d064171 · typeable/pandoc

git.iko.soy/typeable/pandoc

Changes in entity handling:

+ Entities are parsed (and unicode characters returned) in both Markdown and HTML readers. + Parsers characterEntity, namedEntity, decimalEntity, hexEntity added to Entities.hs; these parse a string and return a unicode character. + Changed 'entity' parser in HTML reader to use the 'characterEntity' parser from Entities.hs. + Added new 'entity' parser to Markdown reader, and added '&' as a special character. Adjusted test suite accordingly since now we get 'Str "AT",Str "&",Str "T"' instead of 'Str "AT&T".. + stringToSGML moved to Entities.hs. escapeSGML removed as redundant, given encodeEntities. + stringToSGML, encodeEntities, and specialCharToEntity are given a boolean parameter that causes only numerical entities to be used. This is used in the docbook writer. The HTML writer uses named entities where possible, but not all docbook-consumers know about the named entities without special instructions, so it seems safer to use numerical entities there. + decodeEntities is rewritten in a way that avoids Text.Regex, using the new parsers. + charToEntity and charToNumericalEntity added to Entities.hs. + Moved specialCharToEntity from Shared.hs to Entities.hs. + Removed unneeded 'decodeEntities' from 'str' parser in HTML and Markdown readers. + Removed sgmlHexEntity, sgmlDecimalEntity, sgmlNamedEntity, and sgmlCharacterEntity from Shared.hs. + Modified Docbook writer so that it doesn't rely on Text.Regex for detecting "mailto" links. git-svn-id: https://pandoc.googlecode.com/svn/trunk@515 788f1e2b-df1e-0410-8736-df70ead52e1b

commitd06417125dd4d8cb177abd2d472c0c1cad4c49be authorfiddlosopher <fiddlosopher@788f1e2b-df1e-0410-8736-df70ead52e1b> dateSat 27 Jan 2007, 03:04 (20 years ago) parentf2de088

10 files changed+169−177

compare with master

unifiedsplit

src/Text/Pandoc/Entities.hs+92−45 src/Text/Pandoc/Readers/HTML.hs+5−10 src/Text/Pandoc/Readers/Markdown.hs+9−4 src/Text/Pandoc/Shared.hs+3−55 src/Text/Pandoc/Writers/Docbook.hs+14−16 src/Text/Pandoc/Writers/HTML.hs+10−10 src/Text/Pandoc/Writers/Markdown.hs−1 tests/testsuite.native+3−3 tests/writer.docbook+30−30 tests/writer.native+3−3

src/Text/Pandoc/Entities.hs+92−45

@@ -29,62 +29,109 @@ Functions for encoding unicode characters as entity references,

2929and vice versa.

3030-}

3131module Text.Pandoc.Entities (

32− entityToChar,

3332 charToEntity,

34− decodeEntities,

33+ charToNumericalEntity,

34+ specialCharToEntity,

3535 encodeEntities,

36+ decodeEntities,

37+ stringToSGML,

3638 characterEntity

3739 ) where

3840import Data.Char ( chr, ord )

39−import Text.Regex ( mkRegex, matchRegexAll, Regex )

40−import Maybe ( fromMaybe )

41+import Text.ParserCombinators.Parsec

42+import Data.Maybe ( fromMaybe )

4143

42−-- | Regular expression for numerical coded entity.

43−numericalEntity :: Text.Regex.Regex

44−numericalEntity = mkRegex "&#([0-9]+|[xX][0-9A-Fa-f]+);"

44+-- | Returns a string containing an entity reference for the character.

45+charToEntity :: Char -> String

46+charToEntity char =

47+ let matches = filter (\(entity, character) -> (character == char))

48+ entityTable in

49+ if (length matches) == 0

50+ then charToNumericalEntity char

51+ else fst (head matches)

4552

46−-- | Regular expression for character entity.

47−characterEntity :: Text.Regex.Regex

48−characterEntity = mkRegex "&#[0-9]+;|&#[xX][0-9A-Fa-f]+;|&[A-Za-z0-9]+;"

53+-- | Returns a string containing a numerical entity reference for the char.

54+charToNumericalEntity :: Char -> String

55+charToNumericalEntity ch = "&#" ++ show (ord ch) ++ ";"

4956

50−-- | Return a string with all entity references decoded to unicode characters

51−-- where possible.

52−decodeEntities :: String -> String

53−decodeEntities str =

54− case (matchRegexAll characterEntity str) of

55− Nothing -> str

56− Just (before, match, rest, _) -> before ++ replacement ++

57− (decodeEntities rest)

58− where replacement = case (entityToChar match) of

59− Just ch -> [ch]

60− Nothing -> match

57+-- | Escape special character to SGML entity.

