cleaned up ISO code usage for Kurdish, fdo#63460

* instead of the 'ku' macrolanguage code use proper ISO 639-3 codes and
  use 'Latn' script with 'kmr'
* use MS-LCID 0x0492 for Central Kurdish (Iraq) [ckb-IQ]
* added Southern Kurdish (Iraq) [sdh-IQ]

Change-Id: Iaee8be98d0659a0e7bbf041e60025dd1f771066f

4 files changed, 85 insertions, 7 deletions

diff --git a/i18nlangtag/qa/cppunit/test_languagetag.cxx b/i18nlangtag/qa/cppunit/test_languagetag.cxx
index 17217fd08497..8946969b2562 100644
--- a/i18nlangtag/qa/cppunit/test_languagetag.cxx
+++ b/i18nlangtag/qa/cppunit/test_languagetag.cxx
@@ -625,6 +625,14 @@ static bool checkMapping( const OUString rStr1, const OUString& rStr2 )
     if (rStr1 == "yi-Hebr-IL"  ) return rStr2 == "yi-IL";
     if (rStr1 == "ha-NG"       ) return rStr2 == "ha-Latn-NG";
     if (rStr1 == "ha-GH"       ) return rStr2 == "ha-Latn-GH";
+    if (rStr1 == "ku-Arab-IQ"  ) return rStr2 == "ckb-IQ";
+    if (rStr1 == "ku-Arab"     ) return rStr2 == "ckb";
+    if (rStr1 == "kmr-TR"      ) return rStr2 == "kmr-Latn-TR";
+    if (rStr1 == "ku-TR"       ) return rStr2 == "kmr-Latn-TR";
+    if (rStr1 == "kmr-SY"      ) return rStr2 == "kmr-Latn-SY";
+    if (rStr1 == "ku-SY"       ) return rStr2 == "kmr-Latn-SY";
+    if (rStr1 == "ku-IQ"       ) return rStr2 == "ckb-IQ";
+    if (rStr1 == "ku-IR"       ) return rStr2 == "ckb-IR";
     return rStr1 == rStr2;
 }
 
