Merge abarkov@bk-internal.mysql.com:/home/bk/mysql-4.1

into  mysql.com:/usr/home/bar/mysql-4.1.b15376


include/m_ctype.h:
  Auto merged
strings/ctype-bin.c:
  Auto merged
strings/ctype-euc_kr.c:
  Auto merged
strings/ctype-gb2312.c:
  Auto merged
strings/ctype-ucs2.c:
  Auto merged

6 files changed, 47 insertions, 0 deletions

diff --git a/mysql-test/r/ctype_big5.result b/mysql-test/r/ctype_big5.result
index 4c5832a57e9..6574908101c 100644
--- a/mysql-test/r/ctype_big5.result
+++ b/mysql-test/r/ctype_big5.result
@@ -189,3 +189,6 @@ select hex(a) from t1 where a = _big5 0xF9DC;
 hex(a)
 E5ABBA
 drop table t1;
+select hex(convert(_big5 0xC84041 using ucs2));
+hex(convert(_big5 0xC84041 using ucs2))
+003F0041
diff --git a/mysql-test/r/ctype_gbk.result b/mysql-test/r/ctype_gbk.result
index aaffe692126..241539ecf42 100644
--- a/mysql-test/r/ctype_gbk.result
+++ b/mysql-test/r/ctype_gbk.result
@@ -165,3 +165,6 @@ hex(a)
 A1A1
 A3A0
 DROP TABLE t1;
+select hex(convert(_gbk 0xA14041 using ucs2));
+hex(convert(_gbk 0xA14041 using ucs2))
+003F0041
diff --git a/mysql-test/r/ctype_ujis.result b/mysql-test/r/ctype_ujis.result
index c86cbe9ce53..77258ba1730 100644
--- a/mysql-test/r/ctype_ujis.result
+++ b/mysql-test/r/ctype_ujis.result
@@ -2307,3 +2307,9 @@ select c1 as c2h from t1 where c1 like 'ab#_def' escape '#';
 c2h
 ab_def
 drop table t1;
+select hex(convert(_ujis 0xA5FE41 using ucs2));
+hex(convert(_ujis 0xA5FE41 using ucs2))
+003F0041
+select hex(convert(_ujis 0x8FABF841 using ucs2));
+hex(convert(_ujis 0x8FABF841 using ucs2))
+003F0041
diff --git a/mysql-test/t/ctype_big5.test b/mysql-test/t/ctype_big5.test
index ffe2a12234e..200002cd235 100644
--- a/mysql-test/t/ctype_big5.test
+++ b/mysql-test/t/ctype_big5.test
@@ -53,4 +53,14 @@ alter table t1 convert to character set utf8;
 select hex(a) from t1 where a = _big5 0xF9DC;
 drop table t1;
 
+#
+# Bugs#15375: Unassigned multibyte codes are broken
+# into parts when converting to Unicode.
+# This query should return 0x003F0041. I.e. it should
+# scan unassigned double-byte character 0xC840, convert
+# it as QUESTION MARK 0x003F and then scan the next
+# character, which is a single byte character 0x41.
+#
+select hex(convert(_big5 0xC84041 using ucs2));
+
 # End of 4.1 tests
diff --git a/mysql-test/t/ctype_gbk.test b/mysql-test/t/ctype_gbk.test
index 5eeade96186..7aec48586d8 100644
--- a/mysql-test/t/ctype_gbk.test
+++ b/mysql-test/t/ctype_gbk.test
@@ -31,4 +31,14 @@ INSERT INTO t1 VALUES (0xA3A0),(0xA1A1);
 SELECT hex(a) FROM t1 ORDER BY a;
 DROP TABLE t1;
 
+#
+# Bugs#15375: Unassigned multibyte codes are broken
+# into parts when converting to Unicode.
+# This query should return 0x003F0041. I.e. it should
+# scan unassigned double-byte character 0xA140, convert
+# it as QUESTION MARK 0x003F and then scan the next
+# character, which is a single byte character 0x41.
+#
+select hex(convert(_gbk 0xA14041 using ucs2));
+
 # End of 4.1 tests
diff --git a/mysql-test/t/ctype_ujis.test b/mysql-test/t/ctype_ujis.test
index bf74371510b..12d05f44a94 100644
--- a/mysql-test/t/ctype_ujis.test
+++ b/mysql-test/t/ctype_ujis.test
@@ -1152,4 +1152,19 @@ SET collation_connection='ujis_bin';
 -- source include/ctype_innodb_like.inc
 -- source include/ctype_like_escape.inc
 
+#
+# Bugs#15375: Unassigned multibyte codes are broken
+# into parts when converting to Unicode.
+# This query should return 0x003F0041. I.e. it should
+# scan unassigned double-byte character 0xA5FE, convert
+# it as QUESTION MARK 0x003F and then scan the next
+# character, which is a single byte character 0x41.
+#
+select hex(convert(_ujis 0xA5FE41 using ucs2));
+# This one should return 0x003F0041:
+# scan unassigned three-byte character 0x8FABF8,
+# convert it as QUESTION MARK 0x003F and then scan
+# the next character, which is a single byte character 0x41.
+select hex(convert(_ujis 0x8FABF841 using ucs2));
+
 # End of 4.1 tests