1 -----------------------------------------------------------------------------
2 -- (c) The University of Glasgow, 2006
6 -- This is a combination of an Alex-generated lexer from a regex
7 -- definition, with some hand-coded bits.
9 -- Completely accurate information about token-spans within the source
10 -- file is maintained. Every token has a start and end SrcLoc attached to it.
12 -----------------------------------------------------------------------------
16 -- - parsing integers is a bit slow
17 -- - readRational is a bit slow
19 -- Known bugs, that were also in the previous version:
20 -- - M... should be 3 tokens, not 1.
21 -- - pragma-end should be only valid in a pragma
25 -- The above warning supression flag is a temporary kludge.
26 -- While working on this module you are encouraged to remove it and fix
27 -- any warnings in the module. See
28 -- http://hackage.haskell.org/trac/ghc/wiki/Commentary/CodingStyle#Warnings
31 -- Note that Alex itself generates code with with some unused bindings and
32 -- without type signatures, so removing the flag might not be possible.
34 {-# OPTIONS_GHC -funbox-strict-fields #-}
37 Token(..), lexer, pragState, mkPState, PState(..),
38 P(..), ParseResult(..), getSrcLoc,
39 failLocMsgP, failSpanMsgP, srcParseFail,
41 popContext, pushCurrentContext, setLastToken, setSrcLoc,
42 getLexState, popLexState, pushLexState,
43 extension, standaloneDerivingEnabled, bangPatEnabled,
57 import Util ( maybePrefixMatch, readRational )
61 import Data.Char ( chr, ord, isSpace )
65 import Unicode ( GeneralCategory(..), generalCategory, isPrint, isUpper )
68 $unispace = \x05 -- Trick Alex into handling Unicode. See alexGetChar.
69 $whitechar = [\ \n\r\f\v $unispace]
70 $white_no_nl = $whitechar # \n
74 $unidigit = \x03 -- Trick Alex into handling Unicode. See alexGetChar.
75 $decdigit = $ascdigit -- for now, should really be $digit (ToDo)
76 $digit = [$ascdigit $unidigit]
78 $special = [\(\)\,\;\[\]\`\{\}]
79 $ascsymbol = [\!\#\$\%\&\*\+\.\/\<\=\>\?\@\\\^\|\-\~]
80 $unisymbol = \x04 -- Trick Alex into handling Unicode. See alexGetChar.
81 $symbol = [$ascsymbol $unisymbol] # [$special \_\:\"\']
83 $unilarge = \x01 -- Trick Alex into handling Unicode. See alexGetChar.
85 $large = [$asclarge $unilarge]
87 $unismall = \x02 -- Trick Alex into handling Unicode. See alexGetChar.
89 $small = [$ascsmall $unismall \_]
91 $unigraphic = \x06 -- Trick Alex into handling Unicode. See alexGetChar.
92 $graphic = [$small $large $symbol $digit $special $unigraphic \:\"\']
95 $hexit = [$decdigit A-F a-f]
96 $symchar = [$symbol \:]
98 $idchar = [$small $large $digit \']
100 $docsym = [\| \^ \* \$]
102 @varid = $small $idchar*
103 @conid = $large $idchar*
105 @varsym = $symbol $symchar*
106 @consym = \: $symchar*
108 @decimal = $decdigit+
110 @hexadecimal = $hexit+
111 @exponent = [eE] [\-\+]? @decimal
113 -- we support the hierarchical module name extension:
116 @floating_point = @decimal \. @decimal @exponent? | @decimal @exponent
118 -- normal signed numerical literals can only be explicitly negative,
119 -- not explicitly positive (contrast @exponent)
121 @signed = @negative ?
125 -- everywhere: skip whitespace and comments
127 $tab+ { warn Opt_WarnTabs (text "Tab character") }
129 -- Everywhere: deal with nested comments. We explicitly rule out
130 -- pragmas, "{-#", so that we don't accidentally treat them as comments.
131 -- (this can happen even though pragmas will normally take precedence due to
132 -- longest-match, because pragmas aren't valid in every state, but comments
133 -- are). We also rule out nested Haddock comments, if the -haddock flag is
136 "{-" / { isNormalComment } { nested_comment lexToken }
138 -- Single-line comments are a bit tricky. Haskell 98 says that two or
139 -- more dashes followed by a symbol should be parsed as a varsym, so we
140 -- have to exclude those.
142 -- Since Haddock comments aren't valid in every state, we need to rule them
145 -- The following two rules match comments that begin with two dashes, but
146 -- continue with a different character. The rules test that this character
147 -- is not a symbol (in which case we'd have a varsym), and that it's not a
148 -- space followed by a Haddock comment symbol (docsym) (in which case we'd
149 -- have a Haddock comment). The rules then munch the rest of the line.
151 "-- " ~[$docsym \#] .* ;
152 "--" [^$symbol : \ ] .* ;
154 -- Next, match Haddock comments if no -haddock flag
156 "-- " [$docsym \#] .* / { ifExtension (not . haddockEnabled) } ;
158 -- Now, when we've matched comments that begin with 2 dashes and continue
159 -- with a different character, we need to match comments that begin with three
160 -- or more dashes (which clearly can't be Haddock comments). We only need to
161 -- make sure that the first non-dash character isn't a symbol, and munch the
164 "---"\-* [^$symbol :] .* ;
166 -- Since the previous rules all match dashes followed by at least one
167 -- character, we also need to match a whole line filled with just dashes.
169 "--"\-* / { atEOL } ;
171 -- We need this rule since none of the other single line comment rules
172 -- actually match this case.
176 -- 'bol' state: beginning of a line. Slurp up all the whitespace (including
177 -- blank lines) until we find a non-whitespace character, then do layout
180 -- One slight wibble here: what if the line begins with {-#? In
181 -- theory, we have to lex the pragma to see if it's one we recognise,
182 -- and if it is, then we backtrack and do_bol, otherwise we treat it
183 -- as a nested comment. We don't bother with this: if the line begins
184 -- with {-#, then we'll assume it's a pragma we know about and go for do_bol.
187 ^\# (line)? { begin line_prag1 }
188 ^\# pragma .* \n ; -- GCC 3.3 CPP generated, apparently
189 ^\# \! .* \n ; -- #!, for scripts
193 -- after a layout keyword (let, where, do, of), we begin a new layout
194 -- context if the curly brace is missing.
195 -- Careful! This stuff is quite delicate.
196 <layout, layout_do> {
197 \{ / { notFollowedBy '-' } { pop_and open_brace }
198 -- we might encounter {-# here, but {- has been handled already
200 ^\# (line)? { begin line_prag1 }
203 -- do is treated in a subtly different way, see new_layout_context
204 <layout> () { new_layout_context True }
205 <layout_do> () { new_layout_context False }
207 -- after a new layout context which was found to be to the left of the
208 -- previous context, we have generated a '{' token, and we now need to
209 -- generate a matching '}' token.
210 <layout_left> () { do_layout_left }
212 <0,option_prags> \n { begin bol }
214 "{-#" $whitechar* (line|LINE) { begin line_prag2 }
216 -- single-line line pragmas, of the form
217 -- # <line> "<file>" <extra-stuff> \n
218 <line_prag1> $decdigit+ { setLine line_prag1a }
219 <line_prag1a> \" [$graphic \ ]* \" { setFile line_prag1b }
220 <line_prag1b> .* { pop }
222 -- Haskell-style line pragmas, of the form
223 -- {-# LINE <line> "<file>" #-}
224 <line_prag2> $decdigit+ { setLine line_prag2a }
225 <line_prag2a> \" [$graphic \ ]* \" { setFile line_prag2b }
226 <line_prag2b> "#-}"|"-}" { pop }
227 -- NOTE: accept -} at the end of a LINE pragma, for compatibility
228 -- with older versions of GHC which generated these.
