feat: Implement encoding in rules file for -f *.rules

We use `text-icu` as [recommended by `text`](https://hackage.haskell.org/package/text-2.1.2/docs/Data-Text-Encoding.html):

> To gain access to a much larger family of encodings, use the `text-icu` package.

hledger-lib/Hledger/Read/RulesReader.hs

@@ -44,6 +44,7 @@ where
 --- ** imports
 import Prelude hiding (Applicative(..))
 import Control.Applicative (Applicative(..))
+import qualified Control.Exception as C
 import Control.Monad              (unless, when, void)
 import Control.Monad.Except       (ExceptT(..), liftEither, throwError)
 import qualified Control.Monad.Fail as Fail
@@ -65,6 +66,8 @@ import Data.Text (Text)
 import qualified Data.Text as T
 import qualified Data.Text.Encoding as T
 import qualified Data.Text.IO as T
+import Data.Text.ICU.Error (ICUError)
+import qualified Data.Text.ICU.Convert as UCNV
 import Data.Time ( Day, TimeZone, UTCTime, LocalTime, ZonedTime(ZonedTime),
   defaultTimeLocale, getCurrentTimeZone, localDay, parseTimeM, utcToLocalTime, localTimeToUTC, zonedTimeToUTC)
 import Safe (atMay, headMay, lastMay, readMay)
@@ -132,14 +135,20 @@ parse iopts f _ = do
             Nothing  -> return [maybe err (dbg4 "inferred source") $ dataFileFor f]  -- shouldn't fail, f has .rules extension
               where err = error' $ "could not infer a data file for " <> f
     return $ dbg4 "data file" $ headMay fs
+  mconverter <- do
+    case T.unpack <$> getDirective "encoding" rules of
+      Just enc -> Just <$> do
+        let ioconverter = UCNV.open (dbg4 "encoding" enc) (Just False)
+        ExceptT $ (Right . dbg4 "converter" <$> ioconverter) `C.catch` (\(_::ICUError) -> return $ Left "could not open ICU converter")
+      Nothing  -> return Nothing
   case mdatafile of
     Nothing -> return nulljournal  -- data file specified by source rule was not found
     Just dat -> do
       exists <- liftIO $ doesFileExist dat
       if not (dat=="-" || exists)
       then return nulljournal      -- data file inferred from rules file name was not found
       else do
-        t <- liftIO $ readFileOrStdinPortably dat
+        t <- liftIO $ readFileOrStdinPortably' mconverter dat
         readJournalFromCsv (Just $ Left rules) dat t Nothing
         -- apply any command line account aliases. Can fail with a bad replacement pattern.
         >>= liftEither . journalApplyAliases (aliasesFromOpts iopts)
@@ -500,6 +509,7 @@ directivep = (do
 directives :: [Text]
 directives =
   ["source"
+  ,"encoding"
   ,"date-format"
   ,"decimal-mark"
   ,"separator"

hledger-lib/Hledger/Utils/IO.hs

@@ -5,6 +5,7 @@ terminals, pager output, ANSI colour/styles, etc.
 -}
 
 {-# LANGUAGE LambdaCase          #-}
+{-# LANGUAGE OverloadedStrings #-}
 {-# LANGUAGE PackageImports      #-}
 {-# LANGUAGE ScopedTypeVariables #-}
 
@@ -32,9 +33,11 @@ module Hledger.Utils.IO (
   expandGlob,
   sortByModTime,
   readFileOrStdinPortably,
+  readFileOrStdinPortably',
   readFileStrictly,
   readFilePortably,
   readHandlePortably,
+  readHandlePortably',
   -- hereFileRelative,
 
   -- * Command line parsing
@@ -107,6 +110,7 @@ where
 import           Control.Concurrent (forkIO)
 import           Control.Exception (catch, evaluate, throwIO)
 import           Control.Monad (when, forM, guard, void)
+import qualified Data.ByteString as B
 import           Data.Char (toLower)
 import           Data.Colour.RGBSpace (RGB(RGB))
 import           Data.Colour.RGBSpace.HSL (lightness)
@@ -118,6 +122,7 @@ import           Data.Maybe (isJust, catMaybes)
 import           Data.Ord (comparing, Down (Down))
 import qualified Data.Text as T
 import qualified Data.Text.IO as T
+import qualified Data.Text.ICU.Convert as UCNV
 import qualified Data.Text.Lazy as TL
 import qualified Data.Text.Lazy.Builder as TB
 import           Data.Time.Clock (getCurrentTime)
@@ -280,19 +285,30 @@ readFilePortably f =  openFile f ReadMode >>= readHandlePortably
 
 -- | Like readFilePortably, but read from standard input if the path is "-".
 readFileOrStdinPortably :: String -> IO T.Text
-readFileOrStdinPortably f = openFileOrStdin f ReadMode >>= readHandlePortably
+readFileOrStdinPortably = readFileOrStdinPortably' Nothing
+
+-- | Like readFileOrStdinPortably, but take an optional converter.
+readFileOrStdinPortably' :: Maybe UCNV.Converter -> String -> IO T.Text
+readFileOrStdinPortably' c f = openFileOrStdin f >>= readHandlePortably' c
   where
-    openFileOrStdin :: String -> IOMode -> IO Handle
-    openFileOrStdin "-" _ = return stdin
-    openFileOrStdin f' m   = openFile f' m
+    openFileOrStdin :: String -> IO Handle
+    openFileOrStdin "-" = return stdin
+    openFileOrStdin f' = openFile f' ReadMode
 
 readHandlePortably :: Handle -> IO T.Text
-readHandlePortably h = do
+readHandlePortably = readHandlePortably' Nothing
+
+readHandlePortably' :: Maybe UCNV.Converter -> Handle -> IO T.Text
+readHandlePortably' Nothing h = do
   hSetNewlineMode h universalNewlineMode
   menc <- hGetEncoding h
   when (fmap show menc == Just "UTF-8") $  -- XXX no Eq instance, rely on Show
     hSetEncoding h utf8_bom
   T.hGetContents h
+readHandlePortably' (Just c) h = do
+  -- We need to manually apply the newline mode
+  -- Since we already have a Text
+  T.replace "\r\n" "\n". UCNV.toUnicode c <$> B.hGetContents h
 
 -- | Like embedFile, but takes a path relative to the package directory.
 embedFileRelative :: FilePath -> Q Exp

hledger-lib/hledger-lib.cabal

@@ -160,6 +160,7 @@ library
     , template-haskell
     , terminal-size >=0.3.3
     , text >=1.2.4.1
+    , text-icu >=0.8.0.5
     , time >=1.5
     , timeit
     , transformers >=0.2

hledger-lib/package.yaml

@@ -78,6 +78,7 @@ dependencies:
 - template-haskell
 - terminal-size >=0.3.3
 - text >=1.2.4.1
+- text-icu >=0.8.0.5
 - time >=1.5
 - timeit
 - transformers >=0.2