58+specialCharToEntity :: Bool -- ^ Use numerical entities only.

59+ -> Char -- ^ Character to convert.

60+ -> [Char]

61+specialCharToEntity numericalEntities c =

62+ if (c `elem` "&<>\"") || (ord c > 127)

63+ then if numericalEntities

64+ then charToNumericalEntity c

65+ else charToEntity c

66+ else [c]

6167

62−-- | Returns a string with characters replaced with entity references where

63−-- possible.

64−encodeEntities :: String -> String

65−encodeEntities [] = []

66−encodeEntities (c:cs) = if ord c < 128

67− then c:(encodeEntities cs)

68− else (charToEntity c) ++ (encodeEntities cs)

68+-- | Parse SGML character entity.

69+characterEntity :: GenParser Char st Char

70+characterEntity = namedEntity <|> hexEntity <|> decimalEntity <?> "SGML entity"

6971

70−-- | If the string is a valid entity reference, returns @Just@ the character,

71−-- otherwise @Nothing@.

72−entityToChar :: String -> Maybe Char

73−entityToChar entity =

74− case (lookup entity entityTable) of

75− Just ch -> Just ch

76− Nothing -> case (matchRegexAll numericalEntity entity) of

77− Just (_, _, _, [sub]) -> Just (chr (read ('0':sub)))

78− Nothing -> Nothing

72+-- | Parse SGML character entity.

73+namedEntity :: GenParser Char st Char

74+namedEntity = try $ do

75+ st <- char '&'

76+ body <- many1 alphaNum

77+ end <- char ';'

78+ let entity = "&" ++ body ++ ";"

79+ return $ case (lookup entity entityTable) of

80+ Just ch -> ch

81+ Nothing -> '?'

82+

83+-- | Parse SGML hexadecimal entity.

84+hexEntity :: GenParser Char st Char

85+hexEntity = try $ do

86+ st <- string "&#"

87+ hex <- oneOf "Xx"

88+ body <- many1 (oneOf "0123456789ABCDEFabcdef")

89+ end <- char ';'

90+ return $ chr $ read ('0':'x':body)

7991

80−-- | Returns a string containing an entity reference for the character.

81−charToEntity :: Char -> String

82−charToEntity char =

83− let matches = filter (\(entity, character) -> (character == char))

84− entityTable in

85− if (length matches) == 0

86− then "&#" ++ show (ord char) ++ ";"

87− else fst (head matches)

92+-- | Parse SGML decimal entity.

93+decimalEntity :: GenParser Char st Char

94+decimalEntity = try $ do

95+ st <- string "&#"

96+ body <- many1 digit

97+ end <- char ';'

98+ return $ chr $ read body

99+

100+-- | Escape string as needed for SGML. Entity references are not preserved.

101+encodeEntities :: Bool -- ^ Use only numerical entities.

102+ -> String -- ^ String to convert.

103+ -> String

104+encodeEntities numericalEntities =

105+ concatMap (specialCharToEntity numericalEntities)

106+

107+-- | Escape string as needed for SGML, using only numerical entities.

108+-- Entity references are not preserved.

109+encodeEntitiesNumerical :: String -> String

110+encodeEntitiesNumerical =

111+ concatMap (\c -> "&#" ++ show (ord c) ++ ";")

112+

113+-- | Convert entities in a string to characters.

114+decodeEntities :: String -> String

115+decodeEntities str =

116+ case parse (many (characterEntity <|> anyChar)) str str of

117+ Left err -> error $ "\nError: " ++ show err

118+ Right result -> result

119+

120+-- | Escape string for SGML, preserving entity references.

121+stringToSGML :: Bool -- ^ Use only numerical entities.

122+ -> String -- ^ String to convert.

123+ -> String

124+stringToSGML numericalEntities str =

125+ let nonentity = do

126+ notFollowedBy characterEntity

127+ c <- anyChar

128+ return $ specialCharToEntity numericalEntities c

129+ entity = do

130+ ent <- manyTill anyChar (char ';')

131+ return (ent ++ ";") in

132+ case parse (many (nonentity <|> entity)) str str of

133+ Left err -> error $ "\nError: " ++ show err

134+ Right result -> concat result

88135

89136entityTable :: [(String, Char)]

90137entityTable = [

src/Text/Pandoc/Readers/HTML.hs+5−10

@@ -44,7 +44,7 @@ import Text.ParserCombinators.Parsec

4444import Text.ParserCombinators.Pandoc

4545import Text.Pandoc.Definition

4646import Text.Pandoc.Shared

47−import Text.Pandoc.Entities ( decodeEntities, entityToChar )

