X-Git-Url: http://git.megacz.com/?a=blobdiff_plain;f=compiler%2Fparser%2FLexer.x;h=525d50bd614c0161ecd52aff5019c6188968f752;hb=43f56cdf26f945d068975fa5a372ca87755d9791;hp=1be9aa3443047a383be041bbfddd8527112a72d2;hpb=dd70a2da99ae6b09115ea80163d993e738c581fe;p=ghc-hetmet.git diff --git a/compiler/parser/Lexer.x b/compiler/parser/Lexer.x index 1be9aa3..525d50b 100644 --- a/compiler/parser/Lexer.x +++ b/compiler/parser/Lexer.x @@ -21,6 +21,18 @@ -- - pragma-end should be only valid in a pragma { +{-# OPTIONS -w #-} +-- The above warning supression flag is a temporary kludge. +-- While working on this module you are encouraged to remove it and fix +-- any warnings in the module. See +-- http://hackage.haskell.org/trac/ghc/wiki/Commentary/CodingStyle#Warnings +-- for details +-- +-- Note that Alex itself generates code with with some unused bindings and +-- without type signatures, so removing the flag might not be possible. + +{-# OPTIONS_GHC -funbox-strict-fields #-} + module Lexer ( Token(..), lexer, pragState, mkPState, PState(..), P(..), ParseResult(..), getSrcLoc, @@ -32,8 +44,6 @@ module Lexer ( addWarning ) where -#include "HsVersions.h" - import Bag import ErrUtils import Outputable @@ -48,19 +58,15 @@ import Util ( maybePrefixMatch, readRational ) import Control.Monad import Data.Bits -import Data.Char ( chr, isSpace ) +import Data.Char ( chr, ord, isSpace ) import Data.Ratio import Debug.Trace -#if __GLASGOW_HASKELL__ >= 605 -import Data.Char ( GeneralCategory(..), generalCategory, isPrint, isUpper ) -#else -import Compat.Unicode ( GeneralCategory(..), generalCategory, isPrint, isUpper ) -#endif +import Unicode ( GeneralCategory(..), generalCategory, isPrint, isUpper ) } $unispace = \x05 -- Trick Alex into handling Unicode. See alexGetChar. -$whitechar = [\ \n\r\f\v\xa0 $unispace] +$whitechar = [\ \n\r\f\v $unispace] $white_no_nl = $whitechar # \n $tab = \t @@ -70,16 +76,16 @@ $decdigit = $ascdigit -- for now, should really be $digit (ToDo) $digit = [$ascdigit $unidigit] $special = [\(\)\,\;\[\]\`\{\}] -$ascsymbol = [\!\#\$\%\&\*\+\.\/\<\=\>\?\@\\\^\|\-\~ \xa1-\xbf \xd7 \xf7] +$ascsymbol = [\!\#\$\%\&\*\+\.\/\<\=\>\?\@\\\^\|\-\~] $unisymbol = \x04 -- Trick Alex into handling Unicode. See alexGetChar. $symbol = [$ascsymbol $unisymbol] # [$special \_\:\"\'] $unilarge = \x01 -- Trick Alex into handling Unicode. See alexGetChar. -$asclarge = [A-Z \xc0-\xd6 \xd8-\xde] +$asclarge = [A-Z] $large = [$asclarge $unilarge] $unismall = \x02 -- Trick Alex into handling Unicode. See alexGetChar. -$ascsmall = [a-z \xdf-\xf6 \xf8-\xff] +$ascsmall = [a-z] $small = [$ascsmall $unismall \_] $unigraphic = \x06 -- Trick Alex into handling Unicode. See alexGetChar. @@ -142,12 +148,12 @@ $tab+ { warn Opt_WarnTabs (text "Tab character") } -- space followed by a Haddock comment symbol (docsym) (in which case we'd -- have a Haddock comment). The rules then munch the rest of the line. -"-- " ~$docsym .* ; +"-- " ~[$docsym \#] .* ; "--" [^$symbol : \ ] .* ; -- Next, match Haddock comments if no -haddock flag -"-- " $docsym .* / { ifExtension (not . haddockEnabled) } ; +"-- " [$docsym \#] .