Add full Unicode support

`dconf dump` includes the Unicode and the format exported by dconf2nix
such as `\129315` really means nothing to Nix.

> The functions that deal with GVariant text format absolutely always deal in UTF-8.
> Conceptually, GVariant text format is a string of Unicode characters, not bytes.
> Non-ASCII but otherwise printable Unicode characters are not treated any differently from normal ASCII characters.

Let’s ensure they are printed as such.

Also fix the test data from e2b5065,
they were copied as reported by `parserTraced` but the actual data
was mostly Unicode with few escape sequences.

data/dconf.settings

@@ -303,7 +303,7 @@ locations=[<(uint32 2, <('Gdańsk', 'EPGD', true, [(0.94916821905848536, 0.32230
 region='en_US.UTF-8'
 
 [issue28/desktop/ibus/panel/emoji]
-favorites=['\8211', '\8594', '\8593', '\8595', '\8482', '\\u00ad', '\176', '\\v', '\160', '\171', '\8451']
+favorites=['–', '→', '↑', '↓', '™', '\u00ad', '°', '\v', ' ', '«', '℃']
 
 [issue28/org/gnome/desktop/input-sources]
 mru-sources=[('xkb', 'us+altgr-intl'), ('ibus', 'mozc-jp')]

data/unicode.settings

@@ -0,0 +1,3 @@
+[org/gnome/font-manager]
+compare-preview-text='🤣'
+preview-text='příliš žluťoučký kůň úpěl ďábelské ódy\nprilis zlutoucky kun upel dabelske ody\n◌̍◌̍ff̍ ̍čěů\n◌̎\n'

output/dconf.nix

@@ -361,20 +361,20 @@ with lib.hm.gvariant;
     };
 
     "org/gnome/Weather" = {
-      locations = [ (mkVariant [ (mkUint32 2) (mkVariant [ "Gda\324sk" "EPGD" true [ (mkTuple [ 0.9491682190584854 0.3223041410193837 ]) ] [ (mkTuple [ 0.9485864484589182 0.32579479952337237 ]) ] ]) ]) (mkVariant [ (mkUint32 2) (mkVariant [ "Gdynia, Dzia\322dowo County, Warmian-Masurian Voivodeship" "" false [ (mkTuple [ 0.9302794944578734 0.34699627038777753 ]) ] [ (mkTuple [ 0.938610530426954 0.3574455077502486 ]) ] ]) ]) (mkVariant [ (mkUint32 2) (mkVariant [ "Gdynia, Pomeranian Voivodeship" "" false [ (mkTuple [ 0.9514923902475622 0.3235888220312407 ]) ] [ (mkTuple [ 0.9485864484589182 0.32579479952337237 ]) ] ]) ]) ];
+      locations = [ (mkVariant [ (mkUint32 2) (mkVariant [ "Gdańsk" "EPGD" true [ (mkTuple [ 0.9491682190584854 0.3223041410193837 ]) ] [ (mkTuple [ 0.9485864484589182 0.32579479952337237 ]) ] ]) ]) (mkVariant [ (mkUint32 2) (mkVariant [ "Gdynia, Działdowo County, Warmian-Masurian Voivodeship" "" false [ (mkTuple [ 0.9302794944578734 0.34699627038777753 ]) ] [ (mkTuple [ 0.938610530426954 0.3574455077502486 ]) ] ]) ]) (mkVariant [ (mkUint32 2) (mkVariant [ "Gdynia, Pomeranian Voivodeship" "" false [ (mkTuple [ 0.9514923902475622 0.3235888220312407 ]) ] [ (mkTuple [ 0.9485864484589182 0.32579479952337237 ]) ] ]) ]) ];
     };
 
