ebcebcc · typeable/pandoc

git.iko.soy/typeable/pandoc

on master help sign in

Use utf8-string's System.IO.UTF8 to replace Text.Pandoc.UTF8.

ebcebccfiddlosopher18 years ago

Main.hs+12−10

compare with master

unifiedsplit

@@ -30,7 +30,6 @@ writers.

30-}

31module Main where

32import Text.Pandoc

33−import Text.Pandoc.UTF8

34import Text.Pandoc.ODT

33import Text.Pandoc.ODT

35import Text.Pandoc.Shared ( joinWithSep, HTMLMathMethod (..) )

34import Text.Pandoc.Shared ( joinWithSep, HTMLMathMethod (..) )

36import Text.Pandoc.Highlighting ( languages )

35import Text.Pandoc.Highlighting ( languages )

@@ -38,7 +37,9 @@ import System.Environment ( getArgs, getProgName, getEnvironment )

38import System.Exit ( exitWith, ExitCode (..) )

37import System.Exit ( exitWith, ExitCode (..) )

39import System.FilePath ( takeExtension, takeDirectory )

38import System.FilePath ( takeExtension, takeDirectory )

40import System.Console.GetOpt

39import System.Console.GetOpt

40+import Prelude hiding ( putStrLn, writeFile, readFile, getContents )

41−import System.IO

41+import System.IO ( stdout, stderr )

42+import System.IO.UTF8

42import Data.Maybe ( fromMaybe )

43import Data.Maybe ( fromMaybe )

43import Data.Char ( toLower )

44import Data.Char ( toLower )

44import Control.Monad ( (>>=) )

45import Control.Monad ( (>>=) )

@@ -275,7 +276,7 @@ options =

275 (\arg opt -> do

276 (\arg opt -> do

276 let old = optIncludeInHeader opt

277 let old = optIncludeInHeader opt

277 text <- readFile arg

278 text <- readFile arg

278− return opt { optIncludeInHeader = old ++ fromUTF8 text,

279+ return opt { optIncludeInHeader = old ++ text,

279 optStandalone = True })

280 optStandalone = True })

280 "FILENAME")

281 "FILENAME")

281 "" -- "File to include at end of header (implies -s)"

282 "" -- "File to include at end of header (implies -s)"

@@ -285,7 +286,7 @@ options =

285 (\arg opt -> do

286 (\arg opt -> do

286 let old = optIncludeBeforeBody opt

287 let old = optIncludeBeforeBody opt

287 text <- readFile arg

288 text <- readFile arg

288− return opt { optIncludeBeforeBody = old ++ fromUTF8 text })

289+ return opt { optIncludeBeforeBody = old ++ text })

289 "FILENAME")

290 "FILENAME")

290 "" -- "File to include before document body"

291 "" -- "File to include before document body"

291

292

@@ -294,7 +295,7 @@ options =

294 (\arg opt -> do

295 (\arg opt -> do

295 let old = optIncludeAfterBody opt

296 let old = optIncludeAfterBody opt

296 text <- readFile arg

297 text <- readFile arg

297− return opt { optIncludeAfterBody = old ++ fromUTF8 text })

298+ return opt { optIncludeAfterBody = old ++ text })

298 "FILENAME")

299 "FILENAME")

299 "" -- "File to include after document body"

300 "" -- "File to include after document body"

300

301

@@ -302,7 +303,7 @@ options =

302 (ReqArg

303 (ReqArg

303 (\arg opt -> do

304 (\arg opt -> do

304 text <- readFile arg

305 text <- readFile arg

305− return opt { optCustomHeader = fromUTF8 text,

306+ return opt { optCustomHeader = text,

306 optStandalone = True })

307 optStandalone = True })

307 "FILENAME")

308 "FILENAME")

308 "" -- "File to use for custom header (implies -s)"

309 "" -- "File to use for custom header (implies -s)"

@@ -555,10 +556,11 @@ main = do

555 then putStrLn

556 then putStrLn

556 else writeFile outputFile . (++ "\n")

557 else writeFile outputFile . (++ "\n")

557

558

558− (readSources sources) >>= writeOutput . toUTF8 .