230 -- We only want RULES pragmas to be picked up when explicit forall
231 -- syntax is enabled is on, because the contents of the pragma always
232 -- uses it. If it's not on then we're sure to get a parse error.
233 -- (ToDo: we should really emit a warning when ignoring pragmas)
234 -- XXX Now that we can enable this without the -fglasgow-exts hammer,
235 -- is it better just to let the parse error happen?
237 "{-#" $whitechar* (RULES|rules) / { ifExtension explicitForallEnabled } { token ITrules_prag }
240 "{-#" $whitechar* (INLINE|inline) { token (ITinline_prag True) }
241 "{-#" $whitechar* (NO(T?)INLINE|no(t?)inline)
242 { token (ITinline_prag False) }
243 "{-#" $whitechar* (SPECIALI[SZ]E|speciali[sz]e)
244 { token ITspec_prag }
245 "{-#" $whitechar* (SPECIALI[SZ]E|speciali[sz]e)
246 $whitechar* (INLINE|inline) { token (ITspec_inline_prag True) }
247 "{-#" $whitechar* (SPECIALI[SZ]E|speciali[sz]e)
248 $whitechar* (NO(T?)INLINE|no(t?)inline)
249 { token (ITspec_inline_prag False) }
250 "{-#" $whitechar* (SOURCE|source) { token ITsource_prag }
251 "{-#" $whitechar* (DEPRECATED|deprecated)
252 { token ITdeprecated_prag }
253 "{-#" $whitechar* (SCC|scc) { token ITscc_prag }
254 "{-#" $whitechar* (GENERATED|generated)
255 { token ITgenerated_prag }
256 "{-#" $whitechar* (CORE|core) { token ITcore_prag }
257 "{-#" $whitechar* (UNPACK|unpack) { token ITunpack_prag }
259 "{-#" { nested_comment lexToken }
261 -- ToDo: should only be valid inside a pragma:
262 "#-}" { token ITclose_prag}
266 "{-#" $whitechar* (OPTIONS|options) { lex_string_prag IToptions_prag }
267 "{-#" $whitechar* (OPTIONS_GHC|options_ghc)
268 { lex_string_prag IToptions_prag }
269 "{-#" $whitechar* (OPTIONS_HADDOCK|options_haddock)
270 { lex_string_prag ITdocOptions }
271 "-- #" { multiline_doc_comment }
272 "{-#" $whitechar* (LANGUAGE|language) { token ITlanguage_prag }
273 "{-#" $whitechar* (INCLUDE|include) { lex_string_prag ITinclude_prag }
281 -- This is to catch things like {-# OPTIONS OPTIONS_HUGS ...
282 "{-#" $whitechar* $idchar+ { nested_comment lexToken }
285 -- '0' state: ordinary lexemes
290 "-- " $docsym / { ifExtension haddockEnabled } { multiline_doc_comment }
291 "{-" \ ? $docsym / { ifExtension haddockEnabled } { nested_doc_comment }
297 "[:" / { ifExtension parrEnabled } { token ITopabrack }
298 ":]" / { ifExtension parrEnabled } { token ITcpabrack }
302 "[|" / { ifExtension thEnabled } { token ITopenExpQuote }
303 "[e|" / { ifExtension thEnabled } { token ITopenExpQuote }
304 "[p|" / { ifExtension thEnabled } { token ITopenPatQuote }
305 "[d|" / { ifExtension thEnabled } { layout_token ITopenDecQuote }
306 "[t|" / { ifExtension thEnabled } { token ITopenTypQuote }
307 "|]" / { ifExtension thEnabled } { token ITcloseQuote }
308 \$ @varid / { ifExtension thEnabled } { skip_one_varid ITidEscape }
309 "$(" / { ifExtension thEnabled } { token ITparenEscape }
311 "[$" @varid "|" / { ifExtension qqEnabled }
312 { lex_quasiquote_tok }
316 "(|" / { ifExtension arrowsEnabled `alexAndPred` notFollowedBySymbol }
317 { special IToparenbar }
318 "|)" / { ifExtension arrowsEnabled } { special ITcparenbar }
322 \? @varid / { ifExtension ipEnabled } { skip_one_varid ITdupipvarid }
326 "(#" / { ifExtension unboxedTuplesEnabled `alexAndPred` notFollowedBySymbol }
327 { token IToubxparen }
328 "#)" / { ifExtension unboxedTuplesEnabled }
329 { token ITcubxparen }
333 "{|" / { ifExtension genericsEnabled } { token ITocurlybar }
334 "|}" / { ifExtension genericsEnabled } { token ITccurlybar }
338 \( { special IToparen }
339 \) { special ITcparen }
340 \[ { special ITobrack }
341 \] { special ITcbrack }
342 \, { special ITcomma }
343 \; { special ITsemi }
344 \` { special ITbackquote }
351 @qual @varid { idtoken qvarid }
352 @qual @conid { idtoken qconid }
354 @conid { idtoken conid }
358 @qual @varid "#"+ / { ifExtension magicHashEnabled } { idtoken qvarid }
359 @qual @conid "#"+ / { ifExtension magicHashEnabled } { idtoken qconid }
360 @varid "#"+ / { ifExtension magicHashEnabled } { varid }
361 @conid "#"+ / { ifExtension magicHashEnabled } { idtoken conid }
367 @qual @varsym { idtoken qvarsym }
368 @qual @consym { idtoken qconsym }
373 -- For the normal boxed literals we need to be careful
374 -- when trying to be close to Haskell98
376 -- Normal integral literals (:: Num a => a, from Integer)
377 @decimal { tok_num positive 0 0 decimal }
378 0[oO] @octal { tok_num positive 2 2 octal }
379 0[xX] @hexadecimal { tok_num positive 2 2 hexadecimal }
381 -- Normal rational literals (:: Fractional a => a, from Rational)
382 @floating_point { strtoken tok_float }
386 -- Unboxed ints (:: Int#) and words (:: Word#)
387 -- It's simpler (and faster?) to give separate cases to the negatives,
388 -- especially considering octal/hexadecimal prefixes.
389 @decimal \# / { ifExtension magicHashEnabled } { tok_primint positive 0 1 decimal }
390 0[oO] @octal \# / { ifExtension magicHashEnabled } { tok_primint positive 2 3 octal }
391 0[xX] @hexadecimal \# / { ifExtension magicHashEnabled } { tok_primint positive 2 3 hexadecimal }
392 @negative @decimal \# / { ifExtension magicHashEnabled } { tok_primint negative 1 2 decimal }
393 @negative 0[oO] @octal \# / { ifExtension magicHashEnabled } { tok_primint negative 3 4 octal }
394 @negative 0[xX] @hexadecimal \# / { ifExtension magicHashEnabled } { tok_primint negative 3 4 hexadecimal }
396 @decimal \# \# / { ifExtension magicHashEnabled } { tok_primword 0 2 decimal }
397 0[oO] @octal \# \# / { ifExtension magicHashEnabled } { tok_primword 2 4 octal }
398 0[xX] @hexadecimal \# \# / { ifExtension magicHashEnabled } { tok_primword 2 4 hexadecimal }
400 -- Unboxed floats and doubles (:: Float#, :: Double#)
401 -- prim_{float,double} work with signed literals
402 @signed @floating_point \# / { ifExtension magicHashEnabled } { init_strtoken 1 tok_primfloat }
403 @signed @floating_point \# \# / { ifExtension magicHashEnabled } { init_strtoken 2 tok_primdouble }
406 -- Strings and chars are lexed by hand-written code. The reason is
407 -- that even if we recognise the string or char here in the regex
408 -- lexer, we would still have to parse the string afterward in order
409 -- to convert it to a String.
