Merge pull request #41 from inariksit/master

Arabic orthography

src/arabic/LexiconAra.gf

@@ -73,7 +73,7 @@ flags
   drink_V2     = dirV2 (regV "شَرِب") ;
 --  drink_V2     = dirV2 (v1 "شرب" i a) ;
   easy_A2V     = mkA2 (sndA "سهل" "فَعل") "لِ" ;
-  eat_V2       = dirV2 (mkV "ء‎كل" FormI) ;
+  eat_V2       = dirV2 (mkV "ءكل" FormI) ;
   empty_A      = sndA "فرغ" "فاعِل" ;
   enemy_N      = brkN "عدو" "فَعُلّ" "أَفعاء" Masc Hum ;
   factory_N    = brkN "صنع" "مَفعَل" "مَفاعِل" Masc NoHum ;
@@ -111,7 +111,7 @@ flags
   iron_N       = brkN "حدّ" "فَعِيل" "فَعائِل" Masc NoHum ;
   king_N       = brkN "ملك" "فَعِل" "فُعُول" Masc Hum ;
   know_V2      = dirV2 (regV "عَرِف") ;
-  know_VS      = mkVS (regV "عَرِف") ; -- or with ع ل م‎?
+  know_VS      = mkVS (regV "عَرِف") ; -- or with ع ل م?
   lake_N       = sdfN "بحر" "فُعَيلة" Fem NoHum ;
   lamp_N       = brkN "صبح" "مِفعال" "مَفاعِيل" Masc NoHum ; --qanDIl, fAnUs
   learn_V2     = dirV2 (v5 "علم") ;

src/arabic/OrthoAra.gf

@@ -11,8 +11,15 @@ flags coding=utf8 ;
     -- "Sun letters": assimilate with def. article
     sun : pattern Str = #("ت"|"ث"|"د"|"ذ"|"ر"|"ز"|"س"|"ش"|"ص"|"ض"|"ط"|"ظ"|"ل"|"ن") ;
 
+  -- Shadda: https://www.unicode.org/L2/L2017/17253-arabic-ordering.pdf
+    fixShd : Str -> Str -> Str = \word,vowel ->
+      case word of {
+        x + "ّ" => x + vowel + "ّ" ;
+        x       => x + vowel
+      } ;
+
   -- Hamza
-    hamza : pattern Str = #("ء‎"|"؟") ;
+    hamza : pattern Str = #("ء"|"؟") ;
 
     rectifyHmz: Str -> Str = \word ->
       case word of {

src/arabic/ParadigmsAra.gf

@@ -438,7 +438,7 @@ resource ParadigmsAra = open
 
   -- e.g. al-jamii3, 2a7ad
   regNP : Str -> Number -> NP = \word,n ->
-    { s = \\c => word ++ vowel ! c ; ---- gives strange chars
+    { s = \\c => fixShd word (dec1sg ! Def ! c) ;
       a = {pgn = Per3 Masc n; isPron = False };
       lock_NP = <>
     };

src/arabic/ResAra.gf

@@ -842,13 +842,12 @@ patHollowImp : (_,_ :Str) -> Gender => Number => Str =\xaf,xAf ->
 
     -- takes a singular or broken plural word and tests the ending to
     -- determine the declension and gives the corresponding inf table
-    sing : Str -> State => Case => Str =
+    sing : Str -> State => Case => Str = \word ->
-      \word ->
+      \\s,c => defArt s (case word of {
-      case word of {
+         lemma + "ِي"  => fixShd lemma (dec2sg ! s ! c) ;
-         lemma + "ِي"   => \\s,c => defArt s lemma + dec2sg ! s ! c ;
+         _ + ("ا"|"ى") => fixShd word  (dec3sg ! s ! c) ;
-          _ + ("ا"|"ى") => \\s,c => defArt s word + dec3sg ! s ! c ;
+         _             => fixShd word  (dec1sg ! s ! c) 
-          _             => \\s,c => defArt s word + dec1sg ! s ! c
+      }) ;
-      };
 
 
     -- takes a singular word and tests the ending to
@@ -952,7 +951,7 @@ patHollowImp : (_,_ :Str) -> Gender => Number => Str =\xaf,xAf ->
         Const =>
           table {
             Nom => "َا";
-            _   => "َيْ‎"
+            _   => "َيْ"
           };
         _ =>
           table {

src/arabic/VerbAra.gf

@@ -20,7 +20,7 @@ concrete VerbAra of Verb = CatAra ** open Prelude, ResAra in {
     ComplVV vv vp =  let vvVP = predV vv in --- IL
       vp ** {
         s = \\pgn,vpf => vvVP.s ! pgn ! vpf
-                      ++ vv.c2  -- أَنْ‎
+                      ++ vv.c2  -- أَنْ
                       ++ vp.s ! pgn ! VPImpf Cnj
       } ;