559+ (readSources sources) >>= writeOutput .

559− (writer writerOptions) .

560+ writer writerOptions .

560− (reader startParserState) . tabFilter tabStop .

561+ reader startParserState .

562+ tabFilter tabStop .

561− fromUTF8 . (joinWithSep "\n")

563+ joinWithSep "\n"

562

564

563 where

565 where

564 readSources [] = mapM readSource ["-"]

566 readSources [] = mapM readSource ["-"]

Text/Pandoc.hs+5−7

@@ -36,13 +36,14 @@ inline links:

37> module Main where

38> import Text.Pandoc

39−>

39+> import qualified System.IO.UTF8 as U

40+>

40> markdownToRST :: String -> String

41> markdownToRST :: String -> String

41−> markdownToRST = toUTF8 .

42+> markdownToRST =

42> (writeRST defaultWriterOptions {writerReferenceLinks = True}) .

43> (writeRST defaultWriterOptions {writerReferenceLinks = True}) .

43−> (readMarkdown defaultParserState) . fromUTF8

44+> readMarkdown defaultParserState

44>

45>

45−> main = interact markdownToRST

46+> main = U.getContents >>= U.putStrLn . markdownToRST

47-}

48-}

@@ -84,8 +85,6 @@ module Text.Pandoc

84 , defaultWriterOptions

85 , defaultWriterOptions

85 -- * Default headers for various output formats

86 -- * Default headers for various output formats

86 , module Text.Pandoc.DefaultHeaders

87 , module Text.Pandoc.DefaultHeaders

87− -- * Functions for converting to and from UTF-8

88− , module Text.Pandoc.UTF8

89 -- * Version

88 -- * Version

90 , pandocVersion

89 , pandocVersion

91 ) where

90 ) where

@@ -108,7 +107,6 @@ import Text.Pandoc.Writers.Man

108import Text.Pandoc.Writers.RTF

107import Text.Pandoc.Writers.RTF

109import Text.Pandoc.Writers.MediaWiki

108import Text.Pandoc.Writers.MediaWiki

110import Text.Pandoc.DefaultHeaders

109import Text.Pandoc.DefaultHeaders

111−import Text.Pandoc.UTF8

112import Text.Pandoc.Shared

110import Text.Pandoc.Shared

113

111

114-- | Version number of pandoc library.

112-- | Version number of pandoc library.

Text/Pandoc/ODT.hs−1

@@ -92,7 +92,6 @@ handlePictures tempODT sourceDirRelative xml = do

92 Just x -> x

93 cursor' <- scanPictures tempODT sourceDirRelative cursor

94 let modified = parsed { elContent = toForest $ root cursor' }

95− putStrLn $ showTopElement modified

96 return $ showTopElement modified

95 return $ showTopElement modified

98scanPictures :: FilePath -> FilePath -> Cursor -> IO Cursor

97scanPictures :: FilePath -> FilePath -> Cursor -> IO Cursor

Text/Pandoc/UTF8.hsdeleted−45

@@ -1,45 +0,0 @@

1−-- | Functions for converting Unicode strings to UTF-8 and vice versa.

2−--

3−-- Taken from <http://www.cse.ogi.edu/~hallgren/Talks/LHiH/base/lib/UTF8.hs>.

5−-- Science University.

6−--

7−-- Modified by Martin Norbaeck

8−-- to pass illegal UTF-8 sequences through unchanged.

9−module Text.Pandoc.UTF8 (

10− fromUTF8,

11− toUTF8

12− ) where

13−

14−-- From the Char module supplied with HBC.

15−

16−-- | Take a UTF-8 string and decode it into a Unicode string.