412 \" { lex_string_tok }
416 -- -----------------------------------------------------------------------------
420 = ITas -- Haskell keywords
444 | ITscc -- ToDo: remove (we use {-# SCC "..." #-} now)
446 | ITforall -- GHC extension keywords
464 | ITinline_prag Bool -- True <=> INLINE, False <=> NOINLINE
465 | ITspec_prag -- SPECIALISE
466 | ITspec_inline_prag Bool -- SPECIALISE INLINE (or NOINLINE)
473 | ITcore_prag -- hdaume: core annotations
476 | IToptions_prag String
477 | ITinclude_prag String
480 | ITdotdot -- reserved symbols
496 | ITbiglam -- GHC-extension symbols
498 | ITocurly -- special symbols
500 | ITocurlybar -- {|, for type applications
501 | ITccurlybar -- |}, for type applications
505 | ITopabrack -- [:, for parallel arrays with -XParr
506 | ITcpabrack -- :], for parallel arrays with -XParr
517 | ITvarid FastString -- identifiers
519 | ITvarsym FastString
520 | ITconsym FastString
521 | ITqvarid (FastString,FastString)
522 | ITqconid (FastString,FastString)
523 | ITqvarsym (FastString,FastString)
524 | ITqconsym (FastString,FastString)
526 | ITdupipvarid FastString -- GHC extension: implicit param: ?x
528 | ITpragma StringBuffer
531 | ITstring FastString
533 | ITrational Rational
536 | ITprimstring FastString
539 | ITprimfloat Rational
540 | ITprimdouble Rational
542 -- MetaHaskell extension tokens
543 | ITopenExpQuote -- [| or [e|
544 | ITopenPatQuote -- [p|
545 | ITopenDecQuote -- [d|
546 | ITopenTypQuote -- [t|
548 | ITidEscape FastString -- $x
549 | ITparenEscape -- $(
552 | ITquasiQuote (FastString,FastString,SrcSpan) -- [:...|...|]
554 -- Arrow notation extension
561 | ITLarrowtail -- -<<
562 | ITRarrowtail -- >>-
564 | ITunknown String -- Used when the lexer can't make sense of it
565 | ITeof -- end of file token
567 -- Documentation annotations
568 | ITdocCommentNext String -- something beginning '-- |'
569 | ITdocCommentPrev String -- something beginning '-- ^'
570 | ITdocCommentNamed String -- something beginning '-- $'
571 | ITdocSection Int String -- a section heading
572 | ITdocOptions String -- doc options (prune, ignore-exports, etc)
573 | ITdocOptionsOld String -- doc options declared "-- # ..."-style
576 deriving Show -- debugging
580 isSpecial :: Token -> Bool
581 -- If we see M.x, where x is a keyword, but
582 -- is special, we treat is as just plain M.x,
584 isSpecial ITas = True
585 isSpecial IThiding = True
586 isSpecial ITqualified = True
587 isSpecial ITforall = True
588 isSpecial ITexport = True
589 isSpecial ITlabel = True
590 isSpecial ITdynamic = True
591 isSpecial ITsafe = True
592 isSpecial ITthreadsafe = True
593 isSpecial ITunsafe = True
594 isSpecial ITccallconv = True
595 isSpecial ITstdcallconv = True
596 isSpecial ITmdo = True
597 isSpecial ITfamily = True
598 isSpecial ITgroup = True
599 isSpecial ITby = True
600 isSpecial ITusing = True
604 -- the bitmap provided as the third component indicates whether the
605 -- corresponding extension keyword is valid under the extension options
606 -- provided to the compiler; if the extension corresponding to *any* of the
607 -- bits set in the bitmap is enabled, the keyword is valid (this setup
608 -- facilitates using a keyword in two different extensions that can be
609 -- activated independently)
611 reservedWordsFM = listToUFM $
612 map (\(x, y, z) -> (mkFastString x, (y, z)))
613 [( "_", ITunderscore, 0 ),
615 ( "case", ITcase, 0 ),
616 ( "class", ITclass, 0 ),
617 ( "data", ITdata, 0 ),
618 ( "default", ITdefault, 0 ),
619 ( "deriving", ITderiving, 0 ),
621 ( "else", ITelse, 0 ),
622 ( "hiding", IThiding, 0 ),
624 ( "import", ITimport, 0 ),
626 ( "infix", ITinfix, 0 ),
627 ( "infixl", ITinfixl, 0 ),
628 ( "infixr", ITinfixr, 0 ),
629 ( "instance", ITinstance, 0 ),
631 ( "module", ITmodule, 0 ),
632 ( "newtype", ITnewtype, 0 ),
634 ( "qualified", ITqualified, 0 ),
635 ( "then", ITthen, 0 ),
636 ( "type", ITtype, 0 ),
637 ( "where", ITwhere, 0 ),
638 ( "_scc_", ITscc, 0 ), -- ToDo: remove
640 ( "forall", ITforall, bit explicitForallBit),
641 ( "mdo", ITmdo, bit recursiveDoBit),
642 ( "family", ITfamily, bit tyFamBit),
643 ( "group", ITgroup, bit transformComprehensionsBit),
644 ( "by", ITby, bit transformComprehensionsBit),
645 ( "using", ITusing, bit transformComprehensionsBit),
647 ( "foreign", ITforeign, bit ffiBit),
648 ( "export", ITexport, bit ffiBit),
649 ( "label", ITlabel, bit ffiBit),
650 ( "dynamic", ITdynamic, bit ffiBit),
651 ( "safe", ITsafe, bit ffiBit),
652 ( "threadsafe", ITthreadsafe, bit ffiBit),
653 ( "unsafe", ITunsafe, bit ffiBit),
654 ( "stdcall", ITstdcallconv, bit ffiBit),
655 ( "ccall", ITccallconv, bit ffiBit),
656 ( "dotnet", ITdotnet, bit ffiBit),
658 ( "rec", ITrec, bit arrowsBit),
659 ( "proc", ITproc, bit arrowsBit)
662 reservedSymsFM :: UniqFM (Token, Int -> Bool)
663 reservedSymsFM = listToUFM $
664 map (\ (x,y,z) -> (mkFastString x,(y,z)))
665 [ ("..", ITdotdot, always)
666 -- (:) is a reserved op, meaning only list cons
667 ,(":", ITcolon, always)
668 ,("::", ITdcolon, always)
669 ,("=", ITequal, always)
670 ,("\\", ITlam, always)
671 ,("|", ITvbar, always)
672 ,("<-", ITlarrow, always)
673 ,("->", ITrarrow, always)
675 ,("~", ITtilde, always)
676 ,("=>", ITdarrow, always)
677 ,("-", ITminus, always)
678 ,("!", ITbang, always)
680 -- For data T (a::*) = MkT
681 ,("*", ITstar, \i -> kindSigsEnabled i || tyFamEnabled i)
682 -- For 'forall a . t'
683 ,(".", ITdot, explicitForallEnabled)
685 ,("-<", ITlarrowtail, arrowsEnabled)
686 ,(">-", ITrarrowtail, arrowsEnabled)
687 ,("-<<", ITLarrowtail, arrowsEnabled)
688 ,(">>-", ITRarrowtail, arrowsEnabled)
690 #if __GLASGOW_HASKELL__ >= 605
691 ,("∷", ITdcolon, unicodeSyntaxEnabled)
692 ,("⇒", ITdarrow, unicodeSyntaxEnabled)
693 ,("∀", ITforall, \i -> unicodeSyntaxEnabled i &&
694 explicitForallEnabled i)
695 ,("→", ITrarrow, unicodeSyntaxEnabled)
696 ,("←", ITlarrow, unicodeSyntaxEnabled)
697 ,("⋯", ITdotdot, unicodeSyntaxEnabled)
698 -- ToDo: ideally, → and ∷ should be "specials", so that they cannot
699 -- form part of a large operator. This would let us have a better
700 -- syntax for kinds: ɑ∷*→* would be a legal kind signature. (maybe).
