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.

21 files changed, 437 insertions, 66 deletions

diff --git a/include/m_ctype.h b/include/m_ctype.h
index 8116058d687..603bb3cc2e5 100644
--- a/include/m_ctype.h
+++ b/include/m_ctype.h
@@ -115,12 +115,17 @@ typedef struct my_collation_handler_st
 
   int  (*strcasecmp)(struct charset_info_st *, const char *, const char *);
   
+  int  (*instr)(struct charset_info_st *,
+                const char *big,   uint b_length,
+                const char *small, uint s_length);
+  
   /* Hash calculation */
   void (*hash_sort)(struct charset_info_st *cs, const uchar *key, uint len,
 		    ulong *nr1, ulong *nr2); 
 } MY_COLLATION_HANDLER;
 
-extern MY_COLLATION_HANDLER my_collation_bin_handler;
+extern MY_COLLATION_HANDLER my_collation_mb_bin_handler;
+extern MY_COLLATION_HANDLER my_collation_8bit_bin_handler;
 extern MY_COLLATION_HANDLER my_collation_8bit_simple_ci_handler;
 
 
@@ -243,6 +248,10 @@ extern void my_hash_sort_simple(CHARSET_INFO *cs,
 
 extern uint my_lengthsp_8bit(CHARSET_INFO *cs, const char *ptr, uint length);
 
+extern int  my_instr_simple(struct charset_info_st *,
+                            const char *big,   uint b_length,
+                            const char *small, uint s_length);
+
 
 /* Functions for 8bit */
 extern void my_caseup_str_8bit(CHARSET_INFO *, char *);
@@ -307,6 +316,9 @@ int my_wildcmp_mb(CHARSET_INFO *,
 		  int escape, int w_one, int w_many);
 uint my_numchars_mb(CHARSET_INFO *, const char *b, const char *e);
 uint my_charpos_mb(CHARSET_INFO *, const char *b, const char *e, uint pos);
+int  my_instr_mb(struct charset_info_st *,
+                 const char *big,   uint b_length,
+                 const char *small, uint s_length);
 
 
 extern my_bool my_parse_charset_xml(const char *bug, uint len,
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;
diff --git a/mysys/charset.c b/mysys/charset.c
index 58ce8f5b2fe..e8406173b33 100644
--- a/mysys/charset.c
+++ b/mysys/charset.c
@@ -119,7 +119,7 @@ static void simple_cs_init_functions(CHARSET_INFO *cs)
   
   if (cs->state & MY_CS_BINSORT)
   {
-    cs->coll= &my_collation_bin_handler;
+    cs->coll= &my_collation_8bit_bin_handler;
   }
   else
   {
diff --git a/sql/item_func.cc b/sql/item_func.cc
index fe419745b60..35df0dbe40d 100644
--- a/sql/item_func.cc
+++ b/sql/item_func.cc
@@ -1153,7 +1153,6 @@ longlong Item_func_locate::val_int()
 {
   String *a=args[0]->val_str(&value1);
   String *b=args[1]->val_str(&value2);
-  bool binary_cmp= (cmp_collation.collation->state & MY_CS_BINSORT) ? 1 : 0;
   if (!a || !b)
   {
     null_value=1;
@@ -1161,55 +1160,26 @@ longlong Item_func_locate::val_int()
   }
   null_value=0;
   uint start=0;
-#ifdef USE_MB
   uint start0=0;
-#endif
+  int  ind;
+
   if (arg_count == 3)
   {
-    start=(uint) args[2]->val_int()-1;
-#ifdef USE_MB
-    if (use_mb(cmp_collation.collation))
-    {
-      start0=start;
-      if (!binary_cmp)
-        start=a->charpos(start);
-    }
-#endif
+    start0= start =(uint) args[2]->val_int()-1;
+    start=a->charpos(start);
+    
     if (start > a->length() || start+b->length() > a->length())
       return 0;
   }
+
   if (!b->length())				// Found empty string at start
     return (longlong) (start+1);
-#ifdef USE_MB
-  if (use_mb(cmp_collation.collation) && !binary_cmp)
-  {
-    const char *ptr=a->ptr()+start;
-    const char *search=b->ptr();
-    const char *strend = ptr+a->length();
-    const char *end=strend-b->length()+1;
-    const char *search_end=search+b->length();
-    register  uint32 l;
-    while (ptr < end)
-    {
-      if (*ptr == *search)
-      {
-        register char *i,*j;
-        i=(char*) ptr+1; j=(char*) search+1;
-        while (j != search_end)
-          if (*i++ != *j++) goto skipp;
-        return (longlong) start0+1;
-      }
-  skipp:
-      if ((l=my_ismbchar(cmp_collation.collation,ptr,strend)))
-	ptr+=l;
-      else ++ptr;
-      ++start0;
-    }
-    return 0;
-  }
-#endif /* USE_MB */
-  return (longlong) (binary_cmp ? a->strstr(*b,start) :
-		     (a->strstr_case(*b,start)))+1;
+  
+  ind= cmp_collation.collation->coll->instr(cmp_collation.collation,
+                                            a->ptr()+start, a->length()-start,
+                                            b->ptr(), b->length());
+
+  return (longlong) (ind >= 0 ? ind + start0 + 1 : ind + 1);
 }
 
