d724c6b · typeable/pandoc

git.iko.soy/typeable/pandoc

on master help sign in

html inlines and html blocks handling in textile reader

d724c6bPaul Rivier16 years ago

src/Text/Pandoc/Readers/Textile.hs+26−17

compare with master

unifiedsplit

@@ -28,7 +28,7 @@ Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA

28Conversion from Textile to 'Pandoc' document, based on the spec

29available at ahttp://redcloth.org/hobix.com/textile/

31−Implemented and parsed :

31+Implemented and parsed:

32 - Paragraphs

33 - Code blocks

34 - Lists

@@ -36,8 +36,14 @@ Implemented and parsed :

36 - Inlines : strong, emph, cite, code, deleted, superscript,

37 subscript, links

39−Implemented but discarded :

39+Implemented but discarded:

40 - HTML-specific and CSS-specific attributes

41+

42+Left to be implemented:

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

44+ - footnotes

45+ - should autolink be shared through Parsing.hs ?

46+

41-}

47-}

@@ -47,7 +53,8 @@ module Text.Pandoc.Readers.Textile (

47import Text.Pandoc.Definition

53import Text.Pandoc.Definition

48import Text.Pandoc.Shared

54import Text.Pandoc.Shared

49import Text.Pandoc.Parsing

55import Text.Pandoc.Parsing

50−import Text.Pandoc.Readers.HTML ( htmlTag, htmlEndTag )

56+import Text.Pandoc.Readers.HTML ( htmlTag, htmlEndTag, -- find code blocks

57+ rawHtmlBlock, rawHtmlInline )

51import Text.ParserCombinators.Parsec

58import Text.ParserCombinators.Parsec

52import Data.Char ( digitToInt )

59import Data.Char ( digitToInt )

@@ -83,6 +90,7 @@ blockParsers = [ codeBlock

83 , header

90 , header

84 , blockQuote

91 , blockQuote

85 , anyList

92 , anyList

93+ , rawHtmlBlock'

86 , maybeExplicitBlock "table" table

94 , maybeExplicitBlock "table" table

87 , maybeExplicitBlock "p" para

95 , maybeExplicitBlock "p" para

88 , nullBlock ]

96 , nullBlock ]

@@ -170,17 +178,26 @@ orderedListItemAtDepth depth = try $ do

170 sublist <- option [] (anyListAtDepth (depth + 1) >>= return . (:[]))

178 sublist <- option [] (anyListAtDepth (depth + 1) >>= return . (:[]))

171 return (p:sublist)

179 return (p:sublist)

172

180

173−-- | This terminates a block such as a paragraph.

181+-- | This terminates a block such as a paragraph. Because of raw html

182+-- blocks support, we have to lookAhead for a rawHtmlBlock.

174blockBreak :: GenParser Char ParserState ()

183blockBreak :: GenParser Char ParserState ()

184+blockBreak = try $ choice

175−blockBreak = try $ newline >> blanklines >> return ()

185+ [newline >> blanklines >> return (),

186+ lookAhead rawHtmlBlock' >> return ()]

176

187

188+-- | A raw Html Block, optionally followed by blanklines

189+rawHtmlBlock' :: GenParser Char ParserState Block

190+rawHtmlBlock' = try $ do

191+ b <- rawHtmlBlock

192+ optional blanklines

193+ return b

177

194

178-- | In textile, paragraphs are separated by blank lines.

195-- | In textile, paragraphs are separated by blank lines.

179para :: GenParser Char ParserState Block

196para :: GenParser Char ParserState Block

180para = try $ do

197para = try $ do

181 content <- manyTill inline blockBreak

198 content <- manyTill inline blockBreak

182 return $ Para $ normalizeSpaces content

199 return $ Para $ normalizeSpaces content

183−

200+

184

201

185-- Tables

202-- Tables

186

203

@@ -260,6 +277,7 @@ inlineParsers = [ autoLink

260 , str

277 , str

261 , whitespace

278 , whitespace

262 , endline

279 , endline

280+ , rawHtmlInline

263 , code

281 , code

264 , simpleInline (string "??") (Cite [])