47+import Text.Pandoc.Entities ( characterEntity, decodeEntities )

4848import Maybe ( fromMaybe )

4949import Data.List ( intersect, takeWhile, dropWhile )

5050import Data.Char ( toUpper, toLower, isAlphaNum )

@@ -391,14 +391,9 @@ text = choice [ entity, strong, emph, code, str, linebreak, whitespace ] <?> "t

391391special = choice [ link, image, rawHtmlInline ] <?>

392392 "link, inline html, or image"

393393

394−entity = try (do

395− char '&'

396− body <- choice [(many1 letter), (try (do

397− char '#'

398− num <- many1 digit

399− return ("#" ++ num)))]

400− char ';'

401− return (Str [fromMaybe '?' (entityToChar ("&" ++ body ++ ";"))]))

394+entity = do

395+ ent <- characterEntity

396+ return $ Str [ent]

402397

403398code = try (do

404399 htmlTag "code"

@@ -439,7 +434,7 @@ linebreak = do

439434

440435str = do

441436 result <- many1 (noneOf "<& \t\n")

442− return (Str (decodeEntities result))

437+ return (Str result)

443438

444439--

445440-- links and images

src/Text/Pandoc/Readers/Markdown.hs+9−4

@@ -42,7 +42,7 @@ import Text.Pandoc.Readers.HTML ( rawHtmlBlock,

4242 anyHtmlTag, anyHtmlEndTag,

4343 htmlEndTag, extractTagType,

4444 htmlBlockElement )

45−import Text.Pandoc.Entities ( decodeEntities )

45+import Text.Pandoc.Entities ( characterEntity )

4646import Text.ParserCombinators.Parsec

4747

4848-- | Read markdown from an input string and return a Pandoc document.

@@ -88,12 +88,13 @@ blockQuoteChar = '>'

8888hyphenChar = '-'

8989ellipsesChar = '.'

9090listColSepChar = '|'

91+entityStart = '&'

9192

9293-- treat these as potentially non-text when parsing inline:

9394specialChars = [escapeChar, labelStart, labelEnd, emphStart, emphEnd,

9495 emphStartAlt, emphEndAlt, codeStart, codeEnd, autoLinkEnd,

9596 autoLinkStart, mathStart, mathEnd, imageStart, noteStart,

96− hyphenChar, ellipsesChar] ++ quoteChars

97+ hyphenChar, ellipsesChar, entityStart] ++ quoteChars

9798

9899--

99100-- auxiliary functions

@@ -674,7 +675,7 @@ text = choice [ escapedChar, math, strong, emph, smartPunctuation,

674675 code, ltSign, symbol,

675676 str, linebreak, tabchar, whitespace, endline ] <?> "text"

676677

677−inline = choice [ rawLaTeXInline', escapedChar, special, text ] <?> "inline"

678+inline = choice [ rawLaTeXInline', escapedChar, entity, special, text ] <?> "inline"

678679

679680special = choice [ noteRef, inlineNote, link, referenceLink, rawHtmlInline',

680681 autoLink, image ] <?> "link, inline html, note, or image"

@@ -827,9 +828,13 @@ linebreak = try (do

827828

828829nonEndline = noneOf endLineChars

829830

831+entity = do

832+ ent <- characterEntity

833+ return $ Str [ent]

834+

830835str = do

831836 result <- many1 ((noneOf (specialChars ++ spaceChars ++ endLineChars)))

832− return (Str (decodeEntities result))

837+ return (Str result)

833838

834839-- an endline character that can be treated as a space, not a structural break

835840endline = try (do

src/Text/Pandoc/Shared.hs+3−55

@@ -65,8 +65,6 @@ module Text.Pandoc.Shared (

6565 replaceReferenceLinks,

6666 replaceRefLinksBlockList,

6767 -- * SGML

68− escapeSGML,

69− stringToSGML,

7068 inTags,

7169 selfClosingTag,

7270 inTagsSimple,

@@ -74,7 +72,7 @@ module Text.Pandoc.Shared (

7472 ) where

7573import Text.Pandoc.Definition

7674import Text.ParserCombinators.Parsec as Parsec

77−import Text.Pandoc.Entities ( decodeEntities, charToEntity )

75+import Text.Pandoc.Entities ( decodeEntities, encodeEntities, stringToSGML )

7876import Text.PrettyPrint.HughesPJ as PP ( text, char, (<>),

7977 ($$), nest, Doc, isEmpty )

8078import Data.Char ( toLower, ord )

@@ -523,61 +521,11 @@ replaceRefLinksInline keytable (Quoted t lst) =

523521 Quoted t (map (replaceRefLinksInline keytable) lst)

524522replaceRefLinksInline keytable other = other

525523

526−-- | Parse SGML character entity.

