fix: loading utf-8 file when locale encoding is set to ascii

CHANGELOG.md

@@ -1,5 +1,9 @@
 # Revision history for configurator-pg
 
+# Unreleased
+
+* Fix loading `utf-8` file when locale encoding is set to `ASCII`
+
 ## 0.2.10 -- 2024-03-06
 
 * Allow megaparsec-9.6

configurator-pg.cabal

@@ -34,6 +34,7 @@ library
                        Data.Configurator.Syntax
                        Data.Configurator.Types
   build-depends:       base                 >= 4.9 && < 4.22
+                     , bytestring           >= 0.10.8 && < 0.13
                      , megaparsec           >= 7.0.0 && < 9.8
                      , containers           >= 0.5.6.2 && < 0.8
                      , protolude            >= 0.1.10 && < 0.4

src/Data/Configurator/Load.hs

@@ -6,10 +6,12 @@ import Protolude
 
 import           Control.Exception                (throw)
 import           Text.Megaparsec                  (parse, errorBundlePretty)
+import qualified Data.ByteString                  as BS
 import qualified Data.Map.Strict                  as M
 import           Data.Scientific                  (toBoundedInteger,
                                                    toRealFloat)
 import qualified Data.Text                        as T
+import qualified Data.Text.Encoding               as T
 import qualified Data.Text.Lazy                   as TL
 import           Data.Text.Lazy.Builder           (fromString,
                                                    fromText,
@@ -31,7 +33,7 @@ load path = applyDirective "" "" M.empty (Import $ T.pack path)
 
 loadOne :: Path -> IO [Directive]
 loadOne path = do
-  s <- readFile (T.unpack path)
+  s <- T.decodeUtf8 <$> BS.readFile (T.unpack path)
   case parse topLevel (T.unpack path) s of
     Left err         -> throw $ ParseError $ T.pack $ errorBundlePretty err
     Right directives -> return directives

tests/Test.hs

@@ -1,12 +1,15 @@
 {-# LANGUAGE OverloadedStrings #-}
 {-# LANGUAGE ScopedTypeVariables #-}
+{-# LANGUAGE CPP #-}
 
 module Main where
 
 import Protolude hiding (bool, list, optional)
 
 import           Data.Configurator
 import qualified Data.Text                      as T
+import qualified GHC.IO.Encoding                as E
+import qualified GHC.IO.Encoding.Latin1         as E
 import           System.Environment
 import           System.FilePath
 import           Test.Framework
@@ -20,6 +23,7 @@ tests :: [Test]
 tests =
     [ testCase "read-simple" $ readTest "simple.cfg"
     , testCase "read-pathological" $ readTest "pathological.cfg"
+    , testCase "read-utf-8-with-ascii-locale" $ readTestWithLocale "utf-8.cfg" E.ascii
     , testCase "load" loadTest
     , testCase "load" loadTest
     , testCase "types" typesTest
@@ -51,6 +55,9 @@ testFile name = "tests" </> "resources" </> name
 errorFile :: FilePath -> FilePath
 errorFile name = testFile name <> ".err"
 
+errorFileGHC8 :: FilePath -> FilePath
+errorFileGHC8 name = testFile name <> ".err.ghc8"
+
 parse :: Config -> Parser Value a -> Key -> Either Text a
 parse cfg p key = runParser (required key p) cfg
 
@@ -63,6 +70,10 @@ parseSub cfg p prefix = runParser (subassocs prefix p) cfg
 readTest :: FilePath -> Assertion
 readTest file = load (testFile file) >> return ()
 
+readTestWithLocale :: FilePath -> E.TextEncoding -> Assertion
+readTestWithLocale file locale =
+  E.setLocaleEncoding locale >> load (testFile file) >> return ()
+
 loadTest :: Assertion
 loadTest =
   withLoad "pathological.cfg" $ \cfg -> do
@@ -233,7 +244,11 @@ parseErrorTest file = do
 
 ioErrorTest :: FilePath -> Assertion
 ioErrorTest file = do
+#if __GLASGOW_HASKELL__ >= 900
   err <- readFile $ errorFile file
+#else
+  err <- readFile $ errorFileGHC8 file
+#endif
   (load (testFile file) >> assertFailure "expected an IO error")
     `catch` \ (ex :: IOException) -> do
        assertEqual "" err (show ex)

tests/resources/err-import.cfg.err

@@ -1 +1 @@
-tests/resources/not-exist.cfg: openFile: does not exist (No such file or directory)
+tests/resources/not-exist.cfg: withBinaryFile: does not exist (No such file or directory)

tests/resources/err-import.cfg.err.ghc8

@@ -0,0 +1 @@
+tests/resources/not-exist.cfg: openBinaryFile: does not exist (No such file or directory)

tests/resources/utf-8.cfg

@@ -0,0 +1,2 @@
+# Commènt utf-8 chàrs
+utf-kèy = "utf-8-vàlue"