 
diff --git a/strings/ctype-big5.c b/strings/ctype-big5.c
index ddcec46474d..878493bc31f 100644
--- a/strings/ctype-big5.c
+++ b/strings/ctype-big5.c
@@ -6234,6 +6234,7 @@ static MY_COLLATION_HANDLER my_collation_big5_chinese_ci_handler =
   my_like_range_big5,
   my_wildcmp_mb,
   my_strcasecmp_mb,
+  my_instr_mb,
   my_hash_sort_simple
 };
 
@@ -6305,7 +6306,7 @@ CHARSET_INFO my_charset_big5_bin=
     2,			/* mbmaxlen   */
     0,
     &my_charset_big5_handler,
-    &my_collation_bin_handler
+    &my_collation_mb_bin_handler
 };
 
 
diff --git a/strings/ctype-bin.c b/strings/ctype-bin.c
index 9441268739a..3f74f514c48 100644
--- a/strings/ctype-bin.c
+++ b/strings/ctype-bin.c
@@ -262,8 +262,46 @@ static int my_strnxfrm_bin(CHARSET_INFO *cs __attribute__((unused)),
   return len;
 }
 
+static
+int my_instr_bin(CHARSET_INFO *cs __attribute__((unused)),
+                 const char *big,   uint b_length, 
+		 const char *small, uint s_length)
+{
+  register const uchar *str, *search, *end, *search_end;
+  
+  if (s_length <= b_length)
+  {
+    if (!s_length)
+      return 0;		// Empty string is always found
+    
+    str= (const uchar*) big;
+    search= (const uchar*) small;
+    end= (const uchar*) big+b_length-s_length+1;
+    search_end= (const uchar*) small + s_length;
+    
+skipp:
+    while (str != end)
+    {
+      if ( (*str++) == (*search))
+      {
+	register const uchar *i,*j;
+	
+	i= str; 
+	j= search+1;
+	
+	while (j != search_end)
+	  if ((*i++) != (*j++))
+            goto skipp;
+        
+	return (int) (str- (const uchar*)big) -1;
+      }
+    }
+  }
+  return -1;
+}
+
 
-MY_COLLATION_HANDLER my_collation_bin_handler =
+MY_COLLATION_HANDLER my_collation_8bit_bin_handler =
 {
     my_strnncoll_binary,
     my_strnncollsp_binary,
@@ -271,6 +309,7 @@ MY_COLLATION_HANDLER my_collation_bin_handler =
     my_like_range_simple,
     my_wildcmp_bin,
     my_strcasecmp_bin,
+    my_instr_bin,
     my_hash_sort_bin
 };
 
@@ -317,5 +356,5 @@ CHARSET_INFO my_charset_bin =
     1,				/* mbmaxlen      */
     (char) 255,			/* max_sort_char */
     &my_charset_handler,
-    &my_collation_bin_handler
+    &my_collation_8bit_bin_handler
 };
diff --git a/strings/ctype-czech.c b/strings/ctype-czech.c
index 2398bc33d45..b2e4f1886ed 100644
--- a/strings/ctype-czech.c
+++ b/strings/ctype-czech.c
@@ -612,6 +612,7 @@ static MY_COLLATION_HANDLER my_collation_latin2_czech_ci_handler =
   my_like_range_czech,
   my_wildcmp_8bit,
   my_strcasecmp_8bit,
+  my_instr_simple,
   my_hash_sort_simple,
 };
 
diff --git a/strings/ctype-euc_kr.c b/strings/ctype-euc_kr.c
index bb3e66a8c6f..d47c4268642 100644
--- a/strings/ctype-euc_kr.c
+++ b/strings/ctype-euc_kr.c
@@ -8637,12 +8637,13 @@ my_mb_wc_euc_kr(CHARSET_INFO *cs __attribute__((unused)),
 
 static MY_COLLATION_HANDLER my_collation_ci_handler =
 {
-  my_strnncoll_simple,/* strnncoll  */
+  my_strnncoll_simple,  /* strnncoll  */
   my_strnncollsp_simple,
   my_strnxfrm_simple,	/* strnxfrm   */
-  my_like_range_simple,/* like_range */
+  my_like_range_simple, /* like_range */
   my_wildcmp_mb,	/* wildcmp    */
   my_strcasecmp_mb,
+  my_instr_mb,
   my_hash_sort_simple,
 };
 
