57df7f5 · typeable/pandoc

git.iko.soy/typeable/pandoc

HTML writer: override Text.XHtml's stringToHtml function,

so that characters below 0xff are not converted to numerical entity references. Also convert '\160' to " ". This should aid readability and editability of the HTML source. It does presuppose that the HTML will be served as UTF-8. git-svn-id: https://pandoc.googlecode.com/svn/trunk@1338 788f1e2b-df1e-0410-8736-df70ead52e1b

commit57df7f5eb6c7d9e258ffc017380efd3b3422add9 authorfiddlosopher <fiddlosopher@788f1e2b-df1e-0410-8736-df70ead52e1b> dateSun 27 Jul 2008, 03:54 (18 years ago) parentddaec05

2 files changed+18−4

compare with master

unifiedsplit

Text/Pandoc/Writers/HTML.hs+15−1 tests/writer.html+3−3

Text/Pandoc/Writers/HTML.hs+15−1

@@ -39,7 +39,7 @@ import Data.Char ( ord, toLower, isAlpha )

3939import Data.List ( isPrefixOf, intersperse )

4040import qualified Data.Set as S

4141import Control.Monad.State

42−import Text.XHtml.Transitional

42+import Text.XHtml.Transitional hiding ( stringToHtml )

4343

4444data WriterState = WriterState

4545 { stNotes :: [Html] -- ^ List of notes

@@ -62,6 +62,20 @@ renderFragment opts = if writerWrapText opts

6262 then renderHtmlFragment

6363 else showHtmlFragment

6464

65+-- | Slightly modified version of Text.XHtml's stringToHtml.

66+-- Only uses numerical entities for 0xff and greater.

67+-- Adds  .

68+stringToHtml :: String -> Html

69+stringToHtml = primHtml . concatMap fixChar

70+ where

71+ fixChar '<' = "<"

72+ fixChar '>' = ">"

73+ fixChar '&' = "&"

74+ fixChar '"' = """

75+ fixChar '\160' = " "

76+ fixChar c | ord c < 0xff = [c]

77+ fixChar c = "&#" ++ show (ord c) ++ ";"

78+

6579-- | Convert Pandoc document to Html string.

6680writeHtmlString :: WriterOptions -> Pandoc -> String

6781writeHtmlString opts =

tests/writer.html+3−3

@@ -412,7 +412,7 @@ These should not be escaped: \$ \\ \> \[ \{

412412 ><p

413413 >Should not be a list item:</p

414414 ><p

415− >M.A. 2007</p

415+ >M.A. 2007</p

416416 ><p

417417 >B. Williams</p

418418 ><hr

@@ -678,7 +678,7 @@ Blah

678678 >hello</em

679679 ></sup

680680 > a<sup

681− >hello there</sup

681+ >hello there</sup

682682 >.</p

683683 ><p

684684 >Subscripts: H<sub

@@ -686,7 +686,7 @@ Blah

686686 >O, H<sub

687687 >23</sub

688688 >O, H<sub

689− >many of them</sub

689+ >many of them</sub

690690 >O.</p

691691 ><p

692692 >These should not be superscripts or subscripts, because of the unescaped spaces: a^b c^d, a~b c~d.</p