282 , simpleInline (string "??") (Cite [])

265 , simpleInline (string "**") Strong

283 , simpleInline (string "**") Strong

@@ -313,7 +331,7 @@ image = try $ do

313

331

314-- | Any special symbol defined in specialChars

332-- | Any special symbol defined in specialChars

315symbol :: GenParser Char ParserState Inline

333symbol :: GenParser Char ParserState Inline

316−symbol = do

334+symbol = do

317 result <- oneOf specialChars

335 result <- oneOf specialChars

318 return $ Str [result]

336 return $ Str [result]

319

337

@@ -322,7 +340,6 @@ code :: GenParser Char ParserState Inline

322code = surrounded (char '@') anyChar >>=

340code = surrounded (char '@') anyChar >>=

323 return . Code

341 return . Code

324

342

325−

326-- | Html / CSS attributes

343-- | Html / CSS attributes

327attributes :: GenParser Char ParserState String

344attributes :: GenParser Char ParserState String

328attributes = choice [ enclosed (char '(') (char ')') anyChar,

345attributes = choice [ enclosed (char '(') (char ')') anyChar,

@@ -341,12 +358,4 @@ simpleInline :: GenParser Char ParserState t -- ^ surrounding parser

341 -> GenParser Char ParserState Inline -- ^ content parser (to be used repeatedly)

358 -> GenParser Char ParserState Inline -- ^ content parser (to be used repeatedly)

342simpleInline border construct = surrounded border (inlineWithAttribute) >>=

359simpleInline border construct = surrounded border (inlineWithAttribute) >>=

343 return . construct . normalizeSpaces

360 return . construct . normalizeSpaces

344− where inlineWithAttribute = (try $ optional attributes) >> inline

361+ where inlineWithAttribute = (try $ optional attributes) >> inline

345−

346−

347−-- TODO

348−--

349−-- - Pandoc Meta Information (title, author, date)

350−-- - footnotes

351−-- - should autolink be shared through Parsing.hs ?

352−-- - embeded HTML, both inlines and blocks

tests/textile-reader.native+15−1

@@ -113,4 +113,18 @@ Pandoc (Meta {docTitle = [Str ""], docAuthors = [[Str ""]], docDate = [Str ""]})

113 , [ Plain [Str "sex"] ] ],

114 [ [ Plain [Str "joan"] ]

115 , [ Plain [Str "24"] ]

116− , [ Plain [Str "f"] ] ] ] ]

116+ , [ Plain [Str "f"] ] ] ]

117+, Header 1 [Str "Raw",Space,Str "HTML"]

118+, Para [Str "However,",Space,HtmlInline "<strong>",Space,Str "raw",Space,Str "HTML",Space,Str "inlines",Space,HtmlInline "</strong>",Space,Str "are",Space,Str "accepted,",Space,Str "as",Space,Str "well",Space,Str "as",Space,Str ":"]

119+, RawHtml "<div class=\"foobar\">"

120+, Para [Str "any",Space,Strong [Str "Raw",Space,Str "HTML",Space,Str "Block"],Space,Str "with",Space,Str "bold"]

121+, RawHtml "</div>"

122+, Para [Str "Html",Space,Str "blocks",Space,Str "can",Space,Str "be"]

123+, RawHtml "<div>"

124+, Para [Str "inlined"]

125+, RawHtml "</div>"

126+, Para [Str "as",Space,Str "well."]

127+, BulletList

128+ [ [ 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 ">"] ]

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

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

tests/textile-reader.textile+15

@@ -161,3 +161,18 @@ p{color:green}. and paragraph attributes, and table attributes.

161table{foo:bar}.

162| name | age | sex |

163| joan | 24 | f |

164+

165+h1. Raw HTML

166+

167+However, <strong> raw HTML inlines </strong> are accepted, as well as :

168+

169+<div class="foobar">

170+ any *Raw HTML Block* with bold

171+</div>

172+

173+Html blocks can be <div>inlined</div> as well.

174+

175+* this <div> won't produce raw html blocks </div>

176+* but this <strong> will produce inline html </strong>

177+

178+Can you prove that 2 < 3 ?