ctype-gbk.c:

  Bug #11987
  mysql will truncate the text when the text contain GBK char:"0xA3A0" and "0xA1"
  Allow to store and retrieve even unassigned GBK codes.
  Like we did in Big5 earlier.
have_gbk.inc, have_gbk.require, ctype_gbk.result, ctype_gbk.test:
  new file


strings/ctype-gbk.c:
  Bug #11987
  mysql will truncate the text when the text contain GBK char:"0xA3A0" and "0xA1"
  Allow to store and retrieve even unassigned GBK codes.
  Like we did in Big5 earlier.

5 files changed, 205 insertions, 1 deletions

diff --git a/mysql-test/include/have_gbk.inc b/mysql-test/include/have_gbk.inc
new file mode 100644
index 00000000000..55805da0495
--- /dev/null
+++ b/mysql-test/include/have_gbk.inc
@@ -0,0 +1,4 @@
+-- require r/have_gbk.require
+disable_query_log;
+show collation like "gbk_chinese_ci";
+enable_query_log;
diff --git a/mysql-test/r/ctype_gbk.result b/mysql-test/r/ctype_gbk.result
new file mode 100644
index 00000000000..1a9dea28429
--- /dev/null
+++ b/mysql-test/r/ctype_gbk.result
@@ -0,0 +1,131 @@
+drop table if exists t1;
+SET @test_character_set= 'gbk';
+SET @test_collation= 'gbk_chinese_ci';
+SET @safe_character_set_server= @@character_set_server;
+SET @safe_collation_server= @@collation_server;
+SET character_set_server= @test_character_set;
+SET collation_server= @test_collation;
+CREATE DATABASE d1;
+USE d1;
+CREATE TABLE t1 (c CHAR(10), KEY(c));
+SHOW FULL COLUMNS FROM t1;
+Field	Type	Collation	Null	Key	Default	Extra	Privileges	Comment
+c	char(10)	gbk_chinese_ci	YES	MUL	NULL			
+INSERT INTO t1 VALUES ('aaa'),('aaaa'),('aaaaa');
+SELECT c as want3results FROM t1 WHERE c LIKE 'aaa%';
+want3results
+aaa
+aaaa
+aaaaa
+DROP TABLE t1;
+CREATE TABLE t1 (c1 varchar(15), KEY c1 (c1(2)));
+SHOW FULL COLUMNS FROM t1;
+Field	Type	Collation	Null	Key	Default	Extra	Privileges	Comment
+c1	varchar(15)	gbk_chinese_ci	YES	MUL	NULL			
+INSERT INTO t1 VALUES ('location'),('loberge'),('lotre'),('boabab');
+SELECT c1 as want3results from t1 where c1 like 'l%';
+want3results
+location
+loberge
+lotre
+SELECT c1 as want3results from t1 where c1 like 'lo%';
+want3results
+location
+loberge
+lotre
+SELECT c1 as want1result  from t1 where c1 like 'loc%';
+want1result
+location
+SELECT c1 as want1result  from t1 where c1 like 'loca%';
+want1result
+location
+SELECT c1 as want1result  from t1 where c1 like 'locat%';
+want1result
+location
+SELECT c1 as want1result  from t1 where c1 like 'locati%';
+want1result
+location
+SELECT c1 as want1result  from t1 where c1 like 'locatio%';
+want1result
+location
+SELECT c1 as want1result  from t1 where c1 like 'location%';
+want1result
+location
+DROP TABLE t1;
+DROP DATABASE d1;
+USE test;
+SET character_set_server= @safe_character_set_server;
+SET collation_server= @safe_collation_server;
+SET NAMES gbk;
+SET collation_connection='gbk_chinese_ci';
+create table t1 select repeat('a',4000) a;
+delete from t1;
+insert into t1 values ('a'), ('a '), ('a\t');
+select collation(a),hex(a) from t1 order by a;
+collation(a)	hex(a)
+gbk_chinese_ci	6109
+gbk_chinese_ci	61
+gbk_chinese_ci	6120
+drop table t1;
+create table t1 engine=innodb select repeat('a',50) as c1;
+alter table t1 add index(c1(5));
+insert into t1 values ('abcdefg'),('abcde100'),('abcde110'),('abcde111');
+select collation(c1) from t1 limit 1;
+collation(c1)
+gbk_chinese_ci
+select c1 from t1 where c1 like 'abcdef%' order by c1;
+c1
+abcdefg
+select c1 from t1 where c1 like 'abcde1%' order by c1;
+c1
+abcde100
+abcde110
+abcde111
+select c1 from t1 where c1 like 'abcde11%' order by c1;
+c1
+abcde110
+abcde111
+select c1 from t1 where c1 like 'abcde111%' order by c1;
+c1
+abcde111
+drop table t1;
+SET collation_connection='gbk_bin';
+create table t1 select repeat('a',4000) a;
+delete from t1;
+insert into t1 values ('a'), ('a '), ('a\t');
+select collation(a),hex(a) from t1 order by a;
+collation(a)	hex(a)
+gbk_bin	6109
+gbk_bin	61
+gbk_bin	6120
+drop table t1;
+create table t1 engine=innodb select repeat('a',50) as c1;
+alter table t1 add index(c1(5));
+insert into t1 values ('abcdefg'),('abcde100'),('abcde110'),('abcde111');
+select collation(c1) from t1 limit 1;
+collation(c1)
+gbk_bin
+select c1 from t1 where c1 like 'abcdef%' order by c1;
+c1
+abcdefg
+select c1 from t1 where c1 like 'abcde1%' order by c1;
+c1
+abcde100
+abcde110
+abcde111
+select c1 from t1 where c1 like 'abcde11%' order by c1;
+c1
+abcde110
+abcde111
+select c1 from t1 where c1 like 'abcde111%' order by c1;
+c1
+abcde111
+drop table t1;
+SET NAMES gbk;
+CREATE TABLE t1 (a text) character set gbk;
+INSERT INTO t1 VALUES (0xA3A0),(0xA1A1);
+SELECT hex(a) FROM t1 ORDER BY a;
+hex(a)
+A1A1
+A3A0
+DROP TABLE t1;
diff --git a/mysql-test/r/have_gbk.require b/mysql-test/r/have_gbk.require
new file mode 100644
index 00000000000..e19fead153a
--- /dev/null
+++ b/mysql-test/r/have_gbk.require
@@ -0,0 +1,2 @@
+Collation	Charset	Id	Default	Compiled	Sortlen
+gbk_chinese_ci	gbk	28	Yes	Yes	1
diff --git a/mysql-test/t/ctype_gbk.test b/mysql-test/t/ctype_gbk.test
new file mode 100644
index 00000000000..82a2e61879e
--- /dev/null
+++ b/mysql-test/t/ctype_gbk.test
@@ -0,0 +1,30 @@
+-- source include/have_gbk.inc
+
+#
+# Tests with the gbk character set
+#
+--disable_warnings
+drop table if exists t1;
+--enable_warnings
+
+SET @test_character_set= 'gbk';
+SET @test_collation= 'gbk_chinese_ci';
+-- source include/ctype_common.inc
+
+SET NAMES gbk;
+SET collation_connection='gbk_chinese_ci';
+-- source include/ctype_filesort.inc
+-- source include/ctype_innodb_like.inc
+SET collation_connection='gbk_bin';
+-- source include/ctype_filesort.inc
+-- source include/ctype_innodb_like.inc
+
+#
+# Bug#11987 mysql will truncate the text when
+# the text contain GBK char:"0xA3A0" and "0xA1"
+#
+SET NAMES gbk;
+CREATE TABLE t1 (a text) character set gbk;
+INSERT INTO t1 VALUES (0xA3A0),(0xA1A1);
+SELECT hex(a) FROM t1 ORDER BY a;
+DROP TABLE t1;
diff --git a/strings/ctype-gbk.c b/strings/ctype-gbk.c
index 89d28320fe1..6fb072d266d 100644
--- a/strings/ctype-gbk.c
+++ b/strings/ctype-gbk.c
@@ -9925,6 +9925,43 @@ my_mb_wc_gbk(CHARSET_INFO *cs __attribute__((unused)),
 }
 
