Split CJK full names into name parts correctly.

components/autofill/core/browser/autofill_data_util_unittest.cc

Index: components/autofill/core/browser/autofill_data_util_unittest.cc
diff --git a/components/autofill/core/browser/autofill_data_util_unittest.cc b/components/autofill/core/browser/autofill_data_util_unittest.cc
index 707ac4925f1c69b15012ab4c4aa6e54398fb8212..f472a00c883ded8b049f2da2a22bf7ea764ba04e 100644
--- a/components/autofill/core/browser/autofill_data_util_unittest.cc
+++ b/components/autofill/core/browser/autofill_data_util_unittest.cc
@@ -32,7 +32,33 @@ TEST(AutofillDataUtilTest, SplitName) {
       // Exception to the name suffix removal.
       {"John Ma", "John", "", "Ma"},
       // Common family name prefixes not considered a middle name.
-      {"Milhouse Van Houten", "Milhouse", "", "Van Houten"}};
+      {"Milhouse Van Houten", "Milhouse", "", "Van Houten"},
+
+      // CJK names have reverse order (surname goes first, given name goes
+      // second).
+      {"홍 길동", "길동", "", "홍"}, // Korean name, Hangul
+      {"孫 德明", "德明", "", "孫"}, // Chinese name, Unihan
+      {"山田 貴洋", "貴洋", "", "山田"}, // Japanese name, Unihan
+
+      // CJK names don't usually have a space in the middle, but most of the
+      // time, the surname is only one character (in Chinese & Korean).
+      {"최성훈", "성훈", "", "최"}, // Korean name, Hangul
+      {"강전희", "전희", "", "강"}, // Korean name, Hangul
+      {"刘翔", "翔", "", "刘"}, // (Traditional) Chinese name, Unihan
+      {"劉翔", "翔", "", "劉"}, // (Simplified) Chinese name, Unihan

[gogerald1, 2016/07/14 18:28:52]

nit: I believe this ("劉") is traditional and above ("刘") is simplified,

[nicolaso, 2016/07/14 20:32:17]

Fixed.

+
+      // There are a few exceptions. Occasionally, the surname has two
+      // characters.
+      {"남궁도", "도", "", "남궁"}, // Korean name, Hangul
+      {"황보혜정", "혜정", "", "황보"}, // Korean name, Hangul
+      {"歐陽靖", "靖", "", "歐陽"}, // (Traditional) Chinese name, Unihan
+
+      // Korean full names always have at least 3 characters, so a 2-character

[Jinsuk Kim, 2016/07/13 21:51:09]

Please update the comment here.

[nicolaso, 2016/07/14 18:12:43]

Done.

+      // name can only be a given name. Chinese full names, however, might be
+      // only 2 characters.
+      {"제동", "제동", "", ""}, // Korean name, Hangul
+      {"孫文", "文", "", "孫"} // Chinese name, Unihan
+  };
 
   for (TestCase test_case : test_cases) {
     NameParts name_parts = SplitName(base::UTF8ToUTF16(test_case.full_name));