17−fromUTF8 :: String -> String

18−fromUTF8 "" = ""

19−fromUTF8 ('\xef':'\xbb':'\xbf':cs) = fromUTF8 cs -- skip BOM (byte order marker)

20−fromUTF8 (c:c':cs) | '\xc0' <= c && c <= '\xdf' &&

21− '\x80' <= c' && c' <= '\xbf' =

22− toEnum ((fromEnum c `mod` 0x20) * 0x40 + fromEnum c' `mod` 0x40) : fromUTF8 cs

23−fromUTF8 (c:c':c'':cs) | '\xe0' <= c && c <= '\xef' &&

24− '\x80' <= c' && c' <= '\xbf' &&

25− '\x80' <= c'' && c'' <= '\xbf' =

26− toEnum ((fromEnum c `mod` 0x10 * 0x1000) + (fromEnum c' `mod` 0x40) * 0x40 + fromEnum c'' `mod` 0x40) : fromUTF8 cs

27−fromUTF8 (c:cs) = c : fromUTF8 cs

28−

29−-- | Take a Unicode string and encode it as a UTF-8 string.

30−toUTF8 :: String -> String

31−toUTF8 "" = ""

32−toUTF8 (c:cs) =

33− if c > '\x0000' && c < '\x0080' then

34− c : toUTF8 cs

35− else if c < toEnum 0x0800 then

36− let i = fromEnum c

37− in toEnum (0xc0 + i `div` 0x40) :

38− toEnum (0x80 + i `mod` 0x40) :

39− toUTF8 cs

40− else

41− let i = fromEnum c

42− in toEnum (0xe0 + i `div` 0x1000) :

43− toEnum (0x80 + (i `mod` 0x1000) `div` 0x40) :

44− toEnum (0x80 + i `mod` 0x40) :

45− toUTF8 cs

debian/copyright−36

38Released under the GPL.

40−----------------------------------------------------------------------

41−UTF8.hs

44−

45−Redistribution and use in source and binary forms, with or without

46−modification, are permitted provided that the following conditions

47−are met:

48−

49−- Redistributions of source code must retain the above copyright

50− notice, this list of conditions and the following disclaimer.

51−

52−- Redistributions in binary form must reproduce the above

53− copyright notice, this list of conditions and the following

54− disclaimer in the documentation and/or other materials provided

55− with the distribution.

56−

57−- Neither the name of OGI or OHSU nor the names of its

58− contributors may be used to endorse or promote products derived

59− from this software without specific prior written permission.

60−

61−THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS

62−"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT

63−LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR

64−A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT

65−OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,

66−SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED

67−TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR

68−PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF

69−LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING

70−NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS

71−SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

72−

73−This version of UTF8.hs was Modified by Martin Norbäck, to pass illegal

74−utf-8 sequences through unchanged.

75−

76----------------------------------------------------------------------

40----------------------------------------------------------------------

77ASCIIMathML.js

41ASCIIMathML.js

pandoc.cabal+1−2

@@ -80,14 +80,13 @@ Library

80 Build-depends: highlighting-kate

81 cpp-options: -DHIGHLIGHTING

82 Build-Depends: parsec < 3, xhtml, mtl, network, filepath, process, directory, xml,

83− bytestring, binary

83+ bytestring, binary, utf8-string

84 Hs-Source-Dirs: .

85 Exposed-Modules: Text.Pandoc,

86 Text.Pandoc.Blocks,

87 Text.Pandoc.Definition,

88 Text.Pandoc.CharacterReferences,

89 Text.Pandoc.Shared,

90− Text.Pandoc.UTF8,

91 Text.Pandoc.ODT,

90 Text.Pandoc.ODT,

92 Text.Pandoc.ASCIIMathML,

91 Text.Pandoc.ASCIIMathML,

93 Text.Pandoc.DefaultHeaders,

92 Text.Pandoc.DefaultHeaders,