ISO-2022-JP な行と UTF-8 な行の混ざったファイルを読んでみるテスト

convtest.hs

import Codec.Text.IConv as IConv
import Codec.Binary.UTF8.String as UTF8
import Data.ByteString as B
import Data.ByteString.Lazy as BL
import System.IO as IO

convertDetect :: [ IConv.EncodingName ] -> IConv.EncodingName -> BL.ByteString ->
                 Either BL.ByteString (IConv.EncodingName, BL.ByteString)
convertDetect [] _ src = Left src
convertDetect (srcEnc:srcEncs) dstEnc src = case IConv.convertStrictly srcEnc dstEnc src of
  Left  dst -> Right (srcEnc, dst)
  Right _   -> convertDetect srcEncs dstEnc src

hGetConvertedLine' :: [ IConv.EncodingName ] -> IO.Handle ->
                      IO (Either BL.ByteString (IConv.EncodingName, String))
hGetConvertedLine' srcEncs h = do
  l <- B.hGetLine h
  return $ case convertDetect srcEncs "UTF-8" $ BL.fromChunks [l] of
    Left  srcBin           -> Left srcBin
    Right (srcEnc, dstBin) -> Right (srcEnc, UTF8.decode $ BL.unpack dstBin)

hGetConvertedLine :: [ IConv.EncodingName ] -> IO.Handle -> IO (Either BL.ByteString String)
hGetConvertedLine srcEncs h = do
  result <- hGetConvertedLine' srcEncs h
  return $ case result of
    Left  srcBin      -> Left srcBin
    Right (_, dstStr) -> Right dstStr

hRepeatToEOF :: (IO.Handle -> IO a) -> IO.Handle -> IO [a]
hRepeatToEOF op h = do
  eof <- IO.hIsEOF h
  if eof
    then return []
    else do
         e <- op h
         es <- hRepeatToEOF op h
         return $ e : es

myHGetConvertedLine = hGetConvertedLine' ["ISO-2022-JP","UTF-8"]
testFile = "/home/plaster/work/_testdata/2011/02/02/log"

main = do
  ls <- IO.withFile testFile IO.ReadMode $ hRepeatToEOF myHGetConvertedLine
  mapM print ls