     "org/gnome/shell/weather" = {
       automatic-location = true;
-      locations = [ (mkVariant [ (mkUint32 2) (mkVariant [ "Gda\324sk" "EPGD" true [ (mkTuple [ 0.9491682190584854 0.3223041410193837 ]) ] [ (mkTuple [ 0.9485864484589182 0.32579479952337237 ]) ] ]) ]) (mkVariant [ (mkUint32 2) (mkVariant [ "Gdynia, Dzia\322dowo County, Warmian-Masurian Voivodeship" "" false [ (mkTuple [ 0.9302794944578734 0.34699627038777753 ]) ] [ (mkTuple [ 0.938610530426954 0.3574455077502486 ]) ] ]) ]) (mkVariant [ (mkUint32 2) (mkVariant [ "Gdynia, Pomeranian Voivodeship" "" false [ (mkTuple [ 0.9514923902475622 0.3235888220312407 ]) ] [ (mkTuple [ 0.9485864484589182 0.32579479952337237 ]) ] ]) ]) ];
+      locations = [ (mkVariant [ (mkUint32 2) (mkVariant [ "Gdańsk" "EPGD" true [ (mkTuple [ 0.9491682190584854 0.3223041410193837 ]) ] [ (mkTuple [ 0.9485864484589182 0.32579479952337237 ]) ] ]) ]) (mkVariant [ (mkUint32 2) (mkVariant [ "Gdynia, Działdowo County, Warmian-Masurian Voivodeship" "" false [ (mkTuple [ 0.9302794944578734 0.34699627038777753 ]) ] [ (mkTuple [ 0.938610530426954 0.3574455077502486 ]) ] ]) ]) (mkVariant [ (mkUint32 2) (mkVariant [ "Gdynia, Pomeranian Voivodeship" "" false [ (mkTuple [ 0.9514923902475622 0.3235888220312407 ]) ] [ (mkTuple [ 0.9485864484589182 0.32579479952337237 ]) ] ]) ]) ];
     };
 
     "system/locale" = {
       region = "en_US.UTF-8";
     };
 
     "issue28/desktop/ibus/panel/emoji" = {
-      favorites = [ "\137" "\396" "\395" "\397" "\322" "\\u00ad" "\176" "\\v" "\160" "\171" "\297" ];
+      favorites = [ "–" "→" "↑" "↓" "™" "­" "°" "" " " "«" "℃" ];
     };
 
     "issue28/org/gnome/desktop/input-sources" = {

output/unicode.nix

@@ -0,0 +1,14 @@
+# Generated via dconf2nix: https://github.com/gvolpe/dconf2nix
+{ lib, ... }:
+
+with lib.hm.gvariant;
+
+{
+  dconf.settings = {
+    "org/gnome/font-manager" = {
+      compare-preview-text = "🤣";
+      preview-text = "příliš žluťoučký kůň úpěl ďábelské ódy\nprilis zlutoucky kun upel dabelske ody\n◌̍◌̍ff̍ ̍čěů\n◌̎\n";
+    };
+
+  };
+}