704 -- -----------------------------------------------------------------------------
707 type Action = SrcSpan -> StringBuffer -> Int -> P (Located Token)
709 special :: Token -> Action
710 special tok span _buf _len = return (L span tok)
712 token, layout_token :: Token -> Action
713 token t span _buf _len = return (L span t)
714 layout_token t span _buf _len = pushLexState layout >> return (L span t)
716 idtoken :: (StringBuffer -> Int -> Token) -> Action
717 idtoken f span buf len = return (L span $! (f buf len))
719 skip_one_varid :: (FastString -> Token) -> Action
720 skip_one_varid f span buf len
721 = return (L span $! f (lexemeToFastString (stepOn buf) (len-1)))
723 strtoken :: (String -> Token) -> Action
724 strtoken f span buf len =
725 return (L span $! (f $! lexemeToString buf len))
727 init_strtoken :: Int -> (String -> Token) -> Action
728 -- like strtoken, but drops the last N character(s)
729 init_strtoken drop f span buf len =
730 return (L span $! (f $! lexemeToString buf (len-drop)))
732 begin :: Int -> Action
733 begin code _span _str _len = do pushLexState code; lexToken
736 pop _span _buf _len = do popLexState; lexToken
738 pop_and :: Action -> Action
739 pop_and act span buf len = do popLexState; act span buf len
741 {-# INLINE nextCharIs #-}
742 nextCharIs buf p = not (atEnd buf) && p (currentChar buf)
744 notFollowedBy char _ _ _ (AI _ _ buf)
745 = nextCharIs buf (/=char)
747 notFollowedBySymbol _ _ _ (AI _ _ buf)
748 = nextCharIs buf (`notElem` "!#$%&*+./<=>?@\\^|-~")
750 -- We must reject doc comments as being ordinary comments everywhere.
751 -- In some cases the doc comment will be selected as the lexeme due to
752 -- maximal munch, but not always, because the nested comment rule is
753 -- valid in all states, but the doc-comment rules are only valid in
754 -- the non-layout states.
755 isNormalComment bits _ _ (AI _ _ buf)
756 | haddockEnabled bits = notFollowedByDocOrPragma
757 | otherwise = nextCharIs buf (/='#')
759 notFollowedByDocOrPragma
760 = not $ spaceAndP buf (`nextCharIs` (`elem` "|^*$#"))
762 spaceAndP buf p = p buf || nextCharIs buf (==' ') && p (snd (nextChar buf))
765 haddockDisabledAnd p bits _ _ (AI _ _ buf)
766 = if haddockEnabled bits then False else (p buf)
769 atEOL _ _ _ (AI _ _ buf) = atEnd buf || currentChar buf == '\n'
771 ifExtension pred bits _ _ _ = pred bits
773 multiline_doc_comment :: Action
774 multiline_doc_comment span buf _len = withLexedDocType (worker "")
776 worker commentAcc input docType oneLine = case alexGetChar input of
778 | oneLine -> docCommentEnd input commentAcc docType buf span
779 | otherwise -> case checkIfCommentLine input' of
780 Just input -> worker ('\n':commentAcc) input docType False
781 Nothing -> docCommentEnd input commentAcc docType buf span
782 Just (c, input) -> worker (c:commentAcc) input docType oneLine
783 Nothing -> docCommentEnd input commentAcc docType buf span
785 checkIfCommentLine input = check (dropNonNewlineSpace input)
787 check input = case alexGetChar input of
788 Just ('-', input) -> case alexGetChar input of
789 Just ('-', input) -> case alexGetChar input of
790 Just (c, _) | c /= '-' -> Just input
795 dropNonNewlineSpace input = case alexGetChar input of
797 | isSpace c && c /= '\n' -> dropNonNewlineSpace input'
802 nested comments require traversing by hand, they can't be parsed
803 using regular expressions.
805 nested_comment :: P (Located Token) -> Action
806 nested_comment cont span _str _len = do
810 go 0 input = do setInput input; cont
811 go n input = case alexGetChar input of
812 Nothing -> errBrace input span
813 Just ('-',input) -> case alexGetChar input of
814 Nothing -> errBrace input span
815 Just ('\125',input) -> go (n-1) input
816 Just (_,_) -> go n input
817 Just ('\123',input) -> case alexGetChar input of
818 Nothing -> errBrace input span
819 Just ('-',input) -> go (n+1) input
820 Just (_,_) -> go n input
821 Just (_,input) -> go n input
823 nested_doc_comment :: Action
824 nested_doc_comment span buf _len = withLexedDocType (go "")
826 go commentAcc input docType _ = case alexGetChar input of
827 Nothing -> errBrace input span
828 Just ('-',input) -> case alexGetChar input of
829 Nothing -> errBrace input span
830 Just ('\125',input) ->
831 docCommentEnd input commentAcc docType buf span
832 Just (_,_) -> go ('-':commentAcc) input docType False
833 Just ('\123', input) -> case alexGetChar input of
834 Nothing -> errBrace input span
835 Just ('-',input) -> do
837 let cont = do input <- getInput; go commentAcc input docType False
838 nested_comment cont span buf _len
839 Just (_,_) -> go ('\123':commentAcc) input docType False
840 Just (c,input) -> go (c:commentAcc) input docType False
842 withLexedDocType lexDocComment = do
843 input@(AI _ _ buf) <- getInput
844 case prevChar buf ' ' of
845 '|' -> lexDocComment input ITdocCommentNext False
846 '^' -> lexDocComment input ITdocCommentPrev False
847 '$' -> lexDocComment input ITdocCommentNamed False
848 '*' -> lexDocSection 1 input
849 '#' -> lexDocComment input ITdocOptionsOld False
851 lexDocSection n input = case alexGetChar input of
852 Just ('*', input) -> lexDocSection (n+1) input
853 Just (_, _) -> lexDocComment input (ITdocSection n) True
854 Nothing -> do setInput input; lexToken -- eof reached, lex it normally
857 -------------------------------------------------------------------------------
858 -- This function is quite tricky. We can't just return a new token, we also
859 -- need to update the state of the parser. Why? Because the token is longer
860 -- than what was lexed by Alex, and the lexToken function doesn't know this, so
861 -- it writes the wrong token length to the parser state. This function is
862 -- called afterwards, so it can just update the state.
864 -- This is complicated by the fact that Haddock tokens can span multiple lines,
865 -- which is something that the original lexer didn't account for.
