f917b46 · typeable/pandoc

git.iko.soy/typeable/pandoc

on master help sign in

Textile reader: implemented acronyms, (tm), (r), (c).

f917b46John MacFarlane16 years ago

src/Text/Pandoc/Readers/Textile.hs+29−6

compare with master

unifiedsplit

@@ -43,8 +43,6 @@ Left to be implemented:

43 - Pandoc Meta Information (title, author, date)

44 - footnotes

45 - dimension sign

46− - registered, trademark, and copyright symbols

47− - acronyms

48 - uppercase

46 - uppercase

49 - definition lists

47 - definition lists

50 - continued blocks (ex bq..)

48 - continued blocks (ex bq..)

@@ -86,7 +84,7 @@ readTextile state s = (readWith parseTextile) state (s ++ "\n\n")

87-- | Special chars border strings parsing

85-- | Special chars border strings parsing

88specialChars :: [Char]

86specialChars :: [Char]

89−specialChars = "\\[]<>*#_@~-+^&,.;:!?|\"'%"

87+specialChars = "\\[]<>*#_@~-+^&,.;:!?|\"'%()"

91-- | Generate a Pandoc ADT from a textile document

89-- | Generate a Pandoc ADT from a textile document

92parseTextile :: GenParser Char ParserState Pandoc

90parseTextile :: GenParser Char ParserState Pandoc

@@ -312,6 +310,7 @@ inlines = manyTill inline newline

312-- | Inline parsers tried in order

310-- | Inline parsers tried in order

313inlineParsers :: [GenParser Char ParserState Inline]

311inlineParsers :: [GenParser Char ParserState Inline]

314inlineParsers = [ autoLink

312inlineParsers = [ autoLink

313+ , mark

315 , str

314 , str

316 , htmlSpan

315 , htmlSpan

317-- , smartPunctuation -- from markdown reader

316-- , smartPunctuation -- from markdown reader

@@ -332,16 +331,40 @@ inlineParsers = [ autoLink

332 , symbol

331 , symbol

333 ]

332 ]

334

333

334+-- | Trademark, registered, copyright

335+mark :: GenParser Char st Inline

336+mark = try $ char '(' >> (try tm <|> try reg <|> copy)

337+

338+reg :: GenParser Char st Inline

339+reg = do

340+ oneOf "Rr"

341+ char ')'

342+ return $ Str "\174"

343+

344+tm :: GenParser Char st Inline

345+tm = do

346+ oneOf "Tt"

347+ oneOf "Mm"

348+ char ')'

349+ return $ Str "\8482"

350+

351+copy :: GenParser Char st Inline

352+copy = do

353+ oneOf "Cc"

354+ char ')'

355+ return $ Str "\169"

356+

335-- | Any string

357-- | Any string

336str :: GenParser Char ParserState Inline

358str :: GenParser Char ParserState Inline

337str = do

359str = do

338 xs <- many1 (noneOf (specialChars ++ "\t\n "))

360 xs <- many1 (noneOf (specialChars ++ "\t\n "))

361+ optional $ try $ do

362+ lookAhead (char '(')

363+ notFollowedBy' mark

339− optional $ charsInBalanced '(' ')' -- drop acronym explanation

364+ charsInBalanced '(' ')' -- drop acronym explanation

340− -- e.g. PBS(Public Broadcasting Service)

341 -- parse a following hyphen if followed by a letter

365 -- parse a following hyphen if followed by a letter

342 -- (this prevents unwanted interpretation as starting a strikeout section)

366 -- (this prevents unwanted interpretation as starting a strikeout section)

343 result <- option xs $ try $ do

367 result <- option xs $ try $ do

344− guard $ not . null $ xs

345 char '-'

368 char '-'

346 next <- lookAhead letter

369 next <- lookAhead letter

347 guard $ isLetter (last xs) || isLetter next

370 guard $ isLetter (last xs) || isLetter next

tests/textile-reader.native+7−1

@@ -125,4 +125,10 @@ Pandoc (Meta {docTitle = [Str ""], docAuthors = [[Str ""]], docDate = [Str ""]})

125, BulletList

126 [ [ Plain [Str "this",Space,Str "<",Str "div",Str ">",Space,Str "won",Str "'",Str "t",Space,Str "produce",Space,Str "raw",Space,Str "html",Space,Str "blocks",Space,Str "<",Str "/div",Str ">"] ]

127 , [ Plain [Str "but",Space,Str "this",Space,HtmlInline "<strong>",Space,Str "will",Space,Str "produce",Space,Str "inline",Space,Str "html",Space,HtmlInline "</strong>"] ] ]

128−, Para [Str "Can",Space,Str "you",Space,Str "prove",Space,Str "that",Space,Str "2",Space,Str "<",Space,Str "3",Space,Str "?"] ]

128+, Para [Str "Can",Space,Str "you",Space,Str "prove",Space,Str "that",Space,Str "2",Space,Str "<",Space,Str "3",Space,Str "?"]

129+, Header 1 [Str "Acronyms",Space,Str "and",Space,Str "marks"]

130+, Para [Str "PBS"]

131+, Para [Str "Hi",Str "\8482"]

132+, Para [Str "Hi",Space,Str "\8482"]

133+, Para [Str "\174",Space,Str "Hi",Str "\174"]

134+, Para [Str "Hi",Str "\169",Str "2008",Space,Str "\169",Space,Str "2008"] ]

tests/textile-reader.textile+12

@@ -178,3 +178,15 @@ Html blocks can be <div>inlined</div> as well.

178* but this <strong> will produce inline html </strong>

179

180Can you prove that 2 < 3 ?

181+

182+h1. Acronyms and marks

183+

184+PBS(Public Broadcasting System)

185+

186+Hi(tm)

187+

188+Hi (TM)

189+

190+(r) Hi(r)

191+