noPostTcTable) stmts body)) }
1296 | scc_annot exp { LL $ if opt_SccProfilingOn
1297 then HsSCC (unLoc $1) $2
1299 | hpc_annot exp { LL $ if opt_Hpc
1300 then HsTickPragma (unLoc $1) $2
1303 | 'proc' aexp '->' exp
1304 {% checkPattern $2 >>= \ p ->
1305 return (LL $ HsProc p (LL $ HsCmdTop $4 []
1306 placeHolderType undefined)) }
1307 -- TODO: is LL right here?
1309 | '{-# CORE' STRING '#-}' exp { LL $ HsCoreAnn (getSTRING $2) $4 }
1310 -- hdaume: core annotation
1315 | {- empty -} { False }
1317 scc_annot :: { Located FastString }
1318 : '_scc_' STRING {% (addWarning Opt_WarnWarningsDeprecations (getLoc $1) (text "_scc_ is deprecated; use an SCC pragma instead")) >>= \_ ->
1319 ( do scc <- getSCC $2; return $ LL scc ) }
1320 | '{-# SCC' STRING '#-}' {% do scc <- getSCC $2; return $ LL scc }
1322 hpc_annot :: { Located (FastString,(Int,Int),(Int,Int)) }
1323 : '{-# GENERATED' STRING INTEGER ':' INTEGER '-' INTEGER ':' INTEGER '#-}'
1324 { LL $ (getSTRING $2
1325 ,( fromInteger $ getINTEGER $3
1326 , fromInteger $ getINTEGER $5
1328 ,( fromInteger $ getINTEGER $7
1329 , fromInteger $ getINTEGER $9
1334 fexp :: { LHsExpr RdrName }
1335 : fexp aexp { LL $ HsApp $1 $2 }
1338 aexp :: { LHsExpr RdrName }
1339 : qvar '@' aexp { LL $ EAsPat $1 $3 }
1340 | '~' aexp { LL $ ELazyPat $2 }
1343 aexp1 :: { LHsExpr RdrName }
1344 : aexp1 '{' fbinds '}' {% do { r <- mkRecConstrOrUpdate $1 (comb2 $2 $4) $3
1348 -- Here was the syntax for type applications that I was planning
1349 -- but there are difficulties (e.g. what order for type args)
1350 -- so it's not enabled yet.
1351 -- But this case *is* used for the left hand side of a generic definition,
1352 -- which is parsed as an expression before being munged into a pattern
1353 | qcname '{|' type '|}' { LL $ HsApp (sL (getLoc $1) (HsVar (unLoc $1)))
1354 (sL (getLoc $3) (HsType $3)) }
1356 aexp2 :: { LHsExpr RdrName }
1357 : ipvar { L1 (HsIPVar $! unLoc $1) }
1358 | qcname { L1 (HsVar $! unLoc $1) }
1359 | literal { L1 (HsLit $! unLoc $1) }
1360 -- This will enable overloaded strings permanently. Normally the renamer turns HsString
1361 -- into HsOverLit when -foverloaded-strings is on.
1362 -- | STRING { sL (getLoc $1) (HsOverLit $! mkHsIsString (getSTRING $1) placeHolderType) }
1363 | INTEGER { sL (getLoc $1) (HsOverLit $! mkHsIntegral (getINTEGER $1) placeHolderType) }
1364 | RATIONAL { sL (getLoc $1) (HsOverLit $! mkHsFractional (getRATIONAL $1) placeHolderType) }
1366 -- N.B.: sections get parsed by these next two productions.
1367 -- This allows you to write, e.g., '(+ 3, 4 -)', which isn't correct Haskell98
1368 -- (you'd have to write '((+ 3), (4 -))')
1369 -- but the less cluttered version fell out of having texps.