@@ -8714,7 +8715,7 @@ CHARSET_INFO my_charset_euckr_bin=
     2,			/* mbmaxlen   */
     0,
     &my_charset_handler,
-    &my_collation_bin_handler
+    &my_collation_mb_bin_handler
 };
 
 #endif
diff --git a/strings/ctype-gb2312.c b/strings/ctype-gb2312.c
index b8cee35b186..d429fa34eea 100644
--- a/strings/ctype-gb2312.c
+++ b/strings/ctype-gb2312.c
@@ -5687,12 +5687,13 @@ my_mb_wc_gb2312(CHARSET_INFO *cs  __attribute__((unused)),
 
 static MY_COLLATION_HANDLER my_collation_ci_handler =
 {
-  my_strnncoll_simple,/* strnncoll  */
+  my_strnncoll_simple,  /* strnncoll  */
   my_strnncollsp_simple,
   my_strnxfrm_simple,	/* strnxfrm   */
-  my_like_range_simple,/* like_range */
+  my_like_range_simple, /* like_range */
   my_wildcmp_mb,	/* wildcmp    */
-  my_strcasecmp_mb,
+  my_strcasecmp_mb,     /* instr      */
+  my_instr_mb,
   my_hash_sort_simple,
 };
 
@@ -5763,7 +5764,7 @@ CHARSET_INFO my_charset_gb2312_bin=
     2,			/* mbmaxlen   */
     0,
     &my_charset_handler,
-    &my_collation_bin_handler
+    &my_collation_mb_bin_handler
 };
 
 #endif
diff --git a/strings/ctype-gbk.c b/strings/ctype-gbk.c
index 4a84eabaf8c..fa7aa175103 100644
--- a/strings/ctype-gbk.c
+++ b/strings/ctype-gbk.c
@@ -9890,6 +9890,7 @@ static MY_COLLATION_HANDLER my_collation_ci_handler =
   my_like_range_gbk,
   my_wildcmp_mb,
   my_strcasecmp_mb,
+  my_instr_mb,
   my_hash_sort_simple,
 };
 
@@ -9960,7 +9961,7 @@ CHARSET_INFO my_charset_gbk_bin=
     2,			/* mbmaxlen */
     0,
     &my_charset_handler,
-    &my_collation_bin_handler
+    &my_collation_mb_bin_handler
 };
 
 
diff --git a/strings/ctype-latin1.c b/strings/ctype-latin1.c
index edbd350f364..a8a5329f844 100644
--- a/strings/ctype-latin1.c
+++ b/strings/ctype-latin1.c
@@ -390,6 +390,7 @@ static MY_COLLATION_HANDLER my_collation_german2_ci_handler=
   my_like_range_simple,
   my_wildcmp_8bit,
   my_strcasecmp_8bit,
+  my_instr_simple,
   my_hash_sort_simple
 };
 
@@ -435,6 +436,6 @@ CHARSET_INFO my_charset_latin1_bin=
   1,					/* mbmaxlen  */
   0,
   &my_charset_handler,
-  &my_collation_bin_handler
+  &my_collation_8bit_bin_handler
 };
 
diff --git a/strings/ctype-mb.c b/strings/ctype-mb.c
index 3d1abe95675..03323b3d3a1 100644
--- a/strings/ctype-mb.c
+++ b/strings/ctype-mb.c
@@ -126,11 +126,7 @@ int my_strcasecmp_mb(CHARSET_INFO * cs,const char *s, const char *t)
 #define INC_PTR(cs,A,B) A+=((use_mb_flag && \
                           my_ismbchar(cs,A,B)) ? my_ismbchar(cs,A,B) : 1)
 