 
+/*
+  Returns well formed length of a GBK string.
+*/
+static
+uint my_well_formed_len_gbk(CHARSET_INFO *cs __attribute__((unused)),
+                            const char *b, const char *e,
+                            uint pos, int *error)
+{
+  const char *b0= b;
+  const char *emb= e - 1; /* Last possible end of an MB character */
+
+  *error= 0;
+  while (pos-- && b < e)
+  {
+    if ((uchar) b[0] < 128)
+    {
+      /* Single byte ascii character */
+      b++;
+    }
+    else  if ((b < emb) && isgbkcode((uchar)*b, (uchar)b[1]))
+    {
+      /* Double byte character */
+      b+= 2;
+    }
+    else
+    {
+      /* Wrong byte sequence */
+      *error= 1;
+      break;
+    }
+  }
+  return b - b0;
+}
+
+
+                             
+
 static MY_COLLATION_HANDLER my_collation_ci_handler =
 {
   NULL,			/* init */
@@ -9945,7 +9982,7 @@ static MY_CHARSET_HANDLER my_charset_handler=
   mbcharlen_gbk,
   my_numchars_mb,
   my_charpos_mb,
-  my_well_formed_len_mb,
+  my_well_formed_len_gbk,
   my_lengthsp_8bit,
   my_numcells_8bit,
   my_mb_wc_gbk,