* / { ifExtension (not . haddockEnabled) } ; -- Now, when we've matched comments that begin with 2 dashes and continue -- with a different character, we need to match comments that begin with three @@ -250,9 +256,6 @@ $tab+ { warn Opt_WarnTabs (text "Tab character") } "{-#" $whitechar* (CORE|core) { token ITcore_prag } "{-#" $whitechar* (UNPACK|unpack) { token ITunpack_prag } - "{-#" $whitechar* (DOCOPTIONS|docoptions) - / { ifExtension haddockEnabled } { lex_string_prag ITdocOptions } - "{-#" { nested_comment lexToken } -- ToDo: should only be valid inside a pragma: @@ -260,11 +263,18 @@ $tab+ { warn Opt_WarnTabs (text "Tab character") } } { - "{-#" $whitechar* (OPTIONS|options) { lex_string_prag IToptions_prag } - "{-#" $whitechar* (OPTIONS_GHC|options_ghc) + "{-#" $whitechar* (OPTIONS|options) { lex_string_prag IToptions_prag } + "{-#" $whitechar* (OPTIONS_GHC|options_ghc) { lex_string_prag IToptions_prag } - "{-#" $whitechar* (LANGUAGE|language) { token ITlanguage_prag } - "{-#" $whitechar* (INCLUDE|include) { lex_string_prag ITinclude_prag } + "{-#" $whitechar* (OPTIONS_HADDOCK|options_haddock) + { lex_string_prag ITdocOptions } + "-- #" { multiline_doc_comment } + "{-#" $whitechar* (LANGUAGE|language) { token ITlanguage_prag } + "{-#" $whitechar* (INCLUDE|include) { lex_string_prag ITinclude_prag } +} + +<0> { + "-- #" .* ; } <0,option_prags> { @@ -277,8 +287,8 @@ $tab+ { warn Opt_WarnTabs (text "Tab character") } -- Haddock comments <0> { - "-- " $docsym / { ifExtension haddockEnabled } { multiline_doc_comment } - "{-" \ ? $docsym / { ifExtension haddockEnabled } { nested_doc_comment } + "-- " $docsym / { ifExtension haddockEnabled } { multiline_doc_comment } + "{-" \ ? $docsym / { ifExtension haddockEnabled } { nested_doc_comment } } -- "special" symbols @@ -297,6 +307,9 @@ $tab+ { warn Opt_WarnTabs (text "Tab character") } "|]" / { ifExtension thEnabled } { token ITcloseQuote } \$ @varid / { ifExtension thEnabled } { skip_one_varid ITidEscape } "$(" / { ifExtension thEnabled } { token ITparenEscape } + + "[$" @varid "|" / { ifExtension qqEnabled } + { lex_quasiquote_tok } } <0> { @@ -361,25 +374,29 @@ $tab+ { warn Opt_WarnTabs (text "Tab character") } -- when trying to be close to Haskell98 <0> { -- Normal integral literals (:: Num a => a, from Integer) - @decimal { tok_num positive 0 0 decimal } - 0[oO] @octal { tok_num positive 2 2 octal } - 0[xX] @hexadecimal { tok_num positive 2 2 hexadecimal } + @decimal { tok_num positive 0 0 decimal } + 0[oO] @octal { tok_num positive 2 2 octal } + 0[xX] @hexadecimal { tok_num positive 2 2 hexadecimal } -- Normal rational literals (:: Fractional a => a, from Rational) - @floating_point { strtoken tok_float } + @floating_point { strtoken tok_float } } <0> { - -- Unboxed ints (:: Int#) + -- Unboxed ints (:: Int#) and words (:: Word#) -- It's simpler (and faster?) to give separate cases to the negatives, -- especially considering octal/hexadecimal prefixes. - @decimal \# / { ifExtension magicHashEnabled } { tok_primint positive 0 1 decimal } - 0[oO] @octal \# / { ifExtension magicHashEnabled } { tok_primint positive 2 3 octal } - 0[xX] @hexadecimal \# / { ifExtension magicHashEnabled } { tok_primint positive 2 3 hexadecimal } - @negative @decimal \# / { ifExtension magicHashEnabled } { tok_primint negative 1 2 decimal } - @negative 0[oO] @octal \# / { ifExtension magicHashEnabled } { tok_primint negative 3 4 octal } + @decimal \# / { ifExtension magicHashEnabled } { tok_primint positive 0 1 decimal } + 0[oO] @octal \# / { ifExtension