1370 | '(' texp ')' { LL (HsPar $2) }
1371 | '(' tup_exprs ')' { LL (ExplicitTuple $2 Boxed) }
1373 | '(#' texp '#)' { LL (ExplicitTuple [Present $2] Unboxed) }
1374 | '(#' tup_exprs '#)' { LL (ExplicitTuple $2 Unboxed) }
1376 | '[' list ']' { LL (unLoc $2) }
1377 | '[:' parr ':]' { LL (unLoc $2) }
1378 | '_' { L1 EWildPat }
1380 -- Template Haskell Extension
1381 | TH_ID_SPLICE { L1 $ HsSpliceE (mkHsSplice
1382 (L1 $ HsVar (mkUnqual varName
1383 (getTH_ID_SPLICE $1)))) }
1384 | '$(' exp ')' { LL $ HsSpliceE (mkHsSplice $2) }
1387 | TH_VAR_QUOTE qvar { LL $ HsBracket (VarBr (unLoc $2)) }
1388 | TH_VAR_QUOTE qcon { LL $ HsBracket (VarBr (unLoc $2)) }
1389 | TH_TY_QUOTE tyvar { LL $ HsBracket (VarBr (unLoc $2)) }
1390 | TH_TY_QUOTE gtycon { LL $ HsBracket (VarBr (unLoc $2)) }
1391 | '[|' exp '|]' { LL $ HsBracket (ExpBr $2) }
1392 | '[t|' ctype '|]' { LL $ HsBracket (TypBr $2) }
1393 | '[p|' infixexp '|]' {% checkPattern $2 >>= \p ->
1394 return (LL $ HsBracket (PatBr p)) }
1395 | '[d|' cvtopbody '|]' { LL $ HsBracket (DecBrL $2) }
1396 | quasiquote { L1 (HsQuasiQuoteE (unLoc $1)) }
1398 -- arrow notation extension
1399 | '(|' aexp2 cmdargs '|)' { LL $ HsArrForm $2 Nothing (reverse $3) }
1401 cmdargs :: { [LHsCmdTop RdrName] }
1402 : cmdargs acmd { $2 : $1 }
1403 | {- empty -} { [] }
1405 acmd :: { LHsCmdTop RdrName }
1406 : aexp2 { L1 $ HsCmdTop $1 [] placeHolderType undefined }
1408 cvtopbody :: { [LHsDecl RdrName] }
1409 : '{' cvtopdecls0 '}' { $2 }
1410 | vocurly cvtopdecls0 close { $2 }
1412 cvtopdecls0 :: { [LHsDecl RdrName] }
1413 : {- empty -} { [] }
1416 -----------------------------------------------------------------------------
1417 -- Tuple expressions
1419 -- "texp" is short for tuple expressions:
1420 -- things that can appear unparenthesized as long as they're
1421 -- inside parens or delimitted by commas
1422 texp :: { LHsExpr RdrName }
1425 -- Note [Parsing sections]
1426 -- ~~~~~~~~~~~~~~~~~~~~~~~
1427 -- We include left and right sections here, which isn't
1428 -- technically right according to Haskell 98. For example
1429 -- (3 +, True) isn't legal
1430 -- However, we want to parse bang patterns like
1432 -- and it's convenient to do so here as a section
1433 -- Then when converting expr to pattern we unravel it again
1434 -- Meanwhile, the renamer checks that real sections appear
1436 | infixexp qop { LL $ SectionL $1 $2 }
1437 | qopm infixexp { LL $ SectionR $1 $2 }
1439 -- View patterns get parenthesized above
1440 | exp '->' texp { LL $ EViewPat $1 $3 }
1442 -- Always at least one comma
1443 tup_exprs :: { [HsTupArg RdrName] }
1444 : texp commas_tup_tail { Present $1 : $2 }
1445 | commas tup_tail { replicate $1 missingTupArg ++ $2 }
1447 -- Always starts with commas; always follows an expr
1448 commas_tup_tail :: { [HsTupArg RdrName] }
1449 commas_tup_tail : commas tup_tail { replicate ($1-1) missingTupArg ++ $2 }
1451 -- Always follows a comma
1452 tup_tail :: { [HsTupArg RdrName] }
1453 : texp commas_tup_tail { Present $1 : $2 }
1454 | texp { [Present $1] }
1455 | {- empty -} { [missingTupArg] }
1457 -----------------------------------------------------------------------------
1460 -- The rules below are little bit contorted to keep lexps left-recursive while
1461 -- avoiding another shift/reduce-conflict.