diff --git a/i18nlangtag/source/isolang/isolang.cxx b/i18nlangtag/source/isolang/isolang.cxx
index f9f34a54949e..5505540730ab 100644
--- a/i18nlangtag/source/isolang/isolang.cxx
+++ b/i18nlangtag/source/isolang/isolang.cxx
@@ -506,10 +506,17 @@ static IsoLanguageCountryEntry const aImplIsoLangEntries[] =
     { LANGUAGE_OBSOLETE_USER_LOWER_SORBIAN,"dsb", "DE", 0     },
     { LANGUAGE_OCCITAN_FRANCE,              "oc", "FR", 0     },
     { LANGUAGE_OBSOLETE_USER_OCCITAN,       "oc", "FR", 0     },
-    { LANGUAGE_USER_KURDISH_TURKEY,         "ku", "TR", 0     },
-    { LANGUAGE_USER_KURDISH_SYRIA,          "ku", "SY", 0     },
-    { LANGUAGE_USER_KURDISH_IRAQ,           "ku", "IQ", 0     },
-    { LANGUAGE_USER_KURDISH_IRAN,           "ku", "IR", 0     },
+    { LANGUAGE_USER_KURDISH_TURKEY,        "kmr", "TR", kSAME },
+    { LANGUAGE_USER_KURDISH_TURKEY,         "ku", "TR", kSAME },
+    { LANGUAGE_USER_KURDISH_SYRIA,         "kmr", "SY", kSAME },
+    { LANGUAGE_USER_KURDISH_SYRIA,          "ku", "SY", kSAME },
+    { LANGUAGE_KURDISH_ARABIC_IRAQ,        "ckb", "IQ", 0     },
+    { LANGUAGE_KURDISH_ARABIC_IRAQ,         "ku", "IQ", kSAME },
+    { LANGUAGE_OBSOLETE_USER_KURDISH_IRAQ,  "ku", "IQ", LANGUAGE_KURDISH_ARABIC_IRAQ },
+    { LANGUAGE_USER_KURDISH_SOUTHERN_IRAQ, "sdh", "IQ", 0     },
+    { LANGUAGE_USER_KURDISH_IRAN,          "ckb", "IR", 0     },
+    { LANGUAGE_USER_KURDISH_IRAN,           "ku", "IR", kSAME },
+    { LANGUAGE_KURDISH_ARABIC_LSO,         "ckb", ""  , 0     },
     { LANGUAGE_USER_SARDINIAN,              "sc", "IT", 0     },    // macrolanguage code
     { LANGUAGE_USER_SARDINIAN_CAMPIDANESE, "sro", "IT", 0     },
     { LANGUAGE_USER_SARDINIAN_GALLURESE,   "sdn", "IT", 0     },
@@ -697,8 +704,10 @@ static IsoLanguageScriptCountryEntry const aImplIsoLangScriptEntries[] =
     { LANGUAGE_LATIN_LSO,                           "la-Latn", ""  , kSAME },   // MS, though Latn is suppress-script
     { LANGUAGE_TAI_NUA_CHINA,                      "tdd-Tale", "CN", 0     },   // MS reserved
     { LANGUAGE_LU_CHINA,                           "khb-Talu", "CN", 0     },   // MS reserved
-    { LANGUAGE_KURDISH_ARABIC_IRAQ,                 "ku-Arab", "IQ", 0     },   // macrolanguage code, MS
-    { LANGUAGE_KURDISH_ARABIC_LSO,                  "ku-Arab", ""  , 0     },   // macrolanguage code
+    { LANGUAGE_KURDISH_ARABIC_IRAQ,                 "ku-Arab", "IQ", kSAME },   // macrolanguage code, MS
+    { LANGUAGE_KURDISH_ARABIC_LSO,                  "ku-Arab", ""  , kSAME },   // macrolanguage code, MS
+    { LANGUAGE_USER_KURDISH_TURKEY,                "kmr-Latn", "TR", 0     },
+    { LANGUAGE_USER_KURDISH_SYRIA,                 "kmr-Latn", "SY", 0     },
     { LANGUAGE_PUNJABI_PAKISTAN,                   "pnb-Arab", "PK", 0     },
     { LANGUAGE_PUNJABI_ARABIC_LSO,                 "pnb-Arab", ""  , 0     },
     { LANGUAGE_PUNJABI_PAKISTAN,                    "pa-Arab", "PK", 0     },   // MS, incorrect
diff --git a/i18nlangtag/source/isolang/mslangid.cxx b/i18nlangtag/source/isolang/mslangid.cxx
index 3964c6fd2ed2..d44256de4117 100644
--- a/i18nlangtag/source/isolang/mslangid.cxx
+++ b/i18nlangtag/source/isolang/mslangid.cxx
@@ -216,6 +216,18 @@ bool MsLangId::isRightToLeft( LanguageType nLang )
         default:
             break;
     }
+    switch (nLang)
+    {
+        case LANGUAGE_USER_KURDISH_IRAN:
+        case LANGUAGE_OBSOLETE_USER_KURDISH_IRAQ:
+        case LANGUAGE_KURDISH_ARABIC_IRAQ:
+        case LANGUAGE_KURDISH_ARABIC_LSO:
+        case LANGUAGE_USER_KURDISH_SOUTHERN_IRAQ:
+            return true;
+
+        default:
+            break;
+    }
     return false;
 }
 
@@ -323,7 +335,10 @@ sal_Int16 MsLangId::getScriptType( LanguageType nLang )
         case LANGUAGE_MONGOLIAN_MONGOLIAN_CHINA:
         case LANGUAGE_MONGOLIAN_MONGOLIAN_LSO:
         case LANGUAGE_USER_KURDISH_IRAN:
-        case LANGUAGE_USER_KURDISH_IRAQ:
+        case LANGUAGE_OBSOLETE_USER_KURDISH_IRAQ:
+        case LANGUAGE_KURDISH_ARABIC_IRAQ:
+        case LANGUAGE_KURDISH_ARABIC_LSO:
+        case LANGUAGE_USER_KURDISH_SOUTHERN_IRAQ:
         case LANGUAGE_USER_KYRGYZ_CHINA:
             nScript = ::com::sun::star::i18n::ScriptType::COMPLEX;
             break;
@@ -507,6 +522,9 @@ LanguageType MsLangId::getReplacementForObsoleteLanguage( LanguageType nLang, bo
         case LANGUAGE_OBSOLETE_USER_SERBIAN_CYRILLIC_MONTENEGRO:
             nLang = LANGUAGE_SERBIAN_CYRILLIC_MONTENEGRO;
             break;
+        case LANGUAGE_OBSOLETE_USER_KURDISH_IRAQ:
+            nLang = LANGUAGE_KURDISH_ARABIC_IRAQ;
+            break;
 