src/DConf.hs

@@ -7,7 +7,10 @@ module DConf
   )
 where
 
+import           Control.Monad                  ( replicateM )
+import           Data.Ix                        ( inRange )
 import qualified Data.Map                      as Map
+import           Data.Maybe                     ( catMaybes )
 import           Data.Text                      ( Text )
 import qualified Data.Text                     as T
 import           DConf.Data
@@ -66,24 +69,43 @@ vString = T.pack <$> (single <|> double)
  where
   single = bracket "'" "'" $ inputs "'"
   double = bracket "\"" "\"" $ inputs "\""
+
   lchar :: [Char] -> Parsec Text () Char
-  lchar extra = charExcept $ "\r\n" <> extra
-  octal = do
-    isOctal <- optionMaybe $ char '8'
-    let
-      base = case isOctal of
-        Just _ -> 8
-        Nothing -> 10
-    a <- digit
-    b <- digit
-    c <- digit
-    return $ chr $ (digitToInt a * base * base) + (digitToInt b * base) + digitToInt c
-  qchar :: Parsec Text () Char
+  lchar extra = charExcept $ "\r\n\\" <> extra
+
+  fromHexDigit :: Char -> Int
+  fromHexDigit n | inRange ('A', 'F') n = ord n - ord 'A' + 0xA
+  fromHexDigit n | inRange ('a', 'f') n = ord n - ord 'a' + 0xA
+  fromHexDigit n | inRange ('0', '9') n = ord n - ord '0'
+  fromHexDigit n = error $ "Expected a hexadecimal digit, '" ++ n : "' given"
+
+  hexNum :: Int -> Parsec Text () Int
+  hexNum l = do
+    digits <- replicateM l (fromHexDigit <$> hexDigit)
+    return $ foldl (\acc d -> 16 * acc + d) 0 digits
+
+  qchar :: Parsec Text () (Maybe Char)
   qchar = do
     _ <- char '\\'
-    octal <|> anyChar
+    (
+      -- Unicode escapes of the form `\uxxxx` and `\Uxxxxxxxx` are supported, in hexadecimal.
+      (char 'u' *> (Just <$> (chr <$> hexNum 4)))
+      <|> (char 'U' *> (Just <$> (chr <$> hexNum 8)))
+      -- The usual control sequence escapes `\a`, `\b`, `\f`, `\n`, `\r`, `\t` and `\v` are supported.
+      <|> (char 'a' *> pure (Just '\a'))
+      <|> (char 'b' *> pure (Just '\b'))
+      <|> (char 'f' *> pure (Just '\f'))
+      <|> (char 'n' *> pure (Just '\n'))
+      <|> (char 'r' *> pure (Just '\r'))
+      <|> (char 't' *> pure (Just '\t'))
+      <|> (char 'v' *> pure (Just '\v'))
+      -- Additionally, a `\` before a newline character causes the newline to be ignored.
+      <|> (char '\n' *> pure Nothing)
+      -- Finally, any other character following `\` is copied literally (for example, `\"` or `\\`)
+      <|> (Just <$> anyChar))
+
   inputs :: [Char] -> Parsec Text () String
-  inputs extra = many $ qchar <|> lchar extra
+  inputs extra = catMaybes <$> (many $ qchar <|> (Just <$> lchar extra))
 
 value :: Parsec Text () Value
 value = choice

src/Nix.hs

@@ -9,6 +9,7 @@ module Nix
   )
 where
 
+import           Data.Function                  ( (&) )
 import qualified Data.Map                      as Map
 import qualified Data.Text                     as T
 import           DConf.Data
@@ -68,7 +69,7 @@ renderValue raw = Nix $ renderValue' raw <> ";"
   renderList xs = let
     in "[ " <> T.intercalate " " (renderItem <$> xs) <> " ]"
 
-  renderValue' (S   v) = T.pack $ show v
+  renderValue' (S   v) = renderString v
   renderValue' (B   v) = T.toLower . T.pack $ show v
   renderValue' (I   v) = T.pack $ show v
   renderValue' (D   v) = T.pack $ show v
@@ -81,3 +82,13 @@ renderValue raw = Nix $ renderValue' raw <> ";"
     "''\n" <> mkSpaces 8 <> T.strip v <> "\n" <> mkSpaces 6 <> "''"
   renderValue' (R kvs) =
     "{\n" <> mconcat (fmap (\(k,v) -> mkSpaces 8 <> k <> " = " <> renderValue' v <> ";\n") kvs) <> mkSpaces 6 <> "}"
+
+renderString :: T.Text -> T.Text
+renderString text = "\"" <> escaped <> "\""
+  where
+    escaped =
+      text
+        & T.replace "\\" "\\\\"
+        & T.replace "\n" "\\n"
+        & T.replace "$" "\\$"
+        & T.replace "\"" "\\\""

test/DConf2NixTest.hs

@@ -140,6 +140,17 @@ prop_dconf2nix_tuples :: Property
 prop_dconf2nix_tuples =
   withTests (10 :: TestLimit) dconf2nixTuples
 
+dconf2nixUnicode :: Property
+dconf2nixUnicode =
+  let input  = "data/unicode.settings"
+      output = "output/unicode.nix"
+      root   = Root T.empty
+  in  baseProperty input output root
+
+prop_dconf2nix_unicode :: Property
+prop_dconf2nix_unicode =
+  withTests (10 :: TestLimit) dconf2nixUnicode
+
 dconf2nixEmoji :: Property
 dconf2nixEmoji =
   let input  = "data/emoji.settings"