1463 list :: { LHsExpr RdrName }
1464 : texp { L1 $ ExplicitList placeHolderType [$1] }
1465 | lexps { L1 $ ExplicitList placeHolderType (reverse (unLoc $1)) }
1466 | texp '..' { LL $ ArithSeq noPostTcExpr (From $1) }
1467 | texp ',' exp '..' { LL $ ArithSeq noPostTcExpr (FromThen $1 $3) }
1468 | texp '..' exp { LL $ ArithSeq noPostTcExpr (FromTo $1 $3) }
1469 | texp ',' exp '..' exp { LL $ ArithSeq noPostTcExpr (FromThenTo $1 $3 $5) }
1470 | texp '|' flattenedpquals { sL (comb2 $1 $>) $ mkHsDo ListComp (unLoc $3) $1 }
1472 lexps :: { Located [LHsExpr RdrName] }
1473 : lexps ',' texp { LL (((:) $! $3) $! unLoc $1) }
1474 | texp ',' texp { LL [$3,$1] }
1476 -----------------------------------------------------------------------------
1477 -- List Comprehensions
1479 flattenedpquals :: { Located [LStmt RdrName] }
1480 : pquals { case (unLoc $1) of
1482 -- We just had one thing in our "parallel" list so
1483 -- we simply return that thing directly
1485 qss -> L1 [L1 $ ParStmt [(qs, undefined) | qs <- qss]]
1486 -- We actually found some actual parallel lists so
1487 -- we wrap them into as a ParStmt
1490 pquals :: { Located [[LStmt RdrName]] }
1491 : squals '|' pquals { L (getLoc $2) (reverse (unLoc $1) : unLoc $3) }
1492 | squals { L (getLoc $1) [reverse (unLoc $1)] }
1494 squals :: { Located [LStmt RdrName] } -- In reverse order, because the last
1495 -- one can "grab" the earlier ones
1496 : squals ',' transformqual { LL [L (getLoc $3) ((unLoc $3) (reverse (unLoc $1)))] }
1497 | squals ',' qual { LL ($3 : unLoc $1) }
1498 | transformqual { LL [L (getLoc $1) ((unLoc $1) [])] }
1500 -- | transformquals1 ',' '{|' pquals '|}' { LL ($4 : unLoc $1) }
1501 -- | '{|' pquals '|}' { L1 [$2] }
1504 -- It is possible to enable bracketing (associating) qualifier lists by uncommenting the lines with {| |}
1505 -- above. Due to a lack of consensus on the syntax, this feature is not being used until we get user
1506 -- demand. Note that the {| |} symbols are reused from -XGenerics and hence if you want to compile
1507 -- a program that makes use of this temporary syntax you must supply that flag to GHC
1509 transformqual :: { Located ([LStmt RdrName] -> Stmt RdrName) }
1510 -- Function is applied to a list of stmts *in order*
1511 : 'then' exp { LL $ \leftStmts -> (mkTransformStmt leftStmts $2) }
1513 | 'then' exp 'by' exp { LL $ \leftStmts -> (mkTransformByStmt leftStmts $2 $4) }
1514 | 'then' 'group' 'by' exp { LL $ \leftStmts -> (mkGroupByStmt leftStmts $4) }
1516 -- These two productions deliberately have a shift-reduce conflict. I have made 'group' into a special_id,
1517 -- which means you can enable TransformListComp while still using Data.List.group. However, this makes the two
1518 -- productions ambiguous. I've set things up so that Happy chooses to resolve the conflict in that case by
1519 -- choosing the "group by" variant, which is what we want.
1521 -- This is rather dubious: the user might be confused as to how to parse this statement. However, it is a good
1522 -- practical choice. NB: Data.List.group :: [a] -> [[a]], so using the first production would not even type check
1523 -- if /that/ is the group function we conflict with.
1524 | 'then' 'group' 'using' exp { LL $ \leftStmts -> (mkGroupUsingStmt leftStmts $4) }
1525 | 'then' 'group' 'by' exp 'using' exp { LL $ \leftStmts -> (mkGroupByUsingStmt leftStmts $4 $6) }
1527 -----------------------------------------------------------------------------
1528 -- Parallel array expressions
1530 -- The rules below are little bit contorted; see the list case for details.
1531 -- Note that, in contrast to lists, we only have finite arithmetic sequences.
1532 -- Moreover, we allow explicit arrays with no element (represented by the nil
1533 -- constructor in the list case).
1535 parr :: { LHsExpr RdrName }
1536 : { noLoc (ExplicitPArr placeHolderType []) }
1537 | texp { L1 $ ExplicitPArr placeHolderType [$1] }
1538 | lexps { L1 $ ExplicitPArr placeHolderType
1539 (reverse (unLoc $1)) }
1540 | texp '..' exp { LL $ PArrSeq noPostTcExpr (FromTo $1 $3) }
1541 | texp ',' exp '..' exp { LL $ PArrSeq noPostTcExpr (FromThenTo $1 $3 $5) }
1542 | texp '|' flattenedpquals { LL $ mkHsDo PArrComp (unLoc $3) $1 }
1544 -- We are reusing `lexps' and `flattenedpquals' from the list case.
