Fixed that multibyte charsets didn't honor multibyte

sequence boundaries in functions LIKE and LOCATE in
the case of "binary" collation. Comparison was done
like if the strings were just a binary strings without
character set assumption.

3 files changed, 85 insertions, 1 deletions

diff --git a/mysql-test/r/ctype_ujis.result b/mysql-test/r/ctype_ujis.result
index ffb305a81cf..1730b17eaed 100644
--- a/mysql-test/r/ctype_ujis.result
+++ b/mysql-test/r/ctype_ujis.result
@@ -1,4 +1,5 @@
 drop table if exists t1;
+set names ujis;
 create table t1 (c text character set ujis);
 insert into t1 values (0xa4a2),(0xa4a3);
 select hex(left(c,1)) from t1 group by c;
@@ -6,3 +7,60 @@ hex(left(c,1))
 A4A2
 A4A3
 drop table t1;
+select locate(0xa2a1,0xa1a2a1a3);
+locate(0xa2a1,0xa1a2a1a3)
+2
+select locate(_ujis 0xa2a1,_ujis 0xa1a2a1a3);
+locate(_ujis 0xa2a1,_ujis 0xa1a2a1a3)
+0
+select locate(_ujis 0xa2a1,_ujis 0xa1a2a1a3 collate ujis_bin);
+locate(_ujis 0xa2a1,_ujis 0xa1a2a1a3 collate ujis_bin)
+0
+select locate('he','hello');
+locate('he','hello')
+1
+select locate('he','hello',2);
+locate('he','hello',2)
+0
+select locate('lo','hello',2);
+locate('lo','hello',2)
+4
+select locate('HE','hello');
+locate('HE','hello')
+1
+select locate('HE','hello',2);
+locate('HE','hello',2)
+0
+select locate('LO','hello',2);
+locate('LO','hello',2)
+4
+select locate('HE','hello' collate ujis_bin);
+locate('HE','hello' collate ujis_bin)
+0
+select locate('HE','hello' collate ujis_bin,2);
+locate('HE','hello' collate ujis_bin,2)
+0
+select locate('LO','hello' collate ujis_bin,2);
+locate('LO','hello' collate ujis_bin,2)
+0
+select locate(_ujis 0xa1a3,_ujis 0xa1a2a1a3);
+locate(_ujis 0xa1a3,_ujis 0xa1a2a1a3)
+2
+select 0xa1a2a1a3 like concat(_binary'%',0xa2a1,_binary'%');
+0xa1a2a1a3 like concat(_binary'%',0xa2a1,_binary'%')
+1
+select _ujis 0xa1a2a1a3 like concat(_ujis'%',_ujis 0xa2a1, _ujis'%');
+_ujis 0xa1a2a1a3 like concat(_ujis'%',_ujis 0xa2a1, _ujis'%')
+0
+select _ujis 0xa1a2a1a3 like concat(_ujis'%',_ujis 0xa2a1, _ujis'%') collate ujis_bin;
+_ujis 0xa1a2a1a3 like concat(_ujis'%',_ujis 0xa2a1, _ujis'%') collate ujis_bin
+0
+select 'a' like 'a';
+'a' like 'a'
+1
+select 'A' like 'a';
+'A' like 'a'
+1
+select 'A' like 'a' collate ujis_bin;
+'A' like 'a' collate ujis_bin
+0
diff --git a/mysql-test/r/have_ujis.require b/mysql-test/r/have_ujis.require
index b4de2234ec7..43a309ad74e 100644
--- a/mysql-test/r/have_ujis.require
+++ b/mysql-test/r/have_ujis.require
@@ -1,2 +1,2 @@
 Collation	Charset	Id	Default	Compiled	Sortlen
-ujis_japanese_ci	ujis	12	Yes	Yes	0
+ujis_japanese_ci	ujis	12	Yes	Yes	1
diff --git a/mysql-test/t/ctype_ujis.test b/mysql-test/t/ctype_ujis.test
index e41caf55948..bcf6507b4c7 100644
--- a/mysql-test/t/ctype_ujis.test
+++ b/mysql-test/t/ctype_ujis.test
@@ -7,6 +7,8 @@
 drop table if exists t1;
 --enable_warnings
 
+set names ujis;
+
 #
 # Test problem with LEFT()
 #
@@ -15,3 +17,27 @@ create table t1 (c text character set ujis);
 insert into t1 values (0xa4a2),(0xa4a3);
 select hex(left(c,1)) from t1 group by c;
 drop table t1;
+
+#
+#
+#
+select locate(0xa2a1,0xa1a2a1a3);
+select locate(_ujis 0xa2a1,_ujis 0xa1a2a1a3);
+select locate(_ujis 0xa2a1,_ujis 0xa1a2a1a3 collate ujis_bin);
+select locate('he','hello');
+select locate('he','hello',2);
+select locate('lo','hello',2);
+select locate('HE','hello');
+select locate('HE','hello',2);
+select locate('LO','hello',2);
+select locate('HE','hello' collate ujis_bin);
+select locate('HE','hello' collate ujis_bin,2);
+select locate('LO','hello' collate ujis_bin,2);
+select locate(_ujis 0xa1a3,_ujis 0xa1a2a1a3);
+
+select 0xa1a2a1a3 like concat(_binary'%',0xa2a1,_binary'%');
+select _ujis 0xa1a2a1a3 like concat(_ujis'%',_ujis 0xa2a1, _ujis'%');
+select _ujis 0xa1a2a1a3 like concat(_ujis'%',_ujis 0xa2a1, _ujis'%') collate ujis_bin;
+select 'a' like 'a';
+select 'A' like 'a';
+select 'A' like 'a' collate ujis_bin;