magicHashEnabled } { tok_primint positive 2 3 octal } + 0[xX] @hexadecimal \# / { ifExtension magicHashEnabled } { tok_primint positive 2 3 hexadecimal } + @negative @decimal \# / { ifExtension magicHashEnabled } { tok_primint negative 1 2 decimal } + @negative 0[oO] @octal \# / { ifExtension magicHashEnabled } { tok_primint negative 3 4 octal } @negative 0[xX] @hexadecimal \# / { ifExtension magicHashEnabled } { tok_primint negative 3 4 hexadecimal } + @decimal \# \# / { ifExtension magicHashEnabled } { tok_primword 0 2 decimal } + 0[oO] @octal \# \# / { ifExtension magicHashEnabled } { tok_primword 2 4 octal } + 0[xX] @hexadecimal \# \# / { ifExtension magicHashEnabled } { tok_primword 2 4 hexadecimal } + -- Unboxed floats and doubles (:: Float#, :: Double#) -- prim_{float,double} work with signed literals @signed @floating_point \# / { ifExtension magicHashEnabled } { init_strtoken 1 tok_primfloat } @@ -439,6 +456,9 @@ data Token | ITdotnet | ITmdo | ITfamily + | ITgroup + | ITby + | ITusing -- Pragmas | ITinline_prag Bool -- True <=> INLINE, False <=> NOINLINE @@ -482,8 +502,8 @@ data Token | ITvocurly | ITvccurly | ITobrack - | ITopabrack -- [:, for parallel arrays with -fparr - | ITcpabrack -- :], for parallel arrays with -fparr + | ITopabrack -- [:, for parallel arrays with -XParr + | ITcpabrack -- :], for parallel arrays with -XParr | ITcbrack | IToparen | ITcparen @@ -515,6 +535,7 @@ data Token | ITprimchar Char | ITprimstring FastString | ITprimint Integer + | ITprimword Integer | ITprimfloat Rational | ITprimdouble Rational @@ -528,6 +549,7 @@ data Token | ITparenEscape -- $( | ITvarQuote -- ' | ITtyQuote -- '' + | ITquasiQuote (FastString,FastString,SrcSpan) -- [:...|...|] -- Arrow notation extension | ITproc @@ -548,11 +570,13 @@ data Token | ITdocCommentNamed String -- something beginning '-- $' | ITdocSection Int String -- a section heading | ITdocOptions String -- doc options (prune, ignore-exports, etc) + | ITdocOptionsOld String -- doc options declared "-- # ..."-style #ifdef DEBUG deriving Show -- debugging #endif +{- isSpecial :: Token -> Bool -- If we see M.x, where x is a keyword, but -- is special, we treat is as just plain M.x, @@ -571,7 +595,11 @@ isSpecial ITccallconv = True isSpecial ITstdcallconv = True isSpecial ITmdo = True isSpecial ITfamily = True +isSpecial ITgroup = True +isSpecial ITby = True +isSpecial ITusing = True isSpecial _ = False +-} -- the bitmap provided as the third component indicates whether the -- corresponding extension keyword is valid under the extension options @@ -609,9 +637,12 @@ reservedWordsFM = listToUFM $ ( "where", ITwhere, 0 ), ( "_scc_", ITscc, 0 ), -- ToDo: remove - ( "forall", ITforall, bit explicitForallBit), + ( "forall", ITforall, bit explicitForallBit), ( "mdo", ITmdo, bit recursiveDoBit), ( "family", ITfamily, bit tyFamBit), + ( "group", ITgroup, bit transformComprehensionsBit), + ( "by", ITby, bit transformComprehensionsBit), + ( "using", ITusing, bit transformComprehensionsBit), ( "foreign", ITforeign, bit ffiBit), ( "export", ITexport, bit ffiBit), @@ -676,11 +707,11 @@ reservedSymsFM = listToUFM $ type Action = SrcSpan -> StringBuffer -> Int -> P (Located Token) special :: Token -> Action -special tok span _buf len = return (L span tok) +special tok span _buf _len = return (L span tok) token, layout_token :: Token -> Action -token t span buf len = return (L span t) -layout_token t span buf len = pushLexState layout >> return (L span t) +token t span _buf _len = return (L span t) +layout_token t span _buf _len = pushLexState layout >> return (L span t) idtoken :: (StringBuffer -> Int -> Token) -> Action idtoken f span buf len = return (L span $! (f buf len)) @@ -730,8 +761,10 @@ isNormalComment bits _ _ (AI _ _ buf) spaceAndP buf p = p buf || nextCharIs buf (==' ') && p (snd (nextChar buf)) +{- haddockDisabledAnd p bits _ _ (AI _ _ buf) = if haddockEnabled bits then False else (p buf) +-} atEOL _ _ _ (AI _ _ buf) = atEnd buf || currentChar buf == '\n' @@ -780,12 +813,12 @@ nested_comment cont span _str _len = do Just ('-',input) -> case alexGetChar input of Nothing -> errBrace input span Just ('\125',input) -> go (n-1) input - Just (c,_) -> go n input + Just (_,_) -> go n input Just ('\123',input) -> case alexGetChar input of Nothing -> errBrace input span Just ('-',input) -> go (n+1) input - Just (c,_) -> go n input - Just (c,input) -> go n input + Just (_,_) -> go n input + Just (_,input) -> go n input nested_doc_comment :: Action nested_doc_comment span buf _len = withLexedDocType (go "") @@ -794,16 +827,16 @@ nested_doc_comment span buf _len = withLexedDocType (go "") Nothing -> errBrace input span Just ('-',input) -> case alexGetChar input of Nothing -> errBrace input span - Just ('\125',input@(AI end _ buf2)) -> + Just ('\125',input) -> docCommentEnd input commentAcc docType buf span - Just (c,_) -> go ('-':commentAcc) input docType False + Just (_,_) -> go ('-':commentAcc) input docType False Just ('\123', input) -> case alexGetChar input of Nothing -> errBrace input span Just ('-',input) -> do setInput input let cont = do input <- getInput; go commentAcc input docType False nested_comment cont span buf _len - Just (c,_) -> go ('\123':commentAcc) input docType False + Just (_,_) -> go ('\123':commentAcc) input docType False Just (c,input) -> go (c:commentAcc) input docType False withLexedDocType lexDocComment = do @@ -812,11 +845,12 @@ withLexedDocType lexDocComment = do '|' -> lexDocComment input ITdocCommentNext False '^' -> lexDocComment input ITdocCommentPrev False '$' -> lexDocComment input ITdocCommentNamed False - '*' -> lexDocSection 1 input + '*' -> lexDocSection 1 input + '#' -> lexDocComment input ITdocOptionsOld False where lexDocSection n input = case alexGetChar input of Just ('*', input) -> lexDocSection (n+1) input - Just (c, _) -> lexDocComment input (ITdocSection n) True + Just (_, _) -> lexDocComment input (ITdocSection n) True Nothing -> do setInput input; lexToken -- eof reached, lex it normally -- docCommentEnd @@ -894,6 +928,7 @@ splitQualName orig_buf len = split orig_buf orig_buf qual_size = orig_buf `byteDiff` dot_buf varid span buf len = + fs `seq` case lookupUFM reservedWordsFM fs of Just (keyword,0) -> do maybe_layout keyword @@ -939,6 +974,7 @@ tok_integral itint transint transbuf translen (radix,char_to_int) span buf len = -- some conveniences for use with tok_integral tok_num = tok_integral ITinteger tok_primint = tok_integral ITprimint +tok_primword = tok_integral ITprimword positive positive = id negative = negate decimal = (10,octDecDigit) @@ -1002,7 +1038,7 @@ new_layout_context strict span _buf _len = do -- we must generate a {} sequence now. pushLexState layout_left return (L span ITvocurly) - other -> do + _ -> do setContext (Layout offset : ctx) return (L span ITvocurly) @@ -1036,7 +1072,7 @@ setFile code span buf len = do -- Options, includes and language pragmas. lex_string_prag :: (String -> Token) -> Action -lex_string_prag mkTok span buf len +lex_string_prag mkTok span _buf _len = do input <- getInput start <- getSrcLoc tok <- go [] input @@ -1049,7 +1085,7 @@ lex_string_prag mkTok span buf len else case alexGetChar input of Just (c,i) -> go (c:acc) i Nothing -> err input - isString i [] = True + isString _ [] = True isString i (x:xs) = case alexGetChar i of Just (c,i') | c == x -> isString i' xs @@ -1063,7 +1099,7 @@ lex_string_prag mkTok span buf len -- This stuff is horrible. I hates it. lex_string_tok :: Action -lex_string_tok span buf len = do +lex_string_tok span _buf _len = do tok <- lex_string "" end <- getSrcLoc return (L (mkSrcSpan (srcSpanStart span) end) tok) @@ -1120,7 +1156,7 @@ lex_char_tok :: Action -- but WIHTOUT CONSUMING the x or T part (the parser does that). -- So we have to do two characters of lookahead: when we see 'x we need to -- see if there's a trailing quote -lex_char_tok span buf len = do -- We've seen ' +lex_char_tok span _buf _len = do -- We've seen ' i1 <- getInput -- Look ahead to first character let loc = srcSpanStart span case alexGetChar' i1 of @@ -1133,14 +1169,14 @@ lex_char_tok span buf len = do -- We've seen ' return (L (mkSrcSpan loc end2) ITtyQuote) else lit_error - Just ('\\', i2@(AI end2 _ _)) -> do -- We've seen 'backslash + Just ('\\', i2@(AI _end2 _ _)) -> do -- We've seen 'backslash setInput i2 lit_ch <- lex_escape mc <- getCharOrFail -- Trailing quote if mc == '\'' then finish_char_tok loc lit_ch else do setInput i2; lit_error - Just (c, i2@(AI end2 _ _)) + Just (c, i2@(AI _end2 _ _)) | not (isAny c) -> lit_error | otherwise -> @@ -1180,7 +1216,7 @@ lex_char c inp = do c | isAny c -> do setInput inp; return c _other -> lit_error -isAny c | c > '\xff' = isPrint c +isAny c | c > '\x7f' = isPrint c | otherwise = is_any c lex_escape :: P Char @@ -1204,7 +1240,7 @@ lex_escape = do 'x' -> readNum is_hexdigit 16 hexDigit 'o' -> readNum is_octdigit 8 octDecDigit - x | is_digit x -> readNum2 is_digit 10 octDecDigit (octDecDigit x) + x | is_decdigit x -> readNum2 is_decdigit 10 octDecDigit (octDecDigit x) c1 -> do i <- getInput @@ -1296,6 +1332,42 @@ getCharOrFail = do Just (c,i) -> do setInput i; return c -- ----------------------------------------------------------------------------- +-- QuasiQuote + +lex_quasiquote_tok :: Action +lex_quasiquote_tok span buf len = do + let quoter = reverse $ takeWhile (/= '$') + $ reverse $ lexemeToString buf (len - 1) + quoteStart <- getSrcLoc + quote <- lex_quasiquote "" + end <- getSrcLoc + return (L (mkSrcSpan (srcSpanStart span) end) + (ITquasiQuote (mkFastString quoter, + mkFastString (reverse quote), + mkSrcSpan quoteStart end))) + +lex_quasiquote :: String -> P String +lex_quasiquote s = do + i <- getInput + case alexGetChar' i of + Nothing -> lit_error + + Just ('\\',i) + | Just ('|',i) <- next -> do + setInput i; lex_quasiquote ('|' : s) + | Just (']',i) <- next -> do + setInput i; lex_quasiquote (']' : s) + where next = alexGetChar' i + + Just ('|',i) + | Just (']',i) <- next -> do + setInput i; return s + where next = alexGetChar' i + + Just (c, i) -> do + setInput i; lex_quasiquote (c : s) + +-- ----------------------------------------------------------------------------- -- Warnings warn :: DynFlag -> SDoc -> Action @@ -1348,7 +1420,7 @@ instance Monad P where fail = failP returnP :: a -> P a -returnP a = P $ \s -> POk s a +returnP a = a `seq` (P $ \s -> POk s a) thenP :: P a -> (a -> P b) -> P b (P m) `thenP` k = P $ \ s -> @@ -1363,10 +1435,10 @@ failMsgP :: String -> P a failMsgP msg = P $ \s -> PFailed (last_loc s) (text msg) failLocMsgP :: SrcLoc -> SrcLoc -> String -> P a -failLocMsgP loc1 loc2 str = P $ \s -> PFailed (mkSrcSpan loc1 loc2) (text str) +failLocMsgP loc1 loc2 str = P $ \_ -> PFailed (mkSrcSpan loc1 loc2) (text str) -failSpanMsgP :: SrcSpan -> String -> P a -failSpanMsgP span msg = P $ \s -> PFailed span (text msg) +failSpanMsgP :: SrcSpan -> SDoc -> P a +failSpanMsgP span msg = P $ \_ -> PFailed span msg extension :: (Int -> Bool) -> P Bool extension p = P $ \s -> POk s (p $! extsBitmap s) @@ -1412,7 +1484,7 @@ alexGetChar (AI loc ofs s) adj_c | c <= '\x06' = non_graphic - | c <= '\xff' = c + | c <= '\x7f' = c -- Alex doesn't handle Unicode, so when Unicode -- character is encoutered we output these values -- with the actual character value hidden in the state. @@ -1422,7 +1494,7 @@ alexGetChar (AI loc ofs s) LowercaseLetter -> lower TitlecaseLetter -> upper ModifierLetter -> other_graphic - OtherLetter -> other_graphic + OtherLetter -> lower -- see #1103 NonSpacingMark -> other_graphic SpacingCombiningMark -> other_graphic EnclosingMark -> other_graphic @@ -1456,7 +1528,7 @@ alexGetChar' (AI loc ofs s) ofs' = advanceOffs c ofs advanceOffs :: Char -> Int -> Int -advanceOffs '\n' offs = 0 +advanceOffs '\n' _ = 0 advanceOffs '\t' offs = (offs `quot` 8 + 1) * 8 advanceOffs _ offs = offs + 1 @@ -1473,10 +1545,10 @@ popLexState :: P Int popLexState = P $ \s@PState{ lex_state=ls:l } -> POk s{ lex_state=l } ls getLexState :: P Int -getLexState = P $ \s@PState{ lex_state=ls:l } -> POk s ls +getLexState = P $ \s@PState{ lex_state=ls:_ } -> POk s ls -- for reasons of efficiency, flags indicating language extensions (eg, --- -fglasgow-exts or -fparr) are represented by a bitmap stored in an unboxed +-- -fglasgow-exts or -XParr) are represented by a bitmap stored in an unboxed -- integer genericsBit, ffiBit, parrBit :: Int @@ -1497,6 +1569,8 @@ recursiveDoBit = 13 -- mdo unicodeSyntaxBit = 14 -- the forall symbol, arrow symbols, etc unboxedTuplesBit = 15 -- (# and #) standaloneDerivingBit = 16 -- standalone instance deriving declarations +transformComprehensionsBit = 17 +qqBit = 18 -- enable quasiquoting genericsEnabled, ffiEnabled, parrEnabled :: Int -> Bool always _ = True @@ -1516,17 +1590,17 @@ recursiveDoEnabled flags = testBit flags recursiveDoBit unicodeSyntaxEnabled flags = testBit flags unicodeSyntaxBit unboxedTuplesEnabled flags = testBit flags unboxedTuplesBit standaloneDerivingEnabled flags = testBit flags standaloneDerivingBit +transformComprehensionsEnabled flags = testBit flags transformComprehensionsBit +qqEnabled flags = testBit flags qqBit -- PState for parsing options pragmas -- -pragState :: StringBuffer -> SrcLoc -> PState -pragState buf loc = +pragState :: DynFlags -> StringBuffer -> SrcLoc -> PState +pragState dynflags buf loc = PState { - buffer = buf, + buffer = buf, messages = emptyMessages, - -- XXX defaultDynFlags is not right, but we don't have a real - -- dflags handy - dflags = defaultDynFlags, + dflags = dynflags, last_loc = mkSrcSpan loc loc, last_offs = 0, last_len = 0, @@ -1562,6 +1636,7 @@ mkPState buf loc flags = .