-#ifdef LIKE_CMP_TOUPPER
-#define likeconv(s,A) (uchar) my_toupper(s,A)
-#else
 #define likeconv(s,A) (uchar) (s)->sort_order[(uchar) (A)]
-#endif
 
 int my_wildcmp_mb(CHARSET_INFO *cs,
 		  const char *str,const char *str_end,
@@ -278,5 +274,224 @@ uint my_charpos_mb(CHARSET_INFO *cs __attribute__((unused)),
   return b-b0;
 }
 
+int my_instr_mb(CHARSET_INFO *cs,
+                const char *big,   uint b_length, 
+                const char *small, uint s_length)
+{
+  register const char *end;
+  int res= 0;
+  
+  if (s_length <= b_length)
+  {
+    if (!s_length)
+      return 0;		// Empty string is always found
+    
+    end= big+b_length-s_length+1;
+    
+    while (big < end)
+    {
+      int mblen;
+      
+      if (!cs->coll->strnncoll(cs, (unsigned char*) big,   s_length, 
+      				   (unsigned char*) small, s_length))
+        return res;
+      
+      mblen= (mblen= my_ismbchar(cs, big, end)) ? mblen : 1;
+      big+= mblen;
+      b_length-= mblen;
+      res++;
+    }
+  }
+  return -1;
+}
+
+/* BINARY collations handlers for MB charsets */
+
+static int my_strnncoll_mb_bin(CHARSET_INFO * cs __attribute__((unused)),
+				const uchar *s, uint slen,
+				const uchar *t, uint tlen)
+{
+  int cmp= memcmp(s,t,min(slen,tlen));
+  return cmp ? cmp : (int) (slen - tlen);
+}
+
+static int my_strnncollsp_mb_bin(CHARSET_INFO * cs,
+                               const uchar *s, uint slen,
+                               const uchar *t, uint tlen)
+{
+  int len, cmp;
+
+  for ( ; slen && my_isspace(cs, s[slen-1]) ; slen--);
+  for ( ; tlen && my_isspace(cs, t[tlen-1]) ; tlen--);
+
+  len  = ( slen > tlen ) ? tlen : slen;
+
+  cmp= memcmp(s,t,len);
+  return cmp ? cmp : (int) (slen - tlen);
+}
+
+static int my_strnxfrm_mb_bin(CHARSET_INFO *cs __attribute__((unused)),
+			    uchar * dest, uint len,
+			    const uchar *src, 
+			    uint srclen __attribute__((unused)))
+{
+  if (dest != src)
+    memcpy(dest,src,len= min(len,srclen));
+  return len;
+}
+
+
+static int my_strcasecmp_mb_bin(CHARSET_INFO * cs __attribute__((unused)),
+		      const char *s, const char *t)
+{
+  return strcmp(s,t);
+}
+
+static void my_hash_sort_mb_bin(CHARSET_INFO *cs __attribute__((unused)),
+		      const uchar *key, uint len,ulong *nr1, ulong *nr2)
+{
+  const uchar *pos = key;
+  
+  key+= len;
+  
+  for (; pos < (uchar*) key ; pos++)
+  {
+    nr1[0]^=(ulong) ((((uint) nr1[0] & 63)+nr2[0]) * 
+	     ((uint)*pos)) + (nr1[0] << 8);
+    nr2[0]+=3;
+  }
+}
+
+static int my_wildcmp_mb_bin(CHARSET_INFO *cs,
+		  const char *str,const char *str_end,
+		  const char *wildstr,const char *wildend,
+		  int escape, int w_one, int w_many)
+{
+  int result= -1;				/* Not found, using wildcards */
+
+  bool use_mb_flag=use_mb(cs);
+
+  while (wildstr != wildend)
+  {
+    while (*wildstr != w_many && *wildstr != w_one)
+    {
+      int l;
+      if (*wildstr == escape && wildstr+1 != wildend)
+	wildstr++;
+      if (use_mb_flag &&
+          (l = my_ismbchar(cs, wildstr, wildend)))
+      {
+	  if (str+l > str_end || memcmp(str, wildstr, l) != 0)
+	      return 1;
+	  str += l;
+	  wildstr += l;
+      }
+      else
+      if (str == str_end || *wildstr++ != *str++)
+	return(1);				/* No match */
+      if (wildstr == wildend)
+	return (str != str_end);		/* Match if both are at end */
+      result=1;					/* Found an anchor char */
+    }
+    if (*wildstr == w_one)
+    {
+      do
+      {
+	if (str == str_end)			/* Skip one char if possible */
+	  return (result);
+	INC_PTR(cs,str,str_end);
+      } while (++wildstr < wildend && *wildstr == w_one);
+      if (wildstr == wildend)
+	break;
+    }
+    if (*wildstr == w_many)
+    {						/* Found w_many */
+      uchar cmp;
+      const char* mb = wildstr;
+      int mblen=0;
+      
+      wildstr++;
+      /* Remove any '%' and '_' from the wild search string */
+      for (; wildstr != wildend ; wildstr++)
+      {
+	if (*wildstr == w_many)
+	  continue;
+	if (*wildstr == w_one)
+	{
+	  if (str == str_end)
+	    return (-1);
+	  INC_PTR(cs,str,str_end);
+	  continue;
+	}
+	break;					/* Not a wild character */
+      }
+      if (wildstr == wildend)
+	return(0);				/* Ok if w_many is last */
+      if (str == str_end)
+	return -1;
+      
+      if ((cmp= *wildstr) == escape && wildstr+1 != wildend)
+	cmp= *++wildstr;
+	
+      mb=wildstr;
+      LINT_INIT(mblen);
+      if (use_mb_flag)
+        mblen = my_ismbchar(cs, wildstr, wildend);
+      INC_PTR(cs,wildstr,wildend);		/* This is compared trough cmp */
+      do
+      {
+        if (use_mb_flag)
+	{
+          for (;;)
+          {
+            if (str >= str_end)
+              return -1;
+            if (mblen)
+            {
+              if (str+mblen <= str_end && memcmp(str, mb, mblen) == 0)
+              {
+                str += mblen;
+                break;
+              }
+            }
+            else if (!my_ismbchar(cs, str, str_end) && *str == cmp)
+            {
+              str++;
+              break;
+            }
+            INC_PTR(cs,str, str_end);
+          }
+	}
+        else
+        {
+          while (str != str_end && *str != cmp)
+            str++;
+          if (str++ == str_end) return (-1);
+        }
+	{
+	  int tmp=my_wildcmp_mb(cs,str,str_end,wildstr,wildend,escape,w_one,w_many);
+	  if (tmp <= 0)
+	    return (tmp);
+	}
+      } while (str != str_end && wildstr[0] != w_many);
+      return(-1);
+    }
+  }
+  return (str != str_end ? 1 : 0);
+}
+
+
+MY_COLLATION_HANDLER my_collation_mb_bin_handler =
+{
+    my_strnncoll_mb_bin,
+    my_strnncollsp_mb_bin,
+    my_strnxfrm_mb_bin,
+    my_like_range_simple,
+    my_wildcmp_mb_bin,
+    my_strcasecmp_mb_bin,
+    my_instr_mb,
+    my_hash_sort_mb_bin
+};
+
 
