3abc102 · typeable/pandoc

git.iko.soy/typeable/pandoc

HTML writer: Use toHtml instead of pre-escaping.

We work around the problem that blaze-html unnecessarily escapes `'` by pre-escaping just the `'` characters, instead of the whole string. If blaze-html later stops escaping `'` characters, we can simplify strToHtml to toHtml. Note that this change yields a significant speed boost (111ms to 94ms on one benchmark). Closes #629.

commit3abc1021aae3cbba698ff8e27fc809647461248e authorJohn MacFarlane <jgm@berkeley.edu> dateFri 28 Sep 2012, 11:11 (14 years ago) parent4be1375

1 file changed+7−4

compare with master

unifiedsplit

src/Text/Pandoc/Writers/HTML.hs+7−4

src/Text/Pandoc/Writers/HTML.hs+7−4

@@ -39,7 +39,7 @@ import Text.Pandoc.Readers.TeXMath

3939import Text.Pandoc.Slides

4040import Text.Pandoc.Highlighting ( highlight, styleToCss,

4141 formatHtmlInline, formatHtmlBlock )

42−import Text.Pandoc.XML (stripTags, escapeStringForXML, fromEntities)

42+import Text.Pandoc.XML (stripTags, fromEntities)

4343import Network.HTTP ( urlEncode )

4444import Numeric ( showHex )

4545import Data.Char ( ord, toLower )

@@ -60,7 +60,7 @@ import Text.Blaze.Renderer.String (renderHtml)

6060import Text.TeXMath

6161import Text.XML.Light.Output

6262import System.FilePath (takeExtension)

63−import Data.Monoid (mempty, mconcat)

63+import Data.Monoid

6464

6565data WriterState = WriterState

6666 { stNotes :: [Html] -- ^ List of notes

@@ -77,8 +77,11 @@ defaultWriterState = WriterState {stNotes= [], stMath = False, stQuotes = False,

7777-- Helpers to render HTML with the appropriate function.

7878

7979strToHtml :: String -> Html

80−strToHtml = preEscapedString . escapeStringForXML

81−-- strToHtml = toHtml

80+strToHtml ('\'':xs) = preEscapedString "\'" `mappend` strToHtml xs

81+strToHtml xs@(_:_) = case break (=='\'') xs of

82+ (_ ,[]) -> toHtml xs

83+ (ys,zs) -> toHtml ys `mappend` strToHtml zs

84+strToHtml [] = ""

8285

8386-- | Hard linebreak.

8487nl :: WriterOptions -> Html