1546 -----------------------------------------------------------------------------
1549 guardquals :: { Located [LStmt RdrName] }
1550 : guardquals1 { L (getLoc $1) (reverse (unLoc $1)) }
1552 guardquals1 :: { Located [LStmt RdrName] }
1553 : guardquals1 ',' qual { LL ($3 : unLoc $1) }
1556 -----------------------------------------------------------------------------
1557 -- Case alternatives
1559 altslist :: { Located [LMatch RdrName] }
1560 : '{' alts '}' { LL (reverse (unLoc $2)) }
1561 | vocurly alts close { L (getLoc $2) (reverse (unLoc $2)) }
1563 alts :: { Located [LMatch RdrName] }
1564 : alts1 { L1 (unLoc $1) }
1565 | ';' alts { LL (unLoc $2) }
1567 alts1 :: { Located [LMatch RdrName] }
1568 : alts1 ';' alt { LL ($3 : unLoc $1) }
1569 | alts1 ';' { LL (unLoc $1) }
1572 alt :: { LMatch RdrName }
1573 : pat opt_sig alt_rhs { LL (Match [$1] $2 (unLoc $3)) }
1575 alt_rhs :: { Located (GRHSs RdrName) }
1576 : ralt wherebinds { LL (GRHSs (unLoc $1) (unLoc $2)) }
1578 ralt :: { Located [LGRHS RdrName] }
1579 : '->' exp { LL (unguardedRHS $2) }
1580 | gdpats { L1 (reverse (unLoc $1)) }
1582 gdpats :: { Located [LGRHS RdrName] }
1583 : gdpats gdpat { LL ($2 : unLoc $1) }
1586 gdpat :: { LGRHS RdrName }
1587 : '|' guardquals '->' exp { sL (comb2 $1 $>) $ GRHS (unLoc $2) $4 }
1589 -- 'pat' recognises a pattern, including one with a bang at the top
1590 -- e.g. "!x" or "!(x,y)" or "C a b" etc
1591 -- Bangs inside are parsed as infix operator applications, so that
1592 -- we parse them right when bang-patterns are off
1593 pat :: { LPat RdrName }
1594 pat : exp {% checkPattern $1 }
1595 | '!' aexp {% checkPattern (LL (SectionR (L1 (HsVar bang_RDR)) $2)) }
1597 apat :: { LPat RdrName }
1598 apat : aexp {% checkPattern $1 }
1599 | '!' aexp {% checkPattern (LL (SectionR (L1 (HsVar bang_RDR)) $2)) }
1601 apats :: { [LPat RdrName] }
1602 : apat apats { $1 : $2 }
1603 | {- empty -} { [] }
1605 -----------------------------------------------------------------------------
1606 -- Statement sequences
1608 stmtlist :: { Located [LStmt RdrName] }
1609 : '{' stmts '}' { LL (unLoc $2) }
1610 | vocurly stmts close { $2 }
1612 -- do { ;; s ; s ; ; s ;; }
1613 -- The last Stmt should be an expression, but that's hard to enforce
1614 -- here, because we need too much lookahead if we see do { e ; }
1615 -- So we use ExprStmts throughout, and switch the last one over
1616 -- in ParseUtils.checkDo instead
1617 stmts :: { Located [LStmt RdrName] }
1618 : stmt stmts_help { LL ($1 : unLoc $2) }
1619 | ';' stmts { LL (unLoc $2) }
1620 | {- empty -} { noLoc [] }
1622 stmts_help :: { Located [LStmt RdrName] } -- might be empty
1623 : ';' stmts { LL (unLoc $2) }
1624 | {- empty -} { noLoc [] }
1626 -- For typing stmts at the GHCi prompt, where
1627 -- the input may consist of just comments.
