4b4060b · typeable/pandoc

git.iko.soy/typeable/pandoc

Simplified parsing of reference keys and notes in markdown and RST

readers: + The Reference data structure from Text.Pandoc.Shared is no longer needed, since + referenceKey and noteBlock parses return strings (as many blank lines as are occuried by the key or note) and update state themselves. + getPosition and setPosition are now used to ensure that error messages will give the correct line number. + This yields cleaner (and slightly faster) code, with more accurate parsing error messages. git-svn-id: https://pandoc.googlecode.com/svn/trunk@1012 788f1e2b-df1e-0410-8736-df70ead52e1b

commit4b4060b8ef64da355070ee016235ad3f1d4b1f9f authorfiddlosopher <fiddlosopher@788f1e2b-df1e-0410-8736-df70ead52e1b> dateWed 5 Sep 2007, 23:55 (19 years ago) parentd5e8ab7

3 files changed+54−57

compare with master

unifiedsplit

src/Text/Pandoc/Readers/Markdown.hs+30−18 src/Text/Pandoc/Readers/RST.hs+24−15 src/Text/Pandoc/Shared.hs−24

src/Text/Pandoc/Readers/Markdown.hs+30−18

@@ -140,26 +140,24 @@ titleBlock = try $ do

140140parseMarkdown = do

141141 -- markdown allows raw HTML

142142 updateState (\state -> state { stateParseRaw = True })

143− (title, author, date) <- option ([],[],"") titleBlock

143+ startPos <- getPosition

144144 -- go through once just to get list of reference keys

145− refs <- manyTill (referenceKey <|> (lineClump >>= return . LineClump)) eof

146− let keys = map (\(KeyBlock label target) -> (label, target)) $

147− filter isKeyBlock refs

148− -- strip out keys

149− setInput $ concatMap (\(LineClump ln) -> ln) $ filter isLineClump refs

150− updateState (\state -> state { stateKeys = keys })

145+ -- docMinusKeys is the raw document with blanks where the keys were...

146+ docMinusKeys <- many (referenceKey <|> lineClump) >>= return . concat

147+ setInput docMinusKeys

148+ setPosition startPos

151149 st <- getState

150+ -- go through again for notes unless strict...

152151 if stateStrict st

153152 then return ()

154− else do -- go through for notes (which may contain refs - hence 2nd pass)

155− refs' <- manyTill (noteBlock <|>

156− (lineClump >>= return . LineClump)) eof

157− let notes = map (\(NoteBlock label blocks) -> (label, blocks)) $

158− filter isNoteBlock refs'

159− updateState (\state -> state { stateNotes = notes })

160− setInput $ concatMap (\(LineClump ln) -> ln) $

161− filter isLineClump refs'

162− -- go through again, with note blocks and keys stripped out

153+ else do docMinusNotes <- many (noteBlock <|> lineClump) >>= return . concat

154+ st <- getState

155+ let reversedNotes = stateNotes st

156+ updateState $ \st -> st { stateNotes = reverse reversedNotes }

157+ setInput docMinusNotes

158+ setPosition startPos

159+ -- now parse it for real...

160+ (title, author, date) <- option ([],[],"") titleBlock

163161 blocks <- parseBlocks

164162 return $ Pandoc (Meta title author date) $ filter (/= Null) blocks

165163

@@ -168,6 +166,7 @@ parseMarkdown = do

168166--

169167

170168referenceKey = try $ do

169+ startPos <- getPosition

171170 nonindentSpaces

172171 label <- reference

173172 char ':'

@@ -177,7 +176,13 @@ referenceKey = try $ do

177176 optional (char '>')

178177 tit <- option "" referenceTitle

179178 blanklines

180− return $ KeyBlock label (removeTrailingSpace src, tit)

179+ endPos <- getPosition

180+ let newkey = (label, (removeTrailingSpace src, tit))

181+ st <- getState

182+ let oldkeys = stateKeys st

183+ updateState $ \st -> st { stateKeys = newkey : oldkeys }

184+ -- return blanks so line count isn't affected

185+ return $ replicate (sourceLine endPos - sourceLine startPos) '\n'

181186

182187referenceTitle = try $ do

183188 skipSpaces

@@ -201,15 +206,22 @@ rawLine = do

201206rawLines = many1 rawLine >>= return . concat

202207

203208noteBlock = try $ do

209+ startPos <- getPosition

204210 ref <- noteMarker

205211 char ':'

206212 optional blankline

207213 optional indentSpaces

208214 raw <- sepBy rawLines (try (blankline >> indentSpaces))

209215 optional blanklines

216+ endPos <- getPosition

210217 -- parse the extracted text, which may contain various block elements:

211218 contents <- parseFromString parseBlocks $ (joinWithSep "\n" raw) ++ "\n\n"

212− return $ NoteBlock ref contents

219+ let newnote = (ref, contents)

220+ st <- getState

221+ let oldnotes = stateNotes st

222+ updateState $ \st -> st { stateNotes = newnote : oldnotes }

223+ -- return blanks so line count isn't affected

224+ return $ replicate (sourceLine endPos - sourceLine startPos) '\n'

