Added transliteration test.

2 files changed, 39 insertions, 18 deletions

diff --git a/coding/multilang_utf8_string.cpp b/coding/multilang_utf8_string.cpp
index a3e1723121..107655bca9 100644
--- a/coding/multilang_utf8_string.cpp
+++ b/coding/multilang_utf8_string.cpp
@@ -9,23 +9,38 @@ namespace
 // Languages below were choosen after sorting name:<lang> tags in 2011.
 // Note, that it's not feasible to increase languages number here due to
 // our current encoding (6 bit to store language code).
-StringUtf8Multilang::Languages const g_languages = {{ {"default", "Native for each country", "Any-Latin"},
-    {"en", "English", ""}, {"ja", "日本語", "Any-Latin"}, {"fr", "Français", ""}, {"ko_rm", "Korean (Romanized)", "Korean-Latin/BGN"},
-    {"ar", "العربية", "Any-Latin"}, {"de", "Deutsch", ""}, {"int_name", "International (Latin)", "Any-Latin"}, {"ru", "Русский", "Russian-Latin/BGN"},
-    {"sv", "Svenska", "Any-Latin"}, {"zh", "中文", "Any-Latin"}, {"fi", "Suomi", "Any-Latin"}, {"be", "Беларуская", "Belarusian-Latin/BGN"}, {"ka", "ქართული", "Georgian-Latin"},
-    {"ko", "한국어", "Hangul"}, {"he", "עברית", "Hebrew"}, {"nl", "Nederlands", ""}, {"ga", "Gaeilge", "Any-Latin"},
-    {"ja_rm", "Japanese (Romanized)", "Any-Latin"}, {"el", "Ελληνικά", "Greek-Latin"}, {"it", "Italiano", ""}, {"es", "Español", ""},
-    {"zh_pinyin", "Chinese (Pinyin)", "Any-Latin"}, {"th", "ไทย", "Thai-Latin"}, {"cy", "Cymraeg", "Any-Latin"}, {"sr", "Српски", "Serbian-Latin/BGN"},
-    {"uk", "Українська", "Ukrainian-Latin/BGN"}, {"ca", "Català", "Any-Latin"}, {"hu", "Magyar", "Any-Latin"}, {"hsb", "Hornjoserbšćina", "Any-Latin"}, {"eu", "Euskara", "Any-Latin"},
-    {"fa", "فارسی", "Any-Latin"}, {"br", "Breton", "Any-Latin"}, {"pl", "Polski", "Any-Latin"}, {"hy", "Հայերէն", "Armenian-Latin"}, {"kn", "ಕನ್ನಡ", "Kannada-Latin"},
-    {"sl", "Slovenščina", "Any-Latin"}, {"ro", "Română", "Any-Latin"}, {"sq", "Shqipe", "Any-Latin"}, {"am", "አማርኛ", "Amharic-Latin/BGN"}, {"fy", "Frysk", "Any-Latin"},
-    {"cs", "Čeština", "Any-Latin"}, {"gd", "Gàidhlig", "Any-Latin"}, {"sk", "Slovenčina", "Any-Latin"}, {"af", "Afrikaans", "Any-Latin"},
-    {"ja_kana", "日本語(カタカナ)", "Katakana-Latin"}, {"lb", "Luxembourgish", "Any-Latin"}, {"pt", "Português", "Any-Latin"}, {"hr", "Hrvatski", "Any-Latin"},
-    {"fur", "Friulian", "Any-Latin"}, {"vi", "Tiếng Việt", "Any-Latin"}, {"tr", "Türkçe", "Any-Latin"}, {"bg", "Български", "Bulgarian-Latin/BGN"},
-    {"eo", "Esperanto", "Any-Latin"}, {"lt", "Lietuvių", "Any-Latin"}, {"la", "Latin", ""}, {"kk", "Қазақ", "Kazakh-Latin/BGN"},
-    {"gsw", "Schwiizertüütsch", "Any-Latin"}, {"et", "Eesti", "Any-Latin"}, {"ku", "Kurdish", "Any-Latin"}, {"mn", "Mongolian", "Mongolian-Latin/BGN"},
-    {"mk", "Македонски", "Macedonian-Latin/BGN"}, {"lv", "Latviešu", "Any-Latin"}, {"hi", "हिन्दी", "Any-Latin"}
-}};
+StringUtf8Multilang::Languages const g_languages = {{
+    {"default", "Native for each country", "Any-Latin"},
+    {"en", "English", ""}, {"ja", "日本語", "Any-Latin"}, {"fr", "Français", ""},
+    {"ko_rm", "Korean (Romanized)", "Korean-Latin/BGN"}, {"ar", "العربية", "Any-Latin"},
+    {"de", "Deutsch", ""}, {"int_name", "International (Latin)", "Any-Latin"},