1628 maybe_stmt :: { Maybe (LStmt RdrName) }
1630 | {- nothing -} { Nothing }
1632 stmt :: { LStmt RdrName }
1634 | 'rec' stmtlist { LL $ mkRecStmt (unLoc $2) }
1636 qual :: { LStmt RdrName }
1637 : pat '<-' exp { LL $ mkBindStmt $1 $3 }
1638 | exp { L1 $ mkExprStmt $1 }
1639 | 'let' binds { LL $ LetStmt (unLoc $2) }
1641 -----------------------------------------------------------------------------
1642 -- Record Field Update/Construction
1644 fbinds :: { ([HsRecField RdrName (LHsExpr RdrName)], Bool) }
1646 | {- empty -} { ([], False) }
1648 fbinds1 :: { ([HsRecField RdrName (LHsExpr RdrName)], Bool) }
1649 : fbind ',' fbinds1 { case $3 of (flds, dd) -> ($1 : flds, dd) }
1650 | fbind { ([$1], False) }
1651 | '..' { ([], True) }
1653 fbind :: { HsRecField RdrName (LHsExpr RdrName) }
1654 : qvar '=' exp { HsRecField $1 $3 False }
1655 | qvar { HsRecField $1 placeHolderPunRhs True }
1656 -- In the punning case, use a place-holder
1657 -- The renamer fills in the final value
1659 -----------------------------------------------------------------------------
1660 -- Implicit Parameter Bindings
1662 dbinds :: { Located [LIPBind RdrName] }
1663 : dbinds ';' dbind { let { this = $3; rest = unLoc $1 }
1664 in rest `seq` this `seq` LL (this : rest) }
1665 | dbinds ';' { LL (unLoc $1) }
1666 | dbind { let this = $1 in this `seq` L1 [this] }
1667 -- | {- empty -} { [] }
1669 dbind :: { LIPBind RdrName }
1670 dbind : ipvar '=' exp { LL (IPBind (unLoc $1) $3) }
1672 ipvar :: { Located (IPName RdrName) }
1673 : IPDUPVARID { L1 (IPName (mkUnqual varName (getIPDUPVARID $1))) }
1675 -----------------------------------------------------------------------------
1676 -- Warnings and deprecations
1678 namelist :: { Located [RdrName] }
1679 namelist : name_var { L1 [unLoc $1] }
1680 | name_var ',' namelist { LL (unLoc $1 : unLoc $3) }
1682 name_var :: { Located RdrName }
1683 name_var : var { $1 }
1686 -----------------------------------------
1687 -- Data constructors
1688 qcon :: { Located RdrName }
1690 | '(' qconsym ')' { LL (unLoc $2) }
1691 | sysdcon { L1 $ nameRdrName (dataConName (unLoc $1)) }
1692 -- The case of '[:' ':]' is part of the production `parr'
1694 con :: { Located RdrName }
1696 | '(' consym ')' { LL (unLoc $2) }
1697 | sysdcon { L1 $ nameRdrName (dataConName (unLoc $1)) }
1699 con_list :: { Located [Located RdrName] }
1700 con_list : con { L1 [$1] }
1701 | con ',' con_list { LL ($1 : unLoc $3) }
1703 sysdcon :: { Located DataCon } -- Wired in data constructors
1704 : '(' ')' { LL unitDataCon }
1705 | '(' commas ')' { LL $ tupleCon Boxed ($2 + 1) }
1706 | '(#' '#)' { LL $ unboxedSingletonDataCon }
1707 | '(#' commas '#)' { LL $ tupleCon Unboxed ($2 + 1) }
1708 | '[' ']' { LL nilDataCon }
1710 conop :: { Located RdrName }
1712 | '`' conid '`' { LL (unLoc $2) }
1714 qconop :: { Located RdrName }
1716 | '`' qconid '`' { LL (unLoc $2) }
1718 -----------------------------------------------------------------------------
1719 -- Type constructors
1721 gtycon :: { Located RdrName } -- A "general" qualified tycon
1723 | '(' ')' { LL $ getRdrName unitTyCon }
1724 | '(' commas ')' { LL $ getRdrName (tupleTyCon Boxed ($2 + 1)) }
1725 | '(#' '#)' { LL $ getRdrName unboxedSingletonTyCon }
