4e57451 · typeable/pandoc

git.iko.soy/typeable/pandoc

on master help sign in

Use entities for all characters above 127 in docbook output.

4e57451fiddlosopher20 years ago

src/Text/Pandoc/HtmlEntities.hs+10−7

compare with master

unifiedsplit

@@ -57,7 +57,10 @@ decodeEntities str =

57-- | Returns a string with characters replaced with entity references where

58-- possible.

59encodeEntities :: String -> String

60−encodeEntities = concatMap (\c -> fromMaybe [c] (charToHtmlEntity c))

60+encodeEntities [] = []

61+encodeEntities (c:cs) = if ord c < 127

62+ then c:(encodeEntities cs)

63+ else (charToHtmlEntity c) ++ (encodeEntities cs)

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

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

63-- otherwise @Nothing@.

66-- otherwise @Nothing@.

@@ -69,14 +72,14 @@ htmlEntityToChar entity =

69 Just (_, _, _, [sub]) -> Just (chr (read sub))

72 Just (_, _, _, [sub]) -> Just (chr (read sub))

70 Nothing -> Nothing

73 Nothing -> Nothing

72−-- | If there is an entity reference corresponding to the character, returns

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

73−-- @Just@ the entity reference, otherwise @Nothing@.

74−charToHtmlEntity :: Char -> Maybe String

76+charToHtmlEntity :: Char -> String

75charToHtmlEntity char =

77charToHtmlEntity char =

76− let matches = filter (\(entity, character) -> (character == char)) htmlEntityTable in

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

79+ htmlEntityTable in

77 if (length matches) == 0

80 if (length matches) == 0

78− then Nothing

81+ then "&#" ++ show (ord char) ++ ";"

79− else Just (fst (head matches))

82+ else fst (head matches)

81htmlEntityTable :: [(String, Char)]

84htmlEntityTable :: [(String, Char)]

82htmlEntityTable = [

85htmlEntityTable = [

src/Text/Pandoc/Writers/Docbook.hs+7−5

@@ -33,9 +33,10 @@ module Text.Pandoc.Writers.Docbook (

33import Text.Pandoc.Definition

34import Text.Pandoc.Shared

35import Text.Pandoc.Writers.HTML ( stringToSmartHtml, stringToHtml )

36+import Text.Pandoc.HtmlEntities ( encodeEntities )

36import Text.Html ( stringToHtmlString )

37import Text.Html ( stringToHtmlString )

37import Text.Regex ( mkRegex, matchRegex )

38import Text.Regex ( mkRegex, matchRegex )

38−import Data.Char ( toLower )

39+import Data.Char ( toLower, ord )

39import Data.List ( isPrefixOf, partition )

40import Data.List ( isPrefixOf, partition )

40import Text.PrettyPrint.HughesPJ hiding ( Str )

41import Text.PrettyPrint.HughesPJ hiding ( Str )

@@ -179,9 +180,10 @@ wrap options lst = fsep $ map (hcat . (map (inlineToDocbook options))) (splitByS

179

180

180-- | Escape a string for XML (with "smart" option if specified).

181-- | Escape a string for XML (with "smart" option if specified).

181stringToXML :: WriterOptions -> String -> String

182stringToXML :: WriterOptions -> String -> String

183+stringToXML options = encodeEntities .

182−stringToXML options = if writerSmart options

184+ (if writerSmart options

183− then stringToSmartHtml

185+ then stringToSmartHtml

184− else stringToHtml

186+ else stringToHtml)

185

187

186-- | Escape string to XML appropriate for attributes

188-- | Escape string to XML appropriate for attributes

187attributeStringToXML :: String -> String

189attributeStringToXML :: String -> String

@@ -189,7 +191,7 @@ attributeStringToXML = gsub "\"" """ . codeStringToXML

189

191

190-- | Escape a literal string for XML.

192-- | Escape a literal string for XML.

191codeStringToXML :: String -> String

193codeStringToXML :: String -> String

192−codeStringToXML = gsub "<" "<" . gsub "&" "&"

194+codeStringToXML = encodeEntities . gsub "<" "<" . gsub "&" "&"

193

195

194-- | Convert a list of inline elements to Docbook.

196-- | Convert a list of inline elements to Docbook.

195inlinesToDocbook :: WriterOptions -> [Inline] -> Doc

197inlinesToDocbook :: WriterOptions -> [Inline] -> Doc

tests/writer.docbook+5−5

@@ -701,27 +701,27 @@ Cat & 1 \\ \hline

701 <itemizedlist>

702 <listitem>

703 <para>

704− I hat: Î

704+ I hat: Î

705 </para>

706 </listitem>

707 <listitem>

708 <para>

709− o umlaut: ö

709+ o umlaut: ö

710 </para>

711 </listitem>

712 <listitem>

713 <para>

714− section: §

714+ section: §

715 </para>

716 </listitem>

717 <listitem>

718 <para>

719− set membership: ∈

719+ set membership: ∈

720 </para>

721 </listitem>

722 <listitem>

723 <para>

725 </para>

726 </listitem>

727 </itemizedlist>