Bug #3453 MySQL output formatting in multibyte character sets

6 files changed, 70 insertions, 6 deletions

diff --git a/strings/ctype-big5.c b/strings/ctype-big5.c
index 3f35f7504ac..a7226c10ce3 100644
--- a/strings/ctype-big5.c
+++ b/strings/ctype-big5.c
@@ -6290,7 +6290,7 @@ static MY_CHARSET_HANDLER my_charset_big5_handler=
   my_charpos_mb,
   my_well_formed_len_mb,
   my_lengthsp_8bit,
-  my_numcells_mb,
+  my_numcells_8bit,
   my_mb_wc_big5,	/* mb_wc       */
   my_wc_mb_big5,	/* wc_mb       */
   my_caseup_str_mb,
diff --git a/strings/ctype-euc_kr.c b/strings/ctype-euc_kr.c
index 43a50b0dfbe..0bf9a5210ea 100644
--- a/strings/ctype-euc_kr.c
+++ b/strings/ctype-euc_kr.c
@@ -8657,7 +8657,7 @@ static MY_CHARSET_HANDLER my_charset_handler=
   my_charpos_mb,
   my_well_formed_len_mb,
   my_lengthsp_8bit,
-  my_numcells_mb,
+  my_numcells_8bit,
   my_mb_wc_euc_kr,	/* mb_wc   */
   my_wc_mb_euc_kr,	/* wc_mb   */
   my_caseup_str_mb,
diff --git a/strings/ctype-gb2312.c b/strings/ctype-gb2312.c
index 8d97ac9ca1d..726f9fd464c 100644
--- a/strings/ctype-gb2312.c
+++ b/strings/ctype-gb2312.c
@@ -5708,7 +5708,7 @@ static MY_CHARSET_HANDLER my_charset_handler=
   my_charpos_mb,
   my_well_formed_len_mb,
   my_lengthsp_8bit,
-  my_numcells_mb,
+  my_numcells_8bit,
   my_mb_wc_gb2312,	/* mb_wc      */
   my_wc_mb_gb2312,	/* wc_mb      */
   my_caseup_str_mb,
diff --git a/strings/ctype-gbk.c b/strings/ctype-gbk.c
index 9400fb08f2b..8a7df87c024 100644
--- a/strings/ctype-gbk.c
+++ b/strings/ctype-gbk.c
@@ -9939,7 +9939,7 @@ static MY_CHARSET_HANDLER my_charset_handler=
   my_charpos_mb,
   my_well_formed_len_mb,
   my_lengthsp_8bit,
-  my_numcells_mb,
+  my_numcells_8bit,
   my_mb_wc_gbk,
   my_wc_mb_gbk,
   my_caseup_str_mb,
diff --git a/strings/ctype-sjis.c b/strings/ctype-sjis.c
index b4cfee0f24a..8ed2e9d9c0b 100644
--- a/strings/ctype-sjis.c
+++ b/strings/ctype-sjis.c
@@ -4534,6 +4534,36 @@ my_mb_wc_sjis(CHARSET_INFO *cs  __attribute__((unused)),
   return 2;
 }
 
+static
+uint my_numcells_sjis(CHARSET_INFO *cs __attribute__((unused)),
+                      const char *str, const char *strend)
+{
+  uint clen= 0;
+  const unsigned char *b= (const unsigned char *) str;
+  const unsigned char *e= (const unsigned char *) strend;
+  
+  for (clen= 0; b < e; )
+  {
+    if (*b >= 0xA1 && *b <= 0xDF)
+    {
+      clen++;
+      b++;
+    }
+    else if (*b > 0x7F)
+    {
+      clen+= 2;
+      b+= 2;
+    }
+    else
+    {
+      clen++;
+      b++;
+    }
+  }
+  return clen;
+}
+
+
 
 static MY_COLLATION_HANDLER my_collation_ci_handler =
 {
@@ -4558,7 +4588,7 @@ static MY_CHARSET_HANDLER my_charset_handler=
   my_charpos_mb,
   my_well_formed_len_mb,
   my_lengthsp_8bit,
-  my_numcells_mb,
+  my_numcells_sjis,
   my_mb_wc_sjis,	/* mb_wc */
   my_wc_mb_sjis,	/* wc_mb */
   my_caseup_str_8bit,
diff --git a/strings/ctype-ujis.c b/strings/ctype-ujis.c
index 37c26a3bbc4..514f28275ca 100644
--- a/strings/ctype-ujis.c
+++ b/strings/ctype-ujis.c
@@ -8252,6 +8252,40 @@ my_jisx0212_uni_onechar(int code){
   [xA1-xFE][xA1-xFE]		# JIS X 0208:1997 (two bytes/char)
 */
 
+static
+uint my_numcells_eucjp(CHARSET_INFO *cs __attribute__((unused)),
+