527−sgmlCharacterEntity :: GenParser Char st [Char]

528−sgmlCharacterEntity = sgmlNamedEntity <|> sgmlDecimalEntity <|>

529− sgmlHexEntity <?> "SGML entity"

530−

531−-- | Parse SGML character entity.

532−sgmlNamedEntity :: GenParser Char st [Char]

533−sgmlNamedEntity = try $ do

534− st <- Parsec.char '&'

535− body <- many1 alphaNum

536− end <- Parsec.char ';'

537− return $ (st:body) ++ [end]

538−

539−-- | Parse SGML decimal entity.

540−sgmlDecimalEntity :: GenParser Char st [Char]

541−sgmlDecimalEntity = try $ do

542− st <- string "&#"

543− body <- many1 (oneOf "0123456789")

544− end <- Parsec.char ';'

545− return $ st ++ body ++ [end]

546−

547−-- | Parse SGML hexadecimal entity.

548−sgmlHexEntity :: GenParser Char st [Char]

549−sgmlHexEntity = try $ do

550− st <- string "&#"

551− hex <- oneOf "Xx"

552− body <- many1 (oneOf "0123456789ABCDEFabcdef")

553− end <- Parsec.char ';'

554− return $ st ++ (hex:body) ++ [end]

555−

556−-- | Escape special character to SGML entity.

557−specialCharToEntity :: Char -> [Char]

558−specialCharToEntity c = if (c `elem` "&<>\"") || (ord c > 127)

559− then charToEntity c

560− else [c]

561−

562−-- | Escape string, preserving character entities.

563−stringToSGML :: String -> String

564−stringToSGML str =

565− let segment = sgmlCharacterEntity <|>

566− (do{c <- anyChar;

567− return $ specialCharToEntity c})

568− sgmlString = (do{segs <- many segment; return $ concat segs}) in

569− case parse sgmlString str str of

570− Left err -> error $ "\nError:\n" ++ show err

571− Right result -> result

572−

573−-- | Escape string as needed for SGML. Entity references are not preserved.

574−escapeSGML :: String -> String

575−escapeSGML = concatMap specialCharToEntity

576−

577524-- | Return a text object with a string of formatted SGML attributes.

578525attributeList :: [(String, String)] -> Doc

579526attributeList = text . concatMap

580− (\(a, b) -> " " ++ stringToSGML a ++ "=\"" ++ stringToSGML b ++ "\"")

527+ (\(a, b) -> " " ++ stringToSGML True a ++ "=\"" ++

528+ stringToSGML True b ++ "\"")

581529

582530-- | Put the supplied contents between start and end tags of tagType,

583531-- with specified attributes and (if specified) indentation.

src/Text/Pandoc/Writers/Docbook.hs+14−16

@@ -32,10 +32,9 @@ module Text.Pandoc.Writers.Docbook (

3232 ) where

3333import Text.Pandoc.Definition

3434import Text.Pandoc.Shared

35−import Text.Pandoc.Entities ( encodeEntities )

36−import Text.Regex ( mkRegex, matchRegex )

35+import Text.Pandoc.Entities ( encodeEntities, stringToSGML )

3736import Data.Char ( toLower, ord )

38−import Data.List ( isPrefixOf, partition )

37+import Data.List ( isPrefixOf, partition, drop )

3938import Text.PrettyPrint.HughesPJ hiding ( Str )

4039

4140-- | Data structure for defining hierarchical Pandoc documents

@@ -65,8 +64,8 @@ authorToDocbook name = inTagsIndented "author" $

6564 then -- last name first

6665 let (lastname, rest) = break (==',') name

6766 firstname = removeLeadingSpace rest in

68− inTagsSimple "firstname" (text $ stringToSGML firstname) <>

69− inTagsSimple "surname" (text $ stringToSGML lastname)

67+ inTagsSimple "firstname" (text $ stringToSGML True firstname) <>

68+ inTagsSimple "surname" (text $ stringToSGML True lastname)

7069 else -- last name last

7170 let namewords = words name

7271 lengthname = length namewords

@@ -74,8 +73,8 @@ authorToDocbook name = inTagsIndented "author" $

7473 0 -> ("","")

7574 1 -> ("", name)

7675 n -> (joinWithSep " " (take (n-1) namewords), last namewords) in

77− inTagsSimple "firstname" (text $ stringToSGML firstname) $$

78− inTagsSimple "surname" (text $ stringToSGML lastname)

76+ inTagsSimple "firstname" (text $ stringToSGML True firstname) $$

77+ inTagsSimple "surname" (text $ stringToSGML True lastname)

7978

8079-- | Convert Pandoc document to string in Docbook format.