1726 | '(#' commas '#)' { LL $ getRdrName (tupleTyCon Unboxed ($2 + 1)) }
1727 | '(' '->' ')' { LL $ getRdrName funTyCon }
1728 | '[' ']' { LL $ listTyCon_RDR }
1729 | '[:' ':]' { LL $ parrTyCon_RDR }
1731 oqtycon :: { Located RdrName } -- An "ordinary" qualified tycon
1733 | '(' qtyconsym ')' { LL (unLoc $2) }
1735 qtyconop :: { Located RdrName } -- Qualified or unqualified
1737 | '`' qtycon '`' { LL (unLoc $2) }
1739 qtycon :: { Located RdrName } -- Qualified or unqualified
1740 : QCONID { L1 $! mkQual tcClsName (getQCONID $1) }
1741 | PREFIXQCONSYM { L1 $! mkQual tcClsName (getPREFIXQCONSYM $1) }
1744 tycon :: { Located RdrName } -- Unqualified
1745 : CONID { L1 $! mkUnqual tcClsName (getCONID $1) }
1747 qtyconsym :: { Located RdrName }
1748 : QCONSYM { L1 $! mkQual tcClsName (getQCONSYM $1) }
1751 tyconsym :: { Located RdrName }
1752 : CONSYM { L1 $! mkUnqual tcClsName (getCONSYM $1) }
1754 -----------------------------------------------------------------------------
1757 op :: { Located RdrName } -- used in infix decls
1761 varop :: { Located RdrName }
1763 | '`' varid '`' { LL (unLoc $2) }
1765 qop :: { LHsExpr RdrName } -- used in sections
1766 : qvarop { L1 $ HsVar (unLoc $1) }
1767 | qconop { L1 $ HsVar (unLoc $1) }
1769 qopm :: { LHsExpr RdrName } -- used in sections
1770 : qvaropm { L1 $ HsVar (unLoc $1) }
1771 | qconop { L1 $ HsVar (unLoc $1) }
1773 qvarop :: { Located RdrName }
1775 | '`' qvarid '`' { LL (unLoc $2) }
1777 qvaropm :: { Located RdrName }
1778 : qvarsym_no_minus { $1 }
1779 | '`' qvarid '`' { LL (unLoc $2) }
1781 -----------------------------------------------------------------------------
1784 tyvar :: { Located RdrName }
1785 tyvar : tyvarid { $1 }
1786 | '(' tyvarsym ')' { LL (unLoc $2) }
1788 tyvarop :: { Located RdrName }
1789 tyvarop : '`' tyvarid '`' { LL (unLoc $2) }
1791 | '.' {% parseErrorSDoc (getLoc $1)
1792 (vcat [ptext (sLit "Illegal symbol '.' in type"),
1793 ptext (sLit "Perhaps you intended -XRankNTypes or similar flag"),
1794 ptext (sLit "to enable explicit-forall syntax: forall <tvs>. <type>")])
1797 tyvarid :: { Located RdrName }
1798 : VARID { L1 $! mkUnqual tvName (getVARID $1) }
1799 | special_id { L1 $! mkUnqual tvName (unLoc $1) }
1800 | 'unsafe' { L1 $! mkUnqual tvName (fsLit "unsafe") }
1801 | 'safe' { L1 $! mkUnqual tvName (fsLit "safe") }
1802 | 'interruptible' { L1 $! mkUnqual tvName (fsLit "interruptible") }
1803 | 'threadsafe' { L1 $! mkUnqual tvName (fsLit "threadsafe") }
1805 tyvarsym :: { Located RdrName }
1806 -- Does not include "!", because that is used for strictness marks
1807 -- or ".", because that separates the quantified type vars from the rest
1808 -- or "*", because that's used for kinds
1809 tyvarsym : VARSYM { L1 $! mkUnqual tvName (getVARSYM $1) }
1811 -----------------------------------------------------------------------------
1814 var :: { Located RdrName }
1816 | '(' varsym ')' { LL (unLoc $2) }
1818 qvar :: { Located RdrName }
1820 | '(' varsym ')' { LL (unLoc $2) }
1821 | '(' qvarsym1 ')' { LL (unLoc $2) }
1822 -- We've inlined qvarsym here so that the decision about
1823 -- whether it's a qvar or a var can be postponed until
1824 -- *after* we see the close paren.