         // The following are not strictly obsolete but should be mapped to a
         // replacement locale when encountered.
diff --git a/i18nlangtag/source/languagetag/languagetag.cxx b/i18nlangtag/source/languagetag/languagetag.cxx
index b0a8f4d58517..60e8c1280f55 100644
--- a/i18nlangtag/source/languagetag/languagetag.cxx
+++ b/i18nlangtag/source/languagetag/languagetag.cxx
@@ -2073,6 +2073,11 @@ LanguageTag & LanguageTag::makeFallback()
 }
 
 
+/* TODO: maybe this now could take advantage of the mnOverride field in
+ * isolang.cxx entries and search for kSAME instead of harcoded special
+ * fallbacks. Though iterating through those tables would be slower and even
+ * then there would be some special cases, but we wouldn't lack entries that
+ * were missed out. */
 ::std::vector< OUString > LanguageTag::getFallbackStrings( bool bIncludeFullBcp47 ) const
 {
     ::std::vector< OUString > aVec;
@@ -2111,6 +2116,36 @@ LanguageTag & LanguageTag::makeFallback()
                 aVec.insert( aVec.end(), aRep.begin(), aRep.end());
                 // Already includes 'ca' language fallback.
             }
+            else if (aLanguage == "ku")
+            {
+                if (aCountry == "TR" || aCountry == "SY")
+                {
+                    aVec.push_back( "kmr-Latn-" + aCountry);
+                    aVec.push_back( "kmr-" + aCountry);
+                    aVec.push_back( "kmr-Latn");
+                    aVec.push_back( "kmr");
+                    aVec.push_back( aLanguage);
+                }
+                else if (aCountry == "IQ" || aCountry == "IR")
+                {
+                    aVec.push_back( "ckb-" + aCountry);
+                    aVec.push_back( "ckb");
+                }
+            }
+            else if (aLanguage == "kmr" && (aCountry == "TR" || aCountry == "SY"))
+            {
+                aVec.push_back( "ku-Latn-" + aCountry);
+                aVec.push_back( "ku-" + aCountry);
+                aVec.push_back( aLanguage);
+                aVec.push_back( "ku");
+            }
+            else if (aLanguage == "ckb" && (aCountry == "IQ" || aCountry == "IR"))
+            {
+                aVec.push_back( "ku-Arab-" + aCountry);
+                aVec.push_back( "ku-" + aCountry);
+                aVec.push_back( aLanguage);
+                // not 'ku' only, that was used for Latin script
+            }
             else
                 aVec.push_back( aLanguage);
         }
@@ -2174,6 +2209,8 @@ LanguageTag & LanguageTag::makeFallback()
             }
             else if (aLanguage == "pi" && aScript == "Latn")
                 aVec.push_back( "pli");     // a special case for Pali dictionary, see fdo#41599
+            else if (aLanguage == "krm" && aScript == "Latn" && (aCountry == "TR" || aCountry == "SY"))
+                aVec.push_back( "ku-" + aCountry);
         }
         if (!aVariants.isEmpty() && !bHaveLanguageScriptVariant)
         {
@@ -2184,10 +2221,16 @@ LanguageTag & LanguageTag::makeFallback()
         aTmp = aLanguage + "-" + aScript;
         if (aTmp != maBcp47)
             aVec.push_back( aTmp);
+
         // 'sh' actually denoted a script, so have it here instead of appended
         // at the end as language-only.
         if (aLanguage == "sr" && aScript == "Latn")
             aVec.push_back( "sh");
+        else if (aLanguage == "ku" && aScript == "Arab")
+            aVec.push_back( "ckb");
+        // 'ku' only denoted Latin script
+        else if (aLanguage == "krm" && aScript == "Latn" && aCountry.isEmpty())
+            aVec.push_back( "ku");
     }
     bool bHaveLanguageVariant = false;
     if (!aCountry.isEmpty())