8180writeDocbook :: WriterOptions -> Pandoc -> String

@@ -87,7 +86,7 @@ writeDocbook opts (Pandoc (Meta title authors date) blocks) =

8786 then inTagsIndented "articleinfo" $

8887 (inTagsSimple "title" (wrap opts title)) $$

8988 (vcat (map authorToDocbook authors)) $$

90− (inTagsSimple "date" (text $ stringToSGML date))

89+ (inTagsSimple "date" (text $ stringToSGML True date))

9190 else empty

9291 blocks' = replaceReferenceLinks blocks

9392 (noteBlocks, blocks'') = partition isNoteBlock blocks'

@@ -142,7 +141,7 @@ blockToDocbook opts (Para lst) =

142141blockToDocbook opts (BlockQuote blocks) =

143142 inTagsIndented "blockquote" (blocksToDocbook opts blocks)

144143blockToDocbook opts (CodeBlock str) =

145− text "<screen>\n" <> text (escapeSGML str) <> text "\n</screen>"

144+ text "<screen>\n" <> text (encodeEntities True str) <> text "\n</screen>"

146145blockToDocbook opts (BulletList lst) =

147146 inTagsIndented "itemizedlist" $ listItemsToDocbook opts lst

148147blockToDocbook opts (OrderedList lst) =

@@ -199,7 +198,7 @@ inlinesToDocbook opts lst = hcat (map (inlineToDocbook opts) lst)

199198

200199-- | Convert an inline element to Docbook.

201200inlineToDocbook :: WriterOptions -> Inline -> Doc

202−inlineToDocbook opts (Str str) = text $ stringToSGML str

201+inlineToDocbook opts (Str str) = text $ stringToSGML True str

203202inlineToDocbook opts (Emph lst) =

204203 inTagsSimple "emphasis" (inlinesToDocbook opts lst)

205204inlineToDocbook opts (Strong lst) =

@@ -212,24 +211,23 @@ inlineToDocbook opts Ellipses = text "…"

212211inlineToDocbook opts EmDash = text "—"

213212inlineToDocbook opts EnDash = text "–"

214213inlineToDocbook opts (Code str) =

215− inTagsSimple "literal" $ text (escapeSGML str)

214+ inTagsSimple "literal" $ text (encodeEntities True str)

216215inlineToDocbook opts (TeX str) = inlineToDocbook opts (Code str)

217216inlineToDocbook opts (HtmlInline str) = empty

218217inlineToDocbook opts LineBreak =

219218 text $ "<literallayout></literallayout>"

220219inlineToDocbook opts Space = char ' '

221220inlineToDocbook opts (Link txt (Src src tit)) =

222− case (matchRegex (mkRegex "mailto:(.*)") src) of

223− Just [addr] -> inTagsSimple "email" $ text (escapeSGML addr)

224− Nothing -> inTags False "ulink" [("url", src)] $

225− inlinesToDocbook opts txt

221+ if isPrefixOf "mailto:" src

222+ then inTagsSimple "email" $ text (encodeEntities True $ drop 7 src)

223+ else inTags False "ulink" [("url", src)] $ inlinesToDocbook opts txt

226224inlineToDocbook opts (Link text (Ref ref)) = empty -- shouldn't occur

227225inlineToDocbook opts (Image alt (Src src tit)) =

228226 let titleDoc = if null tit

229227 then empty

230228 else inTagsIndented "objectinfo" $

231229 inTagsIndented "title"

232− (text $ stringToSGML tit) in

230+ (text $ stringToSGML True tit) in

233231 inTagsIndented "inlinemediaobject" $

234232 inTagsIndented "imageobject" $

235233 titleDoc $$ selfClosingTag "imagedata" [("fileref", src)]

src/Text/Pandoc/Writers/HTML.hs+10−10

@@ -32,7 +32,7 @@ module Text.Pandoc.Writers.HTML (

3232 ) where

3333import Text.Pandoc.Definition

3434import Text.Pandoc.Shared

35−import Text.Pandoc.Entities ( encodeEntities )

35+import Text.Pandoc.Entities ( encodeEntities, stringToSGML )

3636import Text.Regex ( mkRegex, matchRegex )

3737import Numeric ( showHex )