|. parrBit `setBitIf` dopt Opt_PArr flags .|. arrowsBit `setBitIf` dopt Opt_Arrows flags .|. thBit `setBitIf` dopt Opt_TemplateHaskell flags + .|. qqBit `setBitIf` dopt Opt_QuasiQuotes flags .|. ipBit `setBitIf` dopt Opt_ImplicitParams flags .|. explicitForallBit `setBitIf` dopt Opt_ScopedTypeVariables flags .|. explicitForallBit `setBitIf` dopt Opt_PolymorphicComponents flags @@ -1577,6 +1652,7 @@ mkPState buf loc flags = .|. unicodeSyntaxBit `setBitIf` dopt Opt_UnicodeSyntax flags .|. unboxedTuplesBit `setBitIf` dopt Opt_UnboxedTuples flags .|. standaloneDerivingBit `setBitIf` dopt Opt_StandaloneDeriving flags + .|. transformComprehensionsBit `setBitIf` dopt Opt_TransformListComp flags -- setBitIf :: Int -> Bool -> Int b `setBitIf` cond | cond = bit b @@ -1600,7 +1676,7 @@ setContext ctx = P $ \s -> POk s{context=ctx} () popContext :: P () popContext = P $ \ s@(PState{ buffer = buf, context = ctx, - loc = loc, last_len = len, last_loc = last_loc }) -> + last_len = len, last_loc = last_loc }) -> case ctx of (_:tl) -> POk s{ context = tl } () [] -> PFailed last_loc (srcParseErr buf len) @@ -1629,8 +1705,8 @@ srcParseErr -> Message srcParseErr buf len = hcat [ if null token - then ptext SLIT("parse error (possibly incorrect indentation)") - else hcat [ptext SLIT("parse error on input "), + then ptext (sLit "parse error (possibly incorrect indentation)") + else hcat [ptext (sLit "parse error on input "), char '`', text token, char '\''] ] where token = lexemeToString (offsetBytes (-len) buf) len @@ -1648,7 +1724,7 @@ srcParseFail = P $ \PState{ buffer = buf, last_len = len, lexError :: String -> P a lexError str = do loc <- getSrcLoc - i@(AI end _ buf) <- getInput + (AI end _ buf) <- getInput reportLexError loc end buf str -- ----------------------------------------------------------------------------- @@ -1657,7 +1733,7 @@ lexError str = do lexer :: (Located Token -> P a) -> P a lexer cont = do - tok@(L span tok__) <- lexToken + tok@(L _span _tok__) <- lexToken -- trace ("token: " ++ show tok__) $ do cont tok @@ -1667,20 +1743,21 @@ lexToken = do sc <- getLexState exts <- getExts case alexScanUser exts inp sc of - AlexEOF -> do let span = mkSrcSpan loc1 loc1 - setLastToken span 0 0 - return (L span ITeof) - AlexError (AI loc2 _ buf) -> do - reportLexError loc1 loc2 buf "lexical error" + AlexEOF -> do + let span = mkSrcSpan loc1 loc1 + setLastToken span 0 0 + return (L span ITeof) + AlexError (AI loc2 _ buf) -> + reportLexError loc1 loc2 buf "lexical error" AlexSkip inp2 _ -> do - setInput inp2 - lexToken - AlexToken inp2@(AI end _ buf2) len t -> do - setInput inp2 - let span = mkSrcSpan loc1 end - let bytes = byteDiff buf buf2 - span `seq` setLastToken span bytes bytes - t span buf bytes + setInput inp2 + lexToken + AlexToken inp2@(AI end _ buf2) _ t -> do + setInput inp2 + let span = mkSrcSpan loc1 end + let bytes = byteDiff buf buf2 + span `seq` setLastToken span bytes bytes + t span buf bytes reportLexError loc1 loc2 buf str | atEnd buf = failLocMsgP loc1 loc2 (str ++ " at end of input")