1826 qvarid :: { Located RdrName }
1828 | QVARID { L1 $! mkQual varName (getQVARID $1) }
1829 | PREFIXQVARSYM { L1 $! mkQual varName (getPREFIXQVARSYM $1) }
1831 varid :: { Located RdrName }
1832 : VARID { L1 $! mkUnqual varName (getVARID $1) }
1833 | special_id { L1 $! mkUnqual varName (unLoc $1) }
1834 | 'unsafe' { L1 $! mkUnqual varName (fsLit "unsafe") }
1835 | 'safe' { L1 $! mkUnqual varName (fsLit "safe") }
1836 | 'interruptible' { L1 $! mkUnqual varName (fsLit "interruptible") }
1837 | 'threadsafe' { L1 $! mkUnqual varName (fsLit "threadsafe") }
1838 | 'forall' { L1 $! mkUnqual varName (fsLit "forall") }
1839 | 'family' { L1 $! mkUnqual varName (fsLit "family") }
1841 qvarsym :: { Located RdrName }
1845 qvarsym_no_minus :: { Located RdrName }
1846 : varsym_no_minus { $1 }
1849 qvarsym1 :: { Located RdrName }
1850 qvarsym1 : QVARSYM { L1 $ mkQual varName (getQVARSYM $1) }
1852 varsym :: { Located RdrName }
1853 : varsym_no_minus { $1 }
1854 | '-' { L1 $ mkUnqual varName (fsLit "-") }
1856 varsym_no_minus :: { Located RdrName } -- varsym not including '-'
1857 : VARSYM { L1 $ mkUnqual varName (getVARSYM $1) }
1858 | special_sym { L1 $ mkUnqual varName (unLoc $1) }
1861 -- These special_ids are treated as keywords in various places,
1862 -- but as ordinary ids elsewhere. 'special_id' collects all these
1863 -- except 'unsafe', 'interruptible', 'forall', and 'family' whose treatment differs
1864 -- depending on context
1865 special_id :: { Located FastString }
1867 : 'as' { L1 (fsLit "as") }
1868 | 'qualified' { L1 (fsLit "qualified") }
1869 | 'hiding' { L1 (fsLit "hiding") }
1870 | 'export' { L1 (fsLit "export") }
1871 | 'label' { L1 (fsLit "label") }
1872 | 'dynamic' { L1 (fsLit "dynamic") }
1873 | 'stdcall' { L1 (fsLit "stdcall") }
1874 | 'ccall' { L1 (fsLit "ccall") }
1875 | 'prim' { L1 (fsLit "prim") }
1876 | 'group' { L1 (fsLit "group") }
1878 special_sym :: { Located FastString }
1879 special_sym : '!' { L1 (fsLit "!") }
1880 | '.' { L1 (fsLit ".") }
1881 | '*' { L1 (fsLit "*") }
1883 -----------------------------------------------------------------------------
1884 -- Data constructors
1886 qconid :: { Located RdrName } -- Qualified or unqualified
1888 | QCONID { L1 $! mkQual dataName (getQCONID $1) }
1889 | PREFIXQCONSYM { L1 $! mkQual dataName (getPREFIXQCONSYM $1) }
1891 conid :: { Located RdrName }
1892 : CONID { L1 $ mkUnqual dataName (getCONID $1) }
1894 qconsym :: { Located RdrName } -- Qualified or unqualified
1896 | QCONSYM { L1 $ mkQual dataName (getQCONSYM $1) }
1898 consym :: { Located RdrName }
1899 : CONSYM { L1 $ mkUnqual dataName (getCONSYM $1) }
1901 -- ':' means only list cons
1902 | ':' { L1 $ consDataCon_RDR }
1905 -----------------------------------------------------------------------------
1908 literal :: { Located HsLit }
1909 : CHAR { L1 $ HsChar $ getCHAR $1 }
1910 | STRING { L1 $ HsString $ getSTRING $1 }
1911 | PRIMINTEGER { L1 $ HsIntPrim $ getPRIMINTEGER $1 }
1912 | PRIMWORD { L1 $ HsWordPrim $ getPRIMWORD $1 }
1913 | PRIMCHAR { L1 $ HsCharPrim $ getPRIMCHAR $1 }
1914 | PRIMSTRING { L1 $ HsStringPrim $ getPRIMSTRING $1 }
1915 | PRIMFLOAT { L1 $ HsFloatPrim $ getPRIMFLOAT $1 }
1916 | PRIMDOUBLE { L1 $ HsDoublePrim $ getPRIMDOUBLE $1 }
1918 -----------------------------------------------------------------------------
1922 : vccurly { () } -- context popped in lexer.