3838import Data.Char ( ord, toLower )

@@ -127,11 +127,11 @@ htmlHeader opts (Meta title authors date) =

127127 then empty

128128 else selfClosingTag "meta" [("name", "author"),

129129 ("content",

130− joinWithSep ", " (map stringToSGML authors))]

130+ joinWithSep ", " (map (stringToSGML False) authors))]

131131 datetext = if (date == "")

132132 then empty

133133 else selfClosingTag "meta" [("name", "date"),

134− ("content", stringToSGML date)] in

134+ ("content", stringToSGML False date)] in

135135 text (writerHeader opts) $$ authortext $$ datetext $$ titletext $$

136136 text "</head>\n<body>"

137137

@@ -168,7 +168,7 @@ blockToHtml opts (Note ref lst) =

168168 (text "↩")

169169blockToHtml opts (Key _ _) = empty

170170blockToHtml opts (CodeBlock str) =

171− text "<pre><code>" <> text (escapeSGML str) <> text "\n</code></pre>"

171+ text "<pre><code>" <> text (encodeEntities False str) <> text "\n</code></pre>"

172172blockToHtml opts (RawHtml str) = text str

173173blockToHtml opts (BulletList lst) =

174174 let attribs = if (writerIncremental opts)

@@ -234,7 +234,7 @@ inlineToHtml opts (Emph lst) =

234234inlineToHtml opts (Strong lst) =

235235 inTagsSimple "strong" (inlineListToHtml opts lst)

236236inlineToHtml opts (Code str) =

237− inTagsSimple "code" $ text (escapeSGML str)

237+ inTagsSimple "code" $ text (encodeEntities False str)

238238inlineToHtml opts (Quoted SingleQuote lst) =

239239 text "‘" <> (inlineListToHtml opts lst) <> text "’"

240240inlineToHtml opts (Quoted DoubleQuote lst) =

@@ -243,16 +243,16 @@ inlineToHtml opts EmDash = text "—"

243243inlineToHtml opts EnDash = text "–"

244244inlineToHtml opts Ellipses = text "…"

245245inlineToHtml opts Apostrophe = text "’"

246−inlineToHtml opts (Str str) = text $ stringToSGML str

247−inlineToHtml opts (TeX str) = text $ escapeSGML str

246+inlineToHtml opts (Str str) = text $ stringToSGML False str

247+inlineToHtml opts (TeX str) = text $ encodeEntities False str

248248inlineToHtml opts (HtmlInline str) = text str

249249inlineToHtml opts (LineBreak) = selfClosingTag "br" []

250250inlineToHtml opts Space = space

251251inlineToHtml opts (Link txt (Src src tit)) =

252− let title = stringToSGML tit in

252+ let title = stringToSGML False tit in

253253 if (isPrefixOf "mailto:" src)

254254 then obfuscateLink opts txt src

255− else inTags False "a" ([("href", escapeSGML src)] ++

255+ else inTags False "a" ([("href", encodeEntities False src)] ++

256256 if null tit then [] else [("title", title)])

257257 (inlineListToHtml opts txt)

258258inlineToHtml opts (Link txt (Ref ref)) =

@@ -260,7 +260,7 @@ inlineToHtml opts (Link txt (Ref ref)) =

260260 (inlineListToHtml opts ref) <> char ']'

261261 -- this is what markdown does, for better or worse

262262inlineToHtml opts (Image alt (Src source tit)) =

263− let title = stringToSGML tit

263+ let title = stringToSGML False tit

264264 alternate = render $ inlineListToHtml opts alt in

265265 selfClosingTag "img" $ [("src", source)] ++

266266 (if null alternate then [] else [("alt", alternate)]) ++

src/Text/Pandoc/Writers/Markdown.hs−1

@@ -32,7 +32,6 @@ Markdown: <http://daringfireball.net/projects/markdown/>

3232module Text.Pandoc.Writers.Markdown (

3333 writeMarkdown

3434 ) where

35−import Text.Regex ( matchRegex, mkRegex )

3635import Text.Pandoc.Definition

3736import Text.Pandoc.Shared

3837import Data.List ( group, isPrefixOf, drop )

tests/testsuite.native+3−3

@@ -240,8 +240,8 @@ Pandoc (Meta [Str "Pandoc",Space,Str "Test",Space,Str "Suite"] ["John MacFarlane

240240 , [ Plain [Str "section:",Space,Str "\167"] ]

241241 , [ Plain [Str "set",Space,Str "membership:",Space,Str "\8712"] ]

242242 , [ Plain [Str "copyright:",Space,Str "\169"] ] ]

243−, Para [Str "AT&T",Space,Str "has",Space,Str "an",Space,Str "ampersand",Space,Str "in",Space,Str "their",Space,Str "name",Str "."]