213225

214226--

215227-- parsing blocks

src/Text/Pandoc/Readers/RST.hs+24−15

@@ -83,15 +83,17 @@ titleTransform ((Header 1 head1):rest) = -- title, no subtitle

8383titleTransform blocks = (blocks, [])

8484

8585parseRST = do

86− -- first pass: get keys

87− refs <- manyTill (referenceKey <|> (lineClump >>= return . LineClump)) eof

88− let keys = map (\(KeyBlock label target) -> (label, target)) $

89− filter isKeyBlock refs

90− -- second pass, with keys stripped out

91− let rawlines = map (\(LineClump ln) -> ln) $ filter isLineClump refs

92− setInput $ concat rawlines

93− updateState (\state -> state { stateKeys = keys })

94− blocks <- parseBlocks

86+ startPos <- getPosition

87+ -- go through once just to get list of reference keys

88+ -- docMinusKeys is the raw document with blanks where the keys were...

89+ docMinusKeys <- many (referenceKey <|> lineClump) >>= return . concat

90+ setInput docMinusKeys

91+ setPosition startPos

92+ st <- getState

93+ let reversedKeys = stateKeys st

94+ updateState $ \st -> st { stateKeys = reverse reversedKeys }

95+ -- now parse it for real...

96+ blocks <- parseBlocks

9597 let blocks' = filter (/= Null) blocks

9698 state <- getState

9799 let (blocks'', title) = if stateStandalone state

@@ -429,9 +431,16 @@ unknownDirective = try $ do

429431-- reference key

430432--

431433

432−referenceKey =

433− choice [imageKey, anonymousKey, regularKeyQuoted, regularKey] >>~

434+referenceKey = do

435+ startPos <- getPosition

436+ key <- choice [imageKey, anonymousKey, regularKeyQuoted, regularKey]

437+ st <- getState

438+ let oldkeys = stateKeys st

439+ updateState $ \st -> st { stateKeys = key : oldkeys }

434440 optional blanklines

441+ endPos <- getPosition

442+ -- return enough blanks to replace key

443+ return $ replicate (sourceLine endPos - sourceLine startPos) '\n'

435444

436445targetURI = do

437446 skipSpaces

@@ -447,26 +456,26 @@ imageKey = try $ do

447456 skipSpaces

448457 string "image::"

449458 src <- targetURI

450− return $ KeyBlock (normalizeSpaces ref) (removeLeadingTrailingSpace src, "")

459+ return (normalizeSpaces ref, (removeLeadingTrailingSpace src, ""))

451460

452461anonymousKey = try $ do

453462 oneOfStrings [".. __:", "__"]

454463 src <- targetURI

455464 state <- getState

456− return $ KeyBlock [Str "_"] (removeLeadingTrailingSpace src, "")

465+ return ([Str "_"], (removeLeadingTrailingSpace src, ""))

457466

458467regularKeyQuoted = try $ do

459468 string ".. _`"

460469 ref <- manyTill inline (char '`')

461470 char ':'

462471 src <- targetURI

463− return $ KeyBlock (normalizeSpaces ref) (removeLeadingTrailingSpace src, "")

472+ return (normalizeSpaces ref, (removeLeadingTrailingSpace src, ""))

464473

465474regularKey = try $ do

466475 string ".. _"

467476 ref <- manyTill inline (char ':')

468477 src <- targetURI

469− return $ KeyBlock (normalizeSpaces ref) (removeLeadingTrailingSpace src, "")

478+ return (normalizeSpaces ref, (removeLeadingTrailingSpace src, ""))

470479

471480 --

472481 -- inline

src/Text/Pandoc/Shared.hs−24

@@ -72,10 +72,6 @@ module Text.Pandoc.Shared (

7272 testStringWith,

7373 ParserState (..),

7474 defaultParserState,

75− Reference (..),

76− isNoteBlock,

77− isKeyBlock,

78− isLineClump,

7975 HeaderType (..),

8076 ParserContext (..),

8177 QuoteContext (..),

@@ -543,26 +539,6 @@ defaultParserState =

543539 stateColumns = 80,

544540 stateHeaderTable = [] }

545541

546−-- | References from preliminary parsing.

547−data Reference

548− = KeyBlock [Inline] Target -- ^ Key for reference-style link (label URL title)

549− | NoteBlock String [Block] -- ^ Footnote reference and contents

550− | LineClump String -- ^ Raw clump of lines with blanks at end

551− deriving (Eq, Read, Show)

552−

553−-- | Auxiliary functions used in preliminary parsing.

554−isNoteBlock :: Reference -> Bool

555−isNoteBlock (NoteBlock _ _) = True

556−isNoteBlock _ = False

557−

558−isKeyBlock :: Reference -> Bool

559−isKeyBlock (KeyBlock _ _) = True

560−isKeyBlock _ = False

561−

562−isLineClump :: Reference -> Bool

563−isLineClump (LineClump _) = True

564−isLineClump _ = False

565−

566542data HeaderType

567543 = SingleHeader Char -- ^ Single line of characters underneath

568544 | DoubleHeader Char -- ^ Lines of characters above and below