1923 | error {% popContext }
1925 -----------------------------------------------------------------------------
1926 -- Miscellaneous (mostly renamings)
1928 modid :: { Located ModuleName }
1929 : CONID { L1 $ mkModuleNameFS (getCONID $1) }
1930 | QCONID { L1 $ let (mod,c) = getQCONID $1 in
1933 (unpackFS mod ++ '.':unpackFS c))
1937 : commas ',' { $1 + 1 }
1940 -----------------------------------------------------------------------------
1941 -- Documentation comments
1943 docnext :: { LHsDocString }
1944 : DOCNEXT {% return (L1 (HsDocString (mkFastString (getDOCNEXT $1)))) }
1946 docprev :: { LHsDocString }
1947 : DOCPREV {% return (L1 (HsDocString (mkFastString (getDOCPREV $1)))) }
1949 docnamed :: { Located (String, HsDocString) }
1951 let string = getDOCNAMED $1
1952 (name, rest) = break isSpace string
1953 in return (L1 (name, HsDocString (mkFastString rest))) }
1955 docsection :: { Located (Int, HsDocString) }
1956 : DOCSECTION {% let (n, doc) = getDOCSECTION $1 in
1957 return (L1 (n, HsDocString (mkFastString doc))) }
1959 moduleheader :: { Maybe LHsDocString }
1960 : DOCNEXT {% let string = getDOCNEXT $1 in
1961 return (Just (L1 (HsDocString (mkFastString string)))) }
1963 maybe_docprev :: { Maybe LHsDocString }
1964 : docprev { Just $1 }
1965 | {- empty -} { Nothing }
1967 maybe_docnext :: { Maybe LHsDocString }
1968 : docnext { Just $1 }
1969 | {- empty -} { Nothing }
1973 happyError = srcParseFail
1975 getVARID (L _ (ITvarid x)) = x
1976 getCONID (L _ (ITconid x)) = x
1977 getVARSYM (L _ (ITvarsym x)) = x
1978 getCONSYM (L _ (ITconsym x)) = x
1979 getQVARID (L _ (ITqvarid x)) = x
1980 getQCONID (L _ (ITqconid x)) = x
1981 getQVARSYM (L _ (ITqvarsym x)) = x
1982 getQCONSYM (L _ (ITqconsym x)) = x
1983 getPREFIXQVARSYM (L _ (ITprefixqvarsym x)) = x
1984 getPREFIXQCONSYM (L _ (ITprefixqconsym x)) = x
1985 getIPDUPVARID (L _ (ITdupipvarid x)) = x
1986 getCHAR (L _ (ITchar x)) = x
1987 getSTRING (L _ (ITstring x)) = x
1988 getINTEGER (L _ (ITinteger x)) = x
1989 getRATIONAL (L _ (ITrational x)) = x
1990 getPRIMCHAR (L _ (ITprimchar x)) = x
1991 getPRIMSTRING (L _ (ITprimstring x)) = x
1992 getPRIMINTEGER (L _ (ITprimint x)) = x
1993 getPRIMWORD (L _ (ITprimword x)) = x
1994 getPRIMFLOAT (L _ (ITprimfloat x)) = x
1995 getPRIMDOUBLE (L _ (ITprimdouble x)) = x
1996 getTH_ID_SPLICE (L _ (ITidEscape x)) = x
1997 getINLINE (L _ (ITinline_prag inl conl)) = (inl,conl)
1998 getSPEC_INLINE (L _ (ITspec_inline_prag True)) = (Inline, FunLike)
1999 getSPEC_INLINE (L _ (ITspec_inline_prag False)) = (NoInline,FunLike)
2001 getDOCNEXT (L _ (ITdocCommentNext x)) = x
2002 getDOCPREV (L _ (ITdocCommentPrev x)) = x
2003 getDOCNAMED (L _ (ITdocCommentNamed x)) = x
2004 getDOCSECTION (L _ (ITdocSection n x)) = (n, x)
2006 getSCC :: Located Token -> P FastString
2007 getSCC lt = do let s = getSTRING lt
2008 err = "Spaces are not allowed in SCCs"
2009 -- We probably actually want to be more restrictive than this
2010 if ' ' `elem` unpackFS s
2011 then failSpanMsgP (getLoc lt) (text err)
2014 -- Utilities for combining source spans
2015 comb2 :: Located a -> Located b -> SrcSpan
2016 comb2 a b = a `seq` b `seq` combineLocs a b
2018 comb3 :: Located a -> Located b -> Located c -> SrcSpan
2019 comb3 a b c = a `seq` b `seq` c `seq`
2020 combineSrcSpans (getLoc a) (combineSrcSpans (getLoc b) (getLoc c))
2022 comb4 :: Located a -> Located b -> Located c -> Located d -> SrcSpan
2023 comb4 a b c d = a `seq` b `seq` c `seq` d `seq`
2024 (combineSrcSpans (getLoc a) $ combineSrcSpans (getLoc b) $
2025 combineSrcSpans (getLoc c) (getLoc d))
2027 -- strict constructor version:
2029 sL :: SrcSpan -> a -> Located a
2030 sL span a = span `seq` a `seq` L span a
2032 -- Make a source location for the file. We're a bit lazy here and just
2033 -- make a point SrcSpan at line 1, column 0. Strictly speaking we should
2034 -- try to find the span of the whole file (ToDo).
2035 fileSrcSpan :: P SrcSpan
2038 let loc = mkSrcLoc (srcLocFile l) 1 1;
2039 return (mkSrcSpan loc loc)