244−, Para [Str "AT&T",Space,Str "is",Space,Str "another",Space,Str "way",Space,Str "to",Space,Str "write",Space,Str "it",Str "."]

243+, Para [Str "AT",Str "&",Str "T",Space,Str "has",Space,Str "an",Space,Str "ampersand",Space,Str "in",Space,Str "their",Space,Str "name",Str "."]

244+, Para [Str "AT",Str "&",Str "T",Space,Str "is",Space,Str "another",Space,Str "way",Space,Str "to",Space,Str "write",Space,Str "it",Str "."]

245245, Para [Str "This",Space,Str "&",Space,Str "that",Str "."]

246246, Para [Str "4",Space,Str "<",Space,Str "5",Str "."]

247247, Para [Str "6",Space,Str ">",Space,Str "5",Str "."]

@@ -294,7 +294,7 @@ Pandoc (Meta [Str "Pandoc",Space,Str "Test",Space,Str "Suite"] ["John MacFarlane

294294, Key [Str "bar"] (Src "/url/" "Title with "quotes" inside")

295295, Header 2 [Str "With",Space,Str "ampersands"]

296296, Para [Str "Here",Apostrophe,Str "s",Space,Str "a",Space,Link [Str "link",Space,Str "with",Space,Str "an",Space,Str "ampersand",Space,Str "in",Space,Str "the",Space,Str "URL"] (Ref [Str "1"]),Str "."]

297−, Para [Str "Here",Apostrophe,Str "s",Space,Str "a",Space,Str "link",Space,Str "with",Space,Str "an",Space,Str "amersand",Space,Str "in",Space,Str "the",Space,Str "link",Space,Str "text:",Space,Link [Str "AT&T"] (Ref [Str "2"]),Str "."]

297+, Para [Str "Here",Apostrophe,Str "s",Space,Str "a",Space,Str "link",Space,Str "with",Space,Str "an",Space,Str "amersand",Space,Str "in",Space,Str "the",Space,Str "link",Space,Str "text:",Space,Link [Str "AT",Str "&",Str "T"] (Ref [Str "2"]),Str "."]

298298, Para [Str "Here",Apostrophe,Str "s",Space,Str "an",Space,Link [Str "inline",Space,Str "link"] (Src "/script?foo=1&bar=2" ""),Str "."]

299299, Para [Str "Here",Apostrophe,Str "s",Space,Str "an",Space,Link [Str "inline",Space,Str "link",Space,Str "in",Space,Str "pointy",Space,Str "braces"] (Src "/script?foo=1&bar=2" ""),Str "."]

300300, Key [Str "1"] (Src "http://example.com/?foo=1&bar=2" "")

tests/writer.docbook+30−30

@@ -89,7 +89,7 @@

8989 </para>

9090 <screen>

9191sub status {

92− print "working";

92+ print "working";

9393}

9494</screen>

9595 <para>

@@ -122,7 +122,7 @@ sub status {

122122 </blockquote>

123123 </blockquote>

124124 <para>

125− This should not be a block quote: 2 > 1.

125+ This should not be a block quote: 2 > 1.

126126 </para>

127127 <para>

128128 Box-style:

@@ -133,7 +133,7 @@ sub status {

133133 </para>

134134 <screen>

135135sub status {

136− print "working";

136+ print "working";

137137}

138138</screen>

139139 </blockquote>

@@ -177,7 +177,7 @@ sub status {

177177---- (should be four hyphens)

178178

179179sub status {

180− print "working";

180+ print "working";

181181}

182182

183183this code block is indented by one tab

@@ -188,7 +188,7 @@ this code block is indented by one tab

188188 <screen>

189189 this code block is indented by two tabs

190190

191−These should not be escaped: \$ \\ \> \[ \{

191+These should not be escaped: \$ \\ \> \[ \{

192192</screen>

193193 </section>

194194 <section>

@@ -577,9 +577,9 @@ These should not be escaped: \$ \\ \> \[ \{

577577 word.

578578 </para>

579579 <para>

580− This is code: <literal>></literal>, <literal>$</literal>,

580+ This is code: <literal>></literal>, <literal>$</literal>,

581581 <literal>\</literal>, <literal>\$</literal>,

582− <literal><html></literal>.

582+ <literal><html></literal>.