866 -- I have added last_line_len in the parser state which represents the length
867 -- of the part of the token that is on the last line. It is now used for layout
868 -- calculation in pushCurrentContext instead of last_len. last_len is, like it
869 -- was before, the full length of the token, and it is now only used for error
872 docCommentEnd :: AlexInput -> String -> (String -> Token) -> StringBuffer ->
873 SrcSpan -> P (Located Token)
874 docCommentEnd input commentAcc docType buf span = do
876 let (AI loc last_offs nextBuf) = input
877 comment = reverse commentAcc
878 span' = mkSrcSpan (srcSpanStart span) loc
879 last_len = byteDiff buf nextBuf
881 last_line_len = if (last_offs - last_len < 0)
885 span `seq` setLastToken span' last_len last_line_len
886 return (L span' (docType comment))
888 errBrace (AI end _ _) span = failLocMsgP (srcSpanStart span) end "unterminated `{-'"
890 open_brace, close_brace :: Action
891 open_brace span _str _len = do
893 setContext (NoLayout:ctx)
894 return (L span ITocurly)
895 close_brace span _str _len = do
897 return (L span ITccurly)
899 qvarid buf len = ITqvarid $! splitQualName buf len
900 qconid buf len = ITqconid $! splitQualName buf len
902 splitQualName :: StringBuffer -> Int -> (FastString,FastString)
903 -- takes a StringBuffer and a length, and returns the module name
904 -- and identifier parts of a qualified name. Splits at the *last* dot,
905 -- because of hierarchical module names.
906 splitQualName orig_buf len = split orig_buf orig_buf
909 | orig_buf `byteDiff` buf >= len = done dot_buf
910 | c == '.' = found_dot buf'
911 | otherwise = split buf' dot_buf
913 (c,buf') = nextChar buf
915 -- careful, we might get names like M....
916 -- so, if the character after the dot is not upper-case, this is
917 -- the end of the qualifier part.
918 found_dot buf -- buf points after the '.'
919 | isUpper c = split buf' buf
920 | otherwise = done buf
922 (c,buf') = nextChar buf
925 (lexemeToFastString orig_buf (qual_size - 1),
926 lexemeToFastString dot_buf (len - qual_size))
928 qual_size = orig_buf `byteDiff` dot_buf
932 case lookupUFM reservedWordsFM fs of
933 Just (keyword,0) -> do
935 return (L span keyword)
936 Just (keyword,exts) -> do
937 b <- extension (\i -> exts .&. i /= 0)
938 if b then do maybe_layout keyword
939 return (L span keyword)
940 else return (L span (ITvarid fs))
941 _other -> return (L span (ITvarid fs))
943 fs = lexemeToFastString buf len
945 conid buf len = ITconid fs
946 where fs = lexemeToFastString buf len
948 qvarsym buf len = ITqvarsym $! splitQualName buf len
949 qconsym buf len = ITqconsym $! splitQualName buf len
951 varsym = sym ITvarsym
952 consym = sym ITconsym
954 sym con span buf len =
955 case lookupUFM reservedSymsFM fs of
956 Just (keyword,exts) -> do
958 if b then return (L span keyword)
959 else return (L span $! con fs)
960 _other -> return (L span $! con fs)
962 fs = lexemeToFastString buf len
964 -- Variations on the integral numeric literal.
965 tok_integral :: (Integer -> Token)
966 -> (Integer -> Integer)
967 -- -> (StringBuffer -> StringBuffer) -> (Int -> Int)
969 -> (Integer, (Char->Int)) -> Action
970 tok_integral itint transint transbuf translen (radix,char_to_int) span buf len =
971 return $ L span $ itint $! transint $ parseUnsignedInteger
972 (offsetBytes transbuf buf) (subtract translen len) radix char_to_int
974 -- some conveniences for use with tok_integral
975 tok_num = tok_integral ITinteger
976 tok_primint = tok_integral ITprimint
977 tok_primword = tok_integral ITprimword positive
980 decimal = (10,octDecDigit)
981 octal = (8,octDecDigit)
982 hexadecimal = (16,hexDigit)
984 -- readRational can understand negative rationals, exponents, everything.
985 tok_float str = ITrational $! readRational str
986 tok_primfloat str = ITprimfloat $! readRational str
987 tok_primdouble str = ITprimdouble $! readRational str
989 -- -----------------------------------------------------------------------------
992 -- we're at the first token on a line, insert layout tokens if necessary
994 do_bol span _str _len = do
998 --trace "layout: inserting '}'" $ do
1000 -- do NOT pop the lex state, we might have a ';' to insert
1001 return (L span ITvccurly)
1003 --trace "layout: inserting ';'" $ do
1005 return (L span ITsemi)
1010 -- certain keywords put us in the "layout" state, where we might
1011 -- add an opening curly brace.
1012 maybe_layout ITdo = pushLexState layout_do
1013 maybe_layout ITmdo = pushLexState layout_do
1014 maybe_layout ITof = pushLexState layout
1015 maybe_layout ITlet = pushLexState layout
1016 maybe_layout ITwhere = pushLexState layout
1017 maybe_layout ITrec = pushLexState layout
1018 maybe_layout _ = return ()
1020 -- Pushing a new implicit layout context. If the indentation of the
1021 -- next token is not greater than the previous layout context, then
1022 -- Haskell 98 says that the new layout context should be empty; that is
1023 -- the lexer must generate {}.
1025 -- We are slightly more lenient than this: when the new context is started
1026 -- by a 'do', then we allow the new context to be at the same indentation as
1027 -- the previous context. This is what the 'strict' argument is for.
1029 new_layout_context strict span _buf _len = do
1031 (AI _ offset _) <- getInput
1034 Layout prev_off : _ |
1035 (strict && prev_off >= offset ||
1036 not strict && prev_off > offset) -> do
1037 -- token is indented to the left of the previous context.
1038 -- we must generate a {} sequence now.
1039 pushLexState layout_left
1040 return (L span ITvocurly)
1042 setContext (Layout offset : ctx)
1043 return (L span ITvocurly)
1045 do_layout_left span _buf _len = do
1047 pushLexState bol -- we must be at the start of a line
1048 return (L span ITvccurly)
1050 -- -----------------------------------------------------------------------------
1053 setLine :: Int -> Action
1054 setLine code span buf len = do
1055 let line = parseUnsignedInteger buf len 10 octDecDigit
1056 setSrcLoc (mkSrcLoc (srcSpanFile span) (fromIntegral line - 1) 0)
1057 -- subtract one: the line number refers to the *following* line
1062 setFile :: Int -> Action
1063 setFile code span buf len = do
1064 let file = lexemeToFastString (stepOn buf) (len-2)
1065 setSrcLoc (mkSrcLoc file (srcSpanEndLine span) (srcSpanEndCol span))
1071 -- -----------------------------------------------------------------------------
1072 -- Options, includes and language pragmas.
1074 lex_string_prag :: (String -> Token) -> Action
1075 lex_string_prag mkTok span _buf _len
1076 = do input <- getInput
1080 return (L (mkSrcSpan start end) tok)
1082 = if isString input "#-}"
1083 then do setInput input
1084 return (mkTok (reverse acc))
1085 else case alexGetChar input of
1086 Just (c,i) -> go (c:acc) i
1087 Nothing -> err input
1088 isString _ [] = True
1090 = case alexGetChar i of
1091 Just (c,i') | c == x -> isString i' xs
1093 err (AI end _ _) = failLocMsgP (srcSpanStart span) end "unterminated options pragma"
1096 -- -----------------------------------------------------------------------------
1099 -- This stuff is horrible. I hates it.
