diff options
| author | John MacFarlane <jgm@berkeley.edu> | 2012-09-06 13:49:43 -0700 |
|---|---|---|
| committer | John MacFarlane <jgm@berkeley.edu> | 2012-09-06 14:08:59 -0700 |
| commit | 37c29bfa3d3dda3cb4a59a3026c28c9cf202daba (patch) | |
| tree | 2f9d57ad82abe43dd906499c7cd16c65056c9bb5 | |
| parent | c2b520fb3602722d39031697227971a57e5bd0a2 (diff) | |
| download | pandoc-37c29bfa3d3dda3cb4a59a3026c28c9cf202daba.tar.gz | |
Textile reader: Improved speed of hyphenedWords.
This speeds up the textile reader by about a factor of 4.
But the reader is still very slow, compared to others readers.
| -rw-r--r-- | Makefile | 2 | ||||
| -rw-r--r-- | src/Text/Pandoc/Readers/Textile.hs | 10 |
2 files changed, 7 insertions, 5 deletions
@@ -7,7 +7,7 @@ all: cabal-dev configure --enable-tests --enable-benchmarks && cabal-dev build prof: - cabal-dev configure --enable-library-profiling --enable-executable-profiling && cabal-dev build + cabal-dev configure --enable-tests --enable-library-profiling --enable-executable-profiling && cabal-dev build prep: pandoc-types citeproc-hs cabal-dev install-deps --enable-library-profiling --enable-tests --enable-benchmarks diff --git a/src/Text/Pandoc/Readers/Textile.hs b/src/Text/Pandoc/Readers/Textile.hs index 1be2e3cd1..dc95d9a56 100644 --- a/src/Text/Pandoc/Readers/Textile.hs +++ b/src/Text/Pandoc/Readers/Textile.hs @@ -61,6 +61,7 @@ import Text.Pandoc.Parsing import Text.Pandoc.Readers.HTML ( htmlTag, isInlineTag, isBlockTag ) import Text.Pandoc.Readers.LaTeX ( rawLaTeXInline, rawLaTeXBlock ) import Text.HTML.TagSoup.Match +import Data.List ( intercalate ) import Data.Char ( digitToInt, isUpper ) import Control.Monad ( guard, liftM ) import Control.Applicative ((<$>), (*>), (<*)) @@ -426,14 +427,15 @@ wordBoundaries = markupChars ++ stringBreakers -- | Parse a hyphened sequence of words hyphenedWords :: Parser [Char] ParserState String -hyphenedWords = try $ do +hyphenedWords = intercalate "-" <$> sepBy1 wordChunk (char '-') + +wordChunk :: Parser [Char] ParserState String +wordChunk = try $ do hd <- noneOf wordBoundaries tl <- many ( (noneOf wordBoundaries) <|> try (notFollowedBy' note *> oneOf markupChars <* lookAhead (noneOf wordBoundaries) ) ) - let wd = hd:tl - option wd $ try $ - (\r -> concat [wd, "-", r]) <$> (char '-' *> hyphenedWords) + return $ hd:tl -- | Any string str :: Parser [Char] ParserState Inline |