583583 </para>

584584 </section>

585585 <section>

@@ -602,7 +602,7 @@ These should not be escaped: \$ \\ \> \[ \{

602602 </para>

603603 <para>

604604 Here is some quoted <quote><literal>code</literal></quote> and a

605− <quote><ulink url="http://example.com/?foo=1&bar=2">quoted link</ulink></quote>.

605+ <quote><ulink url="http://example.com/?foo=1&bar=2">quoted link</ulink></quote>.

606606 </para>

607607 <para>

608608 Some dashes: one—two—three—four—five.

@@ -691,9 +691,9 @@ These should not be escaped: \$ \\ \> \[ \{

691691 </para>

692692 <para>

693693 <literal>\begin{tabular}{|l|l|}\hline

694−Animal & Number \\ \hline

695−Dog & 2 \\

696−Cat & 1 \\ \hline

694+Animal & Number \\ \hline

695+Dog & 2 \\

696+Cat & 1 \\ \hline

697697\end{tabular}</literal>

698698 </para>

699699 </section>

@@ -705,44 +705,44 @@ Cat & 1 \\ \hline

705705 <itemizedlist>

706706 <listitem>

707707 <para>

708− I hat: Î

708+ I hat: Î

709709 </para>

710710 </listitem>

711711 <listitem>

712712 <para>

713− o umlaut: ö

713+ o umlaut: ö

714714 </para>

715715 </listitem>

716716 <listitem>

717717 <para>

718− section: §

718+ section: §

719719 </para>

720720 </listitem>

721721 <listitem>

722722 <para>

723− set membership: ∈

723+ set membership: ∈

724724 </para>

725725 </listitem>

726726 <listitem>

727727 <para>

729729 </para>

730730 </listitem>

731731 </itemizedlist>

732732 <para>

733− AT&T has an ampersand in their name.

733+ AT&T has an ampersand in their name.

734734 </para>

735735 <para>

736− AT&T is another way to write it.

736+ AT&T is another way to write it.

737737 </para>

738738 <para>

739− This & that.

739+ This & that.

740740 </para>

741741 <para>

742− 4 < 5.

742+ 4 < 5.

743743 </para>

744744 <para>

745− 6 > 5.

745+ 6 > 5.

746746 </para>

747747 <para>

748748 Backslash: \

@@ -775,7 +775,7 @@ Cat & 1 \\ \hline

775775 Right paren: )

776776 </para>

777777 <para>

778− Greater-than: >

778+ Greater-than: >

779779 </para>

780780 <para>

781781 Hash: #

@@ -868,25 +868,25 @@ Cat & 1 \\ \hline

868868 <title>With ampersands</title>

869869 <para>

870870 Here's a

871− <ulink url="http://example.com/?foo=1&bar=2">link with an ampersand in the URL</ulink>.

871+ <ulink url="http://example.com/?foo=1&bar=2">link with an ampersand in the URL</ulink>.

872872 </para>

873873 <para>

874874 Here's a link with an amersand in the link text:

875− <ulink url="http://att.com/">AT&T</ulink>.

875+ <ulink url="http://att.com/">AT&T</ulink>.

876876 </para>

877877 <para>

878− Here's an <ulink url="/script?foo=1&bar=2">inline link</ulink>.

878+ Here's an <ulink url="/script?foo=1&bar=2">inline link</ulink>.

879879 </para>

880880 <para>

881881 Here's an

882− <ulink url="/script?foo=1&bar=2">inline link in pointy braces</ulink>.

882+ <ulink url="/script?foo=1&bar=2">inline link in pointy braces</ulink>.

883883 </para>

884884 </section>

885885 <section>

886886 <title>Autolinks</title>

887887 <para>

888888 With an ampersand:

889− <ulink url="http://example.com/?foo=1&bar=2">http://example.com/?foo=1&bar=2</ulink>

889+ <ulink url="http://example.com/?foo=1&bar=2">http://example.com/?foo=1&bar=2</ulink>

890890 </para>

891891 <itemizedlist>

892892 <listitem>

@@ -916,10 +916,10 @@ Cat & 1 \\ \hline

916916 </blockquote>

917917 <para>

918918 Auto-links should not occur here:

919− <literal><http://example.com/></literal>

919+ <literal><http://example.com/></literal>

920920 </para>

921921 <screen>

922−or here: <http://example.com/>

922+or here: <http://example.com/>

923923</screen>

924924 </section>

925925 </section>

@@ -970,7 +970,7 @@ or here: <http://example.com/>

970970 footnote (as with list items).

971971 </para>

972972 <screen>

973− { <code> }

973+ { <code> }

974974</screen>

975975 <para>

976976 If you want, you can indent every line, but you can also be lazy

tests/writer.native+3−3