1101 lex_string_tok :: Action
1102 lex_string_tok span _buf _len = do
1103 tok <- lex_string ""
1105 return (L (mkSrcSpan (srcSpanStart span) end) tok)
1107 lex_string :: String -> P Token
1110 case alexGetChar' i of
1111 Nothing -> lit_error
1115 magicHash <- extension magicHashEnabled
1119 case alexGetChar' i of
1123 then failMsgP "primitive string literal must contain only characters <= \'\\xFF\'"
1124 else let s' = mkZFastString (reverse s) in
1125 return (ITprimstring s')
1126 -- mkZFastString is a hack to avoid encoding the
1127 -- string in UTF-8. We just want the exact bytes.
1129 return (ITstring (mkFastString (reverse s)))
1131 return (ITstring (mkFastString (reverse s)))
1134 | Just ('&',i) <- next -> do
1135 setInput i; lex_string s
1136 | Just (c,i) <- next, is_space c -> do
1137 setInput i; lex_stringgap s
1138 where next = alexGetChar' i
1144 lex_stringgap s = do
1147 '\\' -> lex_string s
1148 c | is_space c -> lex_stringgap s
1152 lex_char_tok :: Action
1153 -- Here we are basically parsing character literals, such as 'x' or '\n'
1154 -- but, when Template Haskell is on, we additionally spot
1155 -- 'x and ''T, returning ITvarQuote and ITtyQuote respectively,
1156 -- but WIHTOUT CONSUMING the x or T part (the parser does that).
1157 -- So we have to do two characters of lookahead: when we see 'x we need to
1158 -- see if there's a trailing quote
1159 lex_char_tok span _buf _len = do -- We've seen '
1160 i1 <- getInput -- Look ahead to first character
1161 let loc = srcSpanStart span
1162 case alexGetChar' i1 of
1163 Nothing -> lit_error
1165 Just ('\'', i2@(AI end2 _ _)) -> do -- We've seen ''
1166 th_exts <- extension thEnabled
1169 return (L (mkSrcSpan loc end2) ITtyQuote)
1172 Just ('\\', i2@(AI _end2 _ _)) -> do -- We've seen 'backslash
1174 lit_ch <- lex_escape
1175 mc <- getCharOrFail -- Trailing quote
1176 if mc == '\'' then finish_char_tok loc lit_ch
1177 else do setInput i2; lit_error
1179 Just (c, i2@(AI _end2 _ _))
1180 | not (isAny c) -> lit_error
1183 -- We've seen 'x, where x is a valid character
1184 -- (i.e. not newline etc) but not a quote or backslash
1185 case alexGetChar' i2 of -- Look ahead one more character
1186 Nothing -> lit_error
1187 Just ('\'', i3) -> do -- We've seen 'x'
1189 finish_char_tok loc c
1190 _other -> do -- We've seen 'x not followed by quote
1191 -- If TH is on, just parse the quote only
1192 th_exts <- extension thEnabled
1193 let (AI end _ _) = i1
1194 if th_exts then return (L (mkSrcSpan loc end) ITvarQuote)
1195 else do setInput i2; lit_error
1197 finish_char_tok :: SrcLoc -> Char -> P (Located Token)
1198 finish_char_tok loc ch -- We've already seen the closing quote
1199 -- Just need to check for trailing #
1200 = do magicHash <- extension magicHashEnabled
1201 i@(AI end _ _) <- getInput
1202 if magicHash then do
1203 case alexGetChar' i of
1204 Just ('#',i@(AI end _ _)) -> do
1206 return (L (mkSrcSpan loc end) (ITprimchar ch))
1208 return (L (mkSrcSpan loc end) (ITchar ch))
1210 return (L (mkSrcSpan loc end) (ITchar ch))
1212 lex_char :: Char -> AlexInput -> P Char
1215 '\\' -> do setInput inp; lex_escape
1216 c | isAny c -> do setInput inp; return c
1219 isAny c | c > '\x7f' = isPrint c
1220 | otherwise = is_any c
1222 lex_escape :: P Char
1236 '^' -> do c <- getCharOrFail
1237 if c >= '@' && c <= '_'
1238 then return (chr (ord c - ord '@'))
1241 'x' -> readNum is_hexdigit 16 hexDigit
1242 'o' -> readNum is_octdigit 8 octDecDigit
1243 x | is_decdigit x -> readNum2 is_decdigit 10 octDecDigit (octDecDigit x)
1247 case alexGetChar' i of
1248 Nothing -> lit_error
1250 case alexGetChar' i2 of
1251 Nothing -> do setInput i2; lit_error
1253 let str = [c1,c2,c3] in
1254 case [ (c,rest) | (p,c) <- silly_escape_chars,
1255 Just rest <- [maybePrefixMatch p str] ] of
1256 (escape_char,[]):_ -> do
1259 (escape_char,_:_):_ -> do
1264 readNum :: (Char -> Bool) -> Int -> (Char -> Int) -> P Char
1265 readNum is_digit base conv = do
1269 then readNum2 is_digit base conv (conv c)
1270 else do setInput i; lit_error
1272 readNum2 is_digit base conv i = do
1275 where read i input = do
1276 case alexGetChar' input of
1277 Just (c,input') | is_digit c -> do
1278 read (i*base + conv c) input'
1280 if i >= 0 && i <= 0x10FFFF
1281 then do setInput input; return (chr i)
1284 silly_escape_chars = [
1321 -- before calling lit_error, ensure that the current input is pointing to
1322 -- the position of the error in the buffer. This is so that we can report
1323 -- a correct location to the user, but also so we can detect UTF-8 decoding
1324 -- errors if they occur.
1325 lit_error = lexError "lexical error in string/character literal"
1327 getCharOrFail :: P Char
1330 case alexGetChar' i of
1331 Nothing -> lexError "unexpected end-of-file in string/character literal"
1332 Just (c,i) -> do setInput i; return c
1334 -- -----------------------------------------------------------------------------
1337 lex_quasiquote_tok :: Action
1338 lex_quasiquote_tok span buf len = do
1339 let quoter = reverse $ takeWhile (/= '$')
1340 $ reverse $ lexemeToString buf (len - 1)
1341 quoteStart <- getSrcLoc
1342 quote <- lex_quasiquote ""
1344 return (L (mkSrcSpan (srcSpanStart span) end)
1345 (ITquasiQuote (mkFastString quoter,
1346 mkFastString (reverse quote),
1347 mkSrcSpan quoteStart end)))
1349 lex_quasiquote :: String -> P String
1350 lex_quasiquote s = do
1352 case alexGetChar' i of
1353 Nothing -> lit_error
1356 | Just ('|',i) <- next -> do
1357 setInput i; lex_quasiquote ('|' : s)
1358 | Just (']',i) <- next -> do
1359 setInput i; lex_quasiquote (']' : s)
1360 where next = alexGetChar' i
1363 | Just (']',i) <- next -> do
1364 setInput i; return s
1365 where next = alexGetChar' i
1368 setInput i; lex_quasiquote (c : s)
1370 -- -----------------------------------------------------------------------------
1373 warn :: DynFlag -> SDoc -> Action
1374 warn option warning srcspan _buf _len = do
1375 addWarning option srcspan warning
1378 -- -----------------------------------------------------------------------------
1389 SrcSpan -- The start and end of the text span related to
1390 -- the error. Might be used in environments which can
1391 -- show this span, e.g. by highlighting it.
1392 Message -- The error message
1394 data PState = PState {
1395 buffer :: StringBuffer,
1397 messages :: Messages,
1398 last_loc :: SrcSpan, -- pos of previous token
1399 last_offs :: !Int, -- offset of the previous token from the
1400 -- beginning of the current line.
1401 -- \t is equal to 8 spaces.