+    {"ru", "Русский", "Russian-Latin/BGN"}, {"sv", "Svenska", "Any-Latin"},
+    {"zh", "中文", "Any-Latin"}, {"fi", "Suomi", "Any-Latin"},
+    {"be", "Беларуская", "Belarusian-Latin/BGN"}, {"ka", "ქართული", "Georgian-Latin"},
+    {"ko", "한국어", "Hangul-Latin/BGN"}, {"he", "עברית", "Hebrew-Latin/BGN"}, {"nl", "Nederlands", ""},
+    {"ga", "Gaeilge", "Any-Latin"}, {"ja_rm", "Japanese (Romanized)", "Any-Latin"},
+    {"el", "Ελληνικά", "Greek-Latin"}, {"it", "Italiano", ""}, {"es", "Español", ""},
+    {"zh_pinyin", "Chinese (Pinyin)", "Any-Latin"}, {"th", "ไทย", "Thai-Latin"},
+    {"cy", "Cymraeg", "Any-Latin"}, {"sr", "Српски", "Serbian-Latin/BGN"},
+    {"uk", "Українська", "Ukrainian-Latin/BGN"}, {"ca", "Català", "Any-Latin"},
+    {"hu", "Magyar", "Any-Latin"}, {"hsb", "Hornjoserbšćina", "Any-Latin"},
+    {"eu", "Euskara", "Any-Latin"}, {"fa", "فارسی", "Any-Latin"}, {"br", "Breton", "Any-Latin"},
+    {"pl", "Polski", "Any-Latin"}, {"hy", "Հայերէն", "Armenian-Latin"},
+    {"kn", "ಕನ್ನಡ", "Kannada-Latin"}, {"sl", "Slovenščina", "Any-Latin"},
+    {"ro", "Română", "Any-Latin"}, {"sq", "Shqipe", "Any-Latin"},
+    {"am", "አማርኛ", "Amharic-Latin/BGN"}, {"fy", "Frysk", "Any-Latin"},
+    {"cs", "Čeština", "Any-Latin"}, {"gd", "Gàidhlig", "Any-Latin"},
+    {"sk", "Slovenčina", "Any-Latin"}, {"af", "Afrikaans", "Any-Latin"},
+    {"ja_kana", "日本語(カタカナ)", "Katakana-Latin"}, {"lb", "Luxembourgish", "Any-Latin"},
+    {"pt", "Português", "Any-Latin"}, {"hr", "Hrvatski", "Any-Latin"},
+    {"fur", "Friulian", "Any-Latin"}, {"vi", "Tiếng Việt", "Any-Latin"},
+    {"tr", "Türkçe", "Any-Latin"}, {"bg", "Български", "Bulgarian-Latin/BGN"},
+    {"eo", "Esperanto", "Any-Latin"}, {"lt", "Lietuvių", "Any-Latin"}, {"la", "Latin", ""},
+    {"kk", "Қазақ", "Kazakh-Latin/BGN"}, {"gsw", "Schwiizertüütsch", "Any-Latin"},
+    {"et", "Eesti", "Any-Latin"}, {"ku", "Kurdish", "Any-Latin"},
+    {"mn", "Mongolian", "Mongolian-Latin/BGN"}, {"mk", "Македонски", "Macedonian-Latin/BGN"},
+    {"lv", "Latviešu", "Any-Latin"}, {"hi", "हिन्दी", "Any-Latin"}
+  }};
 
 static_assert(g_languages.size() == StringUtf8Multilang::kMaxSupportedLanguages,
               "With current encoding we are limited to 64 languages max.");
diff --git a/coding/transliteration.cpp b/coding/transliteration.cpp
index 937dd8106f..c1a9ab1138 100644
--- a/coding/transliteration.cpp
+++ b/coding/transliteration.cpp
@@ -3,6 +3,7 @@
 
 #include "base/logging.hpp"
 
+#include "3party/icu/common/unicode/uclean.h"
 #include "3party/icu/common/unicode/unistr.h"
 #include "3party/icu/common/unicode/utypes.h"
 #include "3party/icu/i18n/unicode/translit.h"
@@ -10,7 +11,12 @@
 
 Transliteration::~Transliteration()
 {
-  //u_cleanup();
+  // The use of u_cleanup() just before an application terminates is optional,
+  // but it should be called only once for performance reasons.
+  // The primary benefit is to eliminate reports of memory or resource leaks originating
+  // in ICU code from the results generated by heap analysis tools.
+  m_transliterators.clear();
+  u_cleanup();
 }
 
 Transliteration & Transliteration::GetInstance()