4e57451 · typeable/pandoc

git.iko.soy/typeable/pandoc

Use entities for all characters above 127 in docbook output.

Though XML tools should support unicode, some people will be using SGML tools, and these do not. Using entities makes the docbook files more portable. Also refactored encodeEntities and charToHtmlEntity in HtmlEntities.hs git-svn-id: https://pandoc.googlecode.com/svn/trunk@394 788f1e2b-df1e-0410-8736-df70ead52e1b

commit4e5745134a8867d68a4678285272869d1c0ebce3 authorfiddlosopher <fiddlosopher@788f1e2b-df1e-0410-8736-df70ead52e1b> dateTue 2 Jan 2007, 00:29 (20 years ago) parent2716943

3 files changed+22−17

compare with master

unifiedsplit

src/Text/Pandoc/HtmlEntities.hs+10−7 src/Text/Pandoc/Writers/Docbook.hs+7−5 tests/writer.docbook+5−5

src/Text/Pandoc/HtmlEntities.hs+10−7

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

5757-- | Returns a string with characters replaced with entity references where

5858-- possible.

5959encodeEntities :: 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)

6164

6265-- | If the string is a valid entity reference, returns @Just@ the character,

6366-- otherwise @Nothing@.

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

6972 Just (_, _, _, [sub]) -> Just (chr (read sub))

7073 Nothing -> Nothing

7174

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

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

74−charToHtmlEntity :: Char -> Maybe String

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

76+charToHtmlEntity :: Char -> String

7577charToHtmlEntity char =

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

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

79+ htmlEntityTable in

7780 if (length matches) == 0

78− then Nothing

79− else Just (fst (head matches))

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

82+ else fst (head matches)

8083

8184htmlEntityTable :: [(String, Char)]

8285htmlEntityTable = [

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

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

3333import Text.Pandoc.Definition

3434import Text.Pandoc.Shared

3535import Text.Pandoc.Writers.HTML ( stringToSmartHtml, stringToHtml )

36+import Text.Pandoc.HtmlEntities ( encodeEntities )

3637import Text.Html ( stringToHtmlString )

3738import Text.Regex ( mkRegex, matchRegex )

38−import Data.Char ( toLower )

39+import Data.Char ( toLower, ord )

3940import Data.List ( isPrefixOf, partition )

4041import Text.PrettyPrint.HughesPJ hiding ( Str )

4142

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

179180

180181-- | Escape a string for XML (with "smart" option if specified).

181182stringToXML :: WriterOptions -> String -> String

182−stringToXML options = if writerSmart options

183− then stringToSmartHtml

184− else stringToHtml

183+stringToXML options = encodeEntities .

184+ (if writerSmart options

185+ then stringToSmartHtml

186+ else stringToHtml)

185187

186188-- | Escape string to XML appropriate for attributes

187189attributeStringToXML :: String -> String

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

189191

190192-- | Escape a literal string for XML.

191193codeStringToXML :: String -> String

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

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

193195

194196-- | Convert a list of inline elements to Docbook.

195197inlinesToDocbook :: WriterOptions -> [Inline] -> Doc

tests/writer.docbook+5−5

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

701701 <itemizedlist>

702702 <listitem>

703703 <para>

704− I hat: Î

704+ I hat: Î

705705 </para>

706706 </listitem>

707707 <listitem>

708708 <para>

709− o umlaut: ö

709+ o umlaut: ö

710710 </para>

711711 </listitem>

712712 <listitem>

713713 <para>

714− section: §

714+ section: §

715715 </para>

716716 </listitem>

717717 <listitem>

718718 <para>

719− set membership: ∈

719+ set membership: ∈

720720 </para>

721721 </listitem>

722722 <listitem>

723723 <para>

725725 </para>

726726 </listitem>

727727 </itemizedlist>