1402 last_len :: !Int, -- len of previous token
1403 last_line_len :: !Int,
1404 loc :: SrcLoc, -- current loc (end of prev token + 1)
1405 extsBitmap :: !Int, -- bitmap that determines permitted extensions
1406 context :: [LayoutContext],
1409 -- last_loc and last_len are used when generating error messages,
1410 -- and in pushCurrentContext only. Sigh, if only Happy passed the
1411 -- current token to happyError, we could at least get rid of last_len.
1412 -- Getting rid of last_loc would require finding another way to
1413 -- implement pushCurrentContext (which is only called from one place).
1415 newtype P a = P { unP :: PState -> ParseResult a }
1417 instance Monad P where
1423 returnP a = a `seq` (P $ \s -> POk s a)
1425 thenP :: P a -> (a -> P b) -> P b
1426 (P m) `thenP` k = P $ \ s ->
1428 POk s1 a -> (unP (k a)) s1
1429 PFailed span err -> PFailed span err
1431 failP :: String -> P a
1432 failP msg = P $ \s -> PFailed (last_loc s) (text msg)
1434 failMsgP :: String -> P a
1435 failMsgP msg = P $ \s -> PFailed (last_loc s) (text msg)
1437 failLocMsgP :: SrcLoc -> SrcLoc -> String -> P a
1438 failLocMsgP loc1 loc2 str = P $ \_ -> PFailed (mkSrcSpan loc1 loc2) (text str)
1440 failSpanMsgP :: SrcSpan -> SDoc -> P a
1441 failSpanMsgP span msg = P $ \_ -> PFailed span msg
1443 extension :: (Int -> Bool) -> P Bool
1444 extension p = P $ \s -> POk s (p $! extsBitmap s)
1447 getExts = P $ \s -> POk s (extsBitmap s)
1449 setSrcLoc :: SrcLoc -> P ()
1450 setSrcLoc new_loc = P $ \s -> POk s{loc=new_loc} ()
1452 getSrcLoc :: P SrcLoc
1453 getSrcLoc = P $ \s@(PState{ loc=loc }) -> POk s loc
1455 setLastToken :: SrcSpan -> Int -> Int -> P ()
1456 setLastToken loc len line_len = P $ \s -> POk s {
1459 last_line_len=line_len
1462 data AlexInput = AI SrcLoc {-#UNPACK#-}!Int StringBuffer
1464 alexInputPrevChar :: AlexInput -> Char
1465 alexInputPrevChar (AI _ _ buf) = prevChar buf '\n'
1467 alexGetChar :: AlexInput -> Maybe (Char,AlexInput)
1468 alexGetChar (AI loc ofs s)
1470 | otherwise = adj_c `seq` loc' `seq` ofs' `seq` s' `seq`
1471 --trace (show (ord c)) $
1472 Just (adj_c, (AI loc' ofs' s'))
1473 where (c,s') = nextChar s
1474 loc' = advanceSrcLoc loc c
1475 ofs' = advanceOffs c ofs
1483 other_graphic = '\x6'
1486 | c <= '\x06' = non_graphic
1488 -- Alex doesn't handle Unicode, so when Unicode
1489 -- character is encoutered we output these values
1490 -- with the actual character value hidden in the state.
1492 case generalCategory c of
1493 UppercaseLetter -> upper
1494 LowercaseLetter -> lower
1495 TitlecaseLetter -> upper
1496 ModifierLetter -> other_graphic
1497 OtherLetter -> lower -- see #1103
1498 NonSpacingMark -> other_graphic
1499 SpacingCombiningMark -> other_graphic
1500 EnclosingMark -> other_graphic
1501 DecimalNumber -> digit
1502 LetterNumber -> other_graphic
1503 OtherNumber -> other_graphic
1504 ConnectorPunctuation -> other_graphic
1505 DashPunctuation -> other_graphic
1506 OpenPunctuation -> other_graphic
1507 ClosePunctuation -> other_graphic
1508 InitialQuote -> other_graphic
1509 FinalQuote -> other_graphic
1510 OtherPunctuation -> other_graphic
1511 MathSymbol -> symbol
1512 CurrencySymbol -> symbol
1513 ModifierSymbol -> symbol
1514 OtherSymbol -> symbol
1516 _other -> non_graphic
1518 -- This version does not squash unicode characters, it is used when
1520 alexGetChar' :: AlexInput -> Maybe (Char,AlexInput)
1521 alexGetChar' (AI loc ofs s)
1523 | otherwise = c `seq` loc' `seq` ofs' `seq` s' `seq`
1524 --trace (show (ord c)) $
1525 Just (c, (AI loc' ofs' s'))
1526 where (c,s') = nextChar s
1527 loc' = advanceSrcLoc loc c
1528 ofs' = advanceOffs c ofs
1530 advanceOffs :: Char -> Int -> Int
1531 advanceOffs '\n' _ = 0
1532 advanceOffs '\t' offs = (offs `quot` 8 + 1) * 8
1533 advanceOffs _ offs = offs + 1
1535 getInput :: P AlexInput
1536 getInput = P $ \s@PState{ loc=l, last_offs=o, buffer=b } -> POk s (AI l o b)
1538 setInput :: AlexInput -> P ()
1539 setInput (AI l o b) = P $ \s -> POk s{ loc=l, last_offs=o, buffer=b } ()
1541 pushLexState :: Int -> P ()
1542 pushLexState ls = P $ \s@PState{ lex_state=l } -> POk s{lex_state=ls:l} ()
1544 popLexState :: P Int
1545 popLexState = P $ \s@PState{ lex_state=ls:l } -> POk s{ lex_state=l } ls
1547 getLexState :: P Int
1548 getLexState = P $ \s@PState{ lex_state=ls:_ } -> POk s ls
1550 -- for reasons of efficiency, flags indicating language extensions (eg,
1551 -- -fglasgow-exts or -XParr) are represented by a bitmap stored in an unboxed
1554 genericsBit, ffiBit, parrBit :: Int
1555 genericsBit = 0 -- {| and |}
1561 explicitForallBit = 7 -- the 'forall' keyword and '.' symbol
1562 bangPatBit = 8 -- Tells the parser to understand bang-patterns
1563 -- (doesn't affect the lexer)
1564 tyFamBit = 9 -- indexed type families: 'family' keyword and kind sigs
1565 haddockBit = 10 -- Lex and parse Haddock comments
1566 magicHashBit = 11 -- # in both functions and operators
1567 kindSigsBit = 12 -- Kind signatures on type variables
1568 recursiveDoBit = 13 -- mdo
1569 unicodeSyntaxBit = 14 -- the forall symbol, arrow symbols, etc
1570 unboxedTuplesBit = 15 -- (# and #)
1571 standaloneDerivingBit = 16 -- standalone instance deriving declarations
1572 transformComprehensionsBit = 17
1573 qqBit = 18 -- enable quasiquoting
1575 genericsEnabled, ffiEnabled, parrEnabled :: Int -> Bool
1577 genericsEnabled flags = testBit flags genericsBit
1578 ffiEnabled flags = testBit flags ffiBit
1579 parrEnabled flags = testBit flags parrBit
1580 arrowsEnabled flags = testBit flags arrowsBit
1581 thEnabled flags = testBit flags thBit
1582 ipEnabled flags = testBit flags ipBit
1583 explicitForallEnabled flags = testBit flags explicitForallBit
1584 bangPatEnabled flags = testBit flags bangPatBit
1585 tyFamEnabled flags = testBit flags tyFamBit
1586 haddockEnabled flags = testBit flags haddockBit
1587 magicHashEnabled flags = testBit flags magicHashBit
1588 kindSigsEnabled flags = testBit flags kindSigsBit