 #endif
diff --git a/strings/ctype-simple.c b/strings/ctype-simple.c
index ca0097579bd..152980dd305 100644
--- a/strings/ctype-simple.c
+++ b/strings/ctype-simple.c
@@ -1030,6 +1030,44 @@ uint my_lengthsp_8bit(CHARSET_INFO *cs __attribute__((unused)),
 }
 
 
+int my_instr_simple(CHARSET_INFO *cs,
+                    const char *big,   uint b_length, 
+		    const char *small, uint s_length)
+{
+  register const uchar *str, *search, *end, *search_end;
+  
+  if (s_length <= b_length)
+  {
+    if (!s_length)
+      return 0;		// Empty string is always found
+    
+    str= (const uchar*) big;
+    search= (const uchar*) small;
+    end= (const uchar*) big+b_length-s_length+1;
+    search_end= (const uchar*) small + s_length;
+    
+skipp:
+    while (str != end)
+    {
+      if (cs->sort_order[*str++] == cs->sort_order[*search])
+      {
+	register const uchar *i,*j;
+	
+	i= str; 
+	j= search+1;
+	
+	while (j != search_end)
+	  if (cs->sort_order[*i++] != cs->sort_order[*j++]) 
+            goto skipp;
+        
+	return (int) (str- (const uchar*)big) -1;
+      }
+    }
+  }
+  return -1;
+}
+
+
 MY_CHARSET_HANDLER my_charset_8bit_handler=
 {
     NULL,			/* ismbchar      */
@@ -1063,5 +1101,6 @@ MY_COLLATION_HANDLER my_collation_8bit_simple_ci_handler =
     my_like_range_simple,
     my_wildcmp_8bit,
     my_strcasecmp_8bit,
+    my_instr_simple,
     my_hash_sort_simple
 };
diff --git a/strings/ctype-sjis.c b/strings/ctype-sjis.c
index 9827c19e7fb..f302e678b9f 100644
--- a/strings/ctype-sjis.c
+++ b/strings/ctype-sjis.c
@@ -4477,6 +4477,7 @@ static MY_COLLATION_HANDLER my_collation_ci_handler =
   my_like_range_sjis,
   my_wildcmp_mb,	/* wildcmp  */
   my_strcasecmp_8bit,
+  my_instr_mb,
   my_hash_sort_simple,
 };
 
