2 -- Haddock - A Haskell Documentation Tool
4 -- (c) Simon Marlow 2002
6 -- This file was modified and integrated into GHC by David Waern 2006
11 -- The above warning supression flag is a temporary kludge.
12 -- While working on this module you are encouraged to remove it and fix
13 -- any warnings in the module. See
14 -- http://hackage.haskell.org/trac/ghc/wiki/Commentary/CodingStyle#Warnings
22 import Lexer hiding (Token)
23 import Parser ( parseIdentifier )
31 import System.IO.Unsafe
36 $hexdigit = [0-9a-fA-F]
38 $alphanum = [A-Za-z0-9]
39 $ident = [$alphanum \'\_\.\!\#\$\%\&\*\+\/\<\=\>\?\@\\\\\^\|\-\~]
43 -- beginning of a paragraph
46 $ws* \> { begin birdtrack }
47 $ws* [\*\-] { token TokBullet `andBegin` string }
48 $ws* \[ { token TokDefStart `andBegin` def }
49 $ws* \( $digit+ \) { token TokNumber `andBegin` string }
53 -- beginning of a line
55 $ws* \> { begin birdtrack }
56 $ws* \n { token TokPara `andBegin` para }
57 -- Here, we really want to be able to say
58 -- $ws* (\n | <eof>) { token TokPara `andBegin` para}
59 -- because otherwise a trailing line of whitespace will result in
60 -- a spurious TokString at the end of a docstring. We don't have <eof>,
61 -- though (NOW I realise what it was for :-). To get around this, we always
62 -- append \n to the end of a docstring.
66 <birdtrack> .* \n? { strtoken TokBirdTrack `andBegin` line }
69 $special { strtoken $ \s -> TokSpecial (head s) }
70 \<.*\> { strtoken $ \s -> TokURL (init (tail s)) }
71 \#.*\# { strtoken $ \s -> TokAName (init (tail s)) }
72 [\'\`] $ident+ [\'\`] { ident }
73 \\ . { strtoken (TokString . tail) }
74 "&#" $digit+ \; { strtoken $ \s -> TokString [chr (read (init (drop 2 s)))] }
75 "&#" [xX] $hexdigit+ \; { strtoken $ \s -> case readHex (init (drop 3 s)) of [(n,_)] -> TokString [chr n] }
76 -- allow special characters through if they don't fit one of the previous
78 [\'\`\<\#\&\\] { strtoken TokString }
79 [^ $special \< \# \n \'\` \& \\ \]]* \n { strtoken TokString `andBegin` line }
80 [^ $special \< \# \n \'\` \& \\ \]]+ { strtoken TokString }
84 \] { token TokDefEnd `andBegin` string }
87 -- ']' doesn't have any special meaning outside of the [...] at the beginning
88 -- of a definition paragraph.
90 \] { strtoken TokString }
105 | TokBirdTrack String
108 -- -----------------------------------------------------------------------------
109 -- Alex support stuff
112 type Action = String -> StartCode -> (StartCode -> [Token]) -> [Token]
114 type AlexInput = (Char,String)
116 alexGetChar (_, []) = Nothing
117 alexGetChar (_, c:cs) = Just (c, (c,cs))
119 alexInputPrevChar (c,_) = c
121 tokenise :: String -> [Token]
122 tokenise str = let toks = go ('\n', eofHack str) para in {-trace (show toks)-} toks
123 where go inp@(_,str) sc =
124 case alexScan inp sc of
126 AlexError _ -> error "lexical error"
127 AlexSkip inp' len -> go inp' sc
128 AlexToken inp' len act -> act (take len str) sc (\sc -> go inp' sc)
130 -- NB. we add a final \n to the string, (see comment in the beginning of line
131 -- production above).
132 eofHack str = str++"\n"
134 andBegin :: Action -> StartCode -> Action
135 andBegin act new_sc = \str sc cont -> act str new_sc cont
137 token :: Token -> Action
138 token t = \str sc cont -> t : cont sc
140 strtoken :: (String -> Token) -> Action
141 strtoken t = \str sc cont -> t str : cont sc
143 begin :: StartCode -> Action
144 begin sc = \str _ cont -> cont sc
146 -- -----------------------------------------------------------------------------
147 -- Lex a string as a Haskell identifier
151 case strToHsQNames id of
152 Just names -> TokIdent names : cont sc
153 Nothing -> TokString str : cont sc
154 where id = init (tail str)
156 strToHsQNames :: String -> Maybe [RdrName]
158 let buffer = unsafePerformIO (stringToStringBuffer str0)
159 pstate = mkPState buffer noSrcLoc defaultDynFlags
160 lex = lexer (\t -> return t)
161 result = unP parseIdentifier pstate
163 POk _ name -> Just [unLoc name]