1589 recursiveDoEnabled flags = testBit flags recursiveDoBit
1590 unicodeSyntaxEnabled flags = testBit flags unicodeSyntaxBit
1591 unboxedTuplesEnabled flags = testBit flags unboxedTuplesBit
1592 standaloneDerivingEnabled flags = testBit flags standaloneDerivingBit
1593 transformComprehensionsEnabled flags = testBit flags transformComprehensionsBit
1594 qqEnabled flags = testBit flags qqBit
1596 -- PState for parsing options pragmas
1598 pragState :: DynFlags -> StringBuffer -> SrcLoc -> PState
1599 pragState dynflags buf loc =
1602 messages = emptyMessages,
1604 last_loc = mkSrcSpan loc loc,
1611 lex_state = [bol, option_prags, 0]
1615 -- create a parse state
1617 mkPState :: StringBuffer -> SrcLoc -> DynFlags -> PState
1618 mkPState buf loc flags =
1622 messages = emptyMessages,
1623 last_loc = mkSrcSpan loc loc,
1628 extsBitmap = fromIntegral bitmap,
1630 lex_state = [bol, 0]
1631 -- we begin in the layout state if toplev_layout is set
1634 bitmap = genericsBit `setBitIf` dopt Opt_Generics flags
1635 .|. ffiBit `setBitIf` dopt Opt_ForeignFunctionInterface flags
1636 .|. parrBit `setBitIf` dopt Opt_PArr flags
1637 .|. arrowsBit `setBitIf` dopt Opt_Arrows flags
1638 .|. thBit `setBitIf` dopt Opt_TemplateHaskell flags
1639 .|. qqBit `setBitIf` dopt Opt_QuasiQuotes flags
1640 .|. ipBit `setBitIf` dopt Opt_ImplicitParams flags
1641 .|. explicitForallBit `setBitIf` dopt Opt_ScopedTypeVariables flags
1642 .|. explicitForallBit `setBitIf` dopt Opt_PolymorphicComponents flags
1643 .|. explicitForallBit `setBitIf` dopt Opt_ExistentialQuantification flags
1644 .|. explicitForallBit `setBitIf` dopt Opt_Rank2Types flags
1645 .|. explicitForallBit `setBitIf` dopt Opt_RankNTypes flags
1646 .|. bangPatBit `setBitIf` dopt Opt_BangPatterns flags
1647 .|. tyFamBit `setBitIf` dopt Opt_TypeFamilies flags
1648 .|. haddockBit `setBitIf` dopt Opt_Haddock flags
1649 .|. magicHashBit `setBitIf` dopt Opt_MagicHash flags
1650 .|. kindSigsBit `setBitIf` dopt Opt_KindSignatures flags
1651 .|. recursiveDoBit `setBitIf` dopt Opt_RecursiveDo flags
1652 .|. unicodeSyntaxBit `setBitIf` dopt Opt_UnicodeSyntax flags
1653 .|. unboxedTuplesBit `setBitIf` dopt Opt_UnboxedTuples flags
1654 .|. standaloneDerivingBit `setBitIf` dopt Opt_StandaloneDeriving flags
1655 .|. transformComprehensionsBit `setBitIf` dopt Opt_TransformListComp flags
1657 setBitIf :: Int -> Bool -> Int
1658 b `setBitIf` cond | cond = bit b
1661 addWarning :: DynFlag -> SrcSpan -> SDoc -> P ()
1662 addWarning option srcspan warning
1663 = P $ \s@PState{messages=(ws,es), dflags=d} ->
1664 let warning' = mkWarnMsg srcspan alwaysQualify warning
1665 ws' = if dopt option d then ws `snocBag` warning' else ws
1666 in POk s{messages=(ws', es)} ()
1668 getMessages :: PState -> Messages
1669 getMessages PState{messages=ms} = ms
1671 getContext :: P [LayoutContext]
1672 getContext = P $ \s@PState{context=ctx} -> POk s ctx
1674 setContext :: [LayoutContext] -> P ()
1675 setContext ctx = P $ \s -> POk s{context=ctx} ()
1678 popContext = P $ \ s@(PState{ buffer = buf, context = ctx,
1679 last_len = len, last_loc = last_loc }) ->
1681 (_:tl) -> POk s{ context = tl } ()
1682 [] -> PFailed last_loc (srcParseErr buf len)
1684 -- Push a new layout context at the indentation of the last token read.
1685 -- This is only used at the outer level of a module when the 'module'
1686 -- keyword is missing.
1687 pushCurrentContext :: P ()
1688 pushCurrentContext = P $ \ s@PState{ last_offs=offs, last_line_len=len, context=ctx } ->
1689 POk s{context = Layout (offs-len) : ctx} ()
1690 --trace ("off: " ++ show offs ++ ", len: " ++ show len) $ POk s{context = Layout (offs-len) : ctx} ()
1692 getOffside :: P Ordering
1693 getOffside = P $ \s@PState{last_offs=offs, context=stk} ->
1694 let ord = case stk of
1695 (Layout n:_) -> compare offs n
1699 -- ---------------------------------------------------------------------------
1700 -- Construct a parse error
1703 :: StringBuffer -- current buffer (placed just after the last token)
1704 -> Int -- length of the previous token
1707 = hcat [ if null token
1708 then ptext (sLit "parse error (possibly incorrect indentation)")
1709 else hcat [ptext (sLit "parse error on input "),
1710 char '`', text token, char '\'']
1712 where token = lexemeToString (offsetBytes (-len) buf) len
1714 -- Report a parse failure, giving the span of the previous token as
1715 -- the location of the error. This is the entry point for errors
1716 -- detected during parsing.
1718 srcParseFail = P $ \PState{ buffer = buf, last_len = len,
1719 last_loc = last_loc } ->
1720 PFailed last_loc (srcParseErr buf len)
1722 -- A lexical error is reported at a particular position in the source file,
1723 -- not over a token range.
1724 lexError :: String -> P a
1727 (AI end _ buf) <- getInput
1728 reportLexError loc end buf str
1730 -- -----------------------------------------------------------------------------
1731 -- This is the top-level function: called from the parser each time a
1732 -- new token is to be read from the input.
1734 lexer :: (Located Token -> P a) -> P a
1736 tok@(L _span _tok__) <- lexToken
1737 -- trace ("token: " ++ show tok__) $ do
1740 lexToken :: P (Located Token)
1742 inp@(AI loc1 _ buf) <- getInput
1745 case alexScanUser exts inp sc of
1747 let span = mkSrcSpan loc1 loc1
1748 setLastToken span 0 0
1749 return (L span ITeof)
1750 AlexError (AI loc2 _ buf) ->
1751 reportLexError loc1 loc2 buf "lexical error"
1752 AlexSkip inp2 _ -> do
1755 AlexToken inp2@(AI end _ buf2) _ t -> do
1757 let span = mkSrcSpan loc1 end
1758 let bytes = byteDiff buf buf2
1759 span `seq` setLastToken span bytes bytes
1762 reportLexError loc1 loc2 buf str
1763 | atEnd buf = failLocMsgP loc1 loc2 (str ++ " at end of input")
1766 c = fst (nextChar buf)
1768 if c == '\0' -- decoding errors are mapped to '\0', see utf8DecodeChar#
1769 then failLocMsgP loc2 loc2 (str ++ " (UTF-8 decoding error)")
1770 else failLocMsgP loc1 loc2 (str ++ " at character " ++ show c)