@@ -4547,7 +4548,7 @@ CHARSET_INFO my_charset_sjis_bin=
     2,			/* mbmaxlen */
     0,
     &my_charset_handler,
-    &my_collation_bin_handler
+    &my_collation_mb_bin_handler
 };
 
 #endif
diff --git a/strings/ctype-tis620.c b/strings/ctype-tis620.c
index 8a6e00b973a..a4d8a7d1f79 100644
--- a/strings/ctype-tis620.c
+++ b/strings/ctype-tis620.c
@@ -710,6 +710,7 @@ static MY_COLLATION_HANDLER my_collation_ci_handler =
     my_like_range_tis620,
     my_wildcmp_8bit,	/* wildcmp   */
     my_strcasecmp_8bit,
+    NULL,
     my_hash_sort_simple,
 };
 
@@ -781,7 +782,7 @@ CHARSET_INFO my_charset_tis620_bin=
     1,			/* mbmaxlen  */
     0,
     &my_charset_handler,
-    &my_collation_bin_handler
+    &my_collation_8bit_bin_handler
 };
 
 
diff --git a/strings/ctype-ucs2.c b/strings/ctype-ucs2.c
index a20502c65d4..31c0f063529 100644
--- a/strings/ctype-ucs2.c
+++ b/strings/ctype-ucs2.c
@@ -1028,6 +1028,7 @@ static MY_COLLATION_HANDLER my_collation_ci_handler =
     my_like_range_simple,
     my_wildcmp_mb,
     my_strcasecmp_ucs2,
+    my_instr_mb,
     my_hash_sort_ucs2
 };
 
@@ -1100,7 +1101,7 @@ CHARSET_INFO my_charset_ucs2_bin=
     2,			/* mbmaxlen     */
     0,
     &my_charset_handler,
-    &my_collation_bin_handler
+    &my_collation_mb_bin_handler
 };
 
 
diff --git a/strings/ctype-ujis.c b/strings/ctype-ujis.c
index 5ef6c1b7486..29375aca727 100644
--- a/strings/ctype-ujis.c
+++ b/strings/ctype-ujis.c
@@ -8434,6 +8434,7 @@ static MY_COLLATION_HANDLER my_collation_ci_handler =
     my_like_range_simple,/* like_range   */
     my_wildcmp_mb,	/* wildcmp      */
     my_strcasecmp_mb,
+    my_instr_mb,
     my_hash_sort_simple,
 };
 
@@ -8504,7 +8505,7 @@ CHARSET_INFO my_charset_ujis_bin=
     3,			/* mbmaxlen     */
     0,
     &my_charset_handler,
-    &my_collation_bin_handler
+    &my_collation_mb_bin_handler
 };
 
 
diff --git a/strings/ctype-utf8.c b/strings/ctype-utf8.c
index 450c2d7aa93..3ede1aa26f6 100644
--- a/strings/ctype-utf8.c
+++ b/strings/ctype-utf8.c
@@ -1959,6 +1959,7 @@ static MY_COLLATION_HANDLER my_collation_ci_handler =
     my_like_range_simple,
     my_wildcmp_mb,
     my_strcasecmp_utf8,
+    my_instr_mb,
     my_hash_sort_utf8
 };
 
@@ -2031,7 +2032,7 @@ CHARSET_INFO my_charset_utf8_bin=
     3,			/* mbmaxlen     */
     0,
     &my_charset_handler,
-    &my_collation_bin_handler
+    &my_collation_mb_bin_handler
 };
 
 
diff --git a/strings/ctype-win1250ch.c b/strings/ctype-win1250ch.c
index 12a8edc4a8f..60a5737009f 100644
--- a/strings/ctype-win1250ch.c
+++ b/strings/ctype-win1250ch.c
@@ -652,6 +652,7 @@ static MY_COLLATION_HANDLER my_collation_czech_ci_handler =
     my_like_range_win1250ch,
     my_wildcmp_8bit,
     my_strcasecmp_8bit,
+    my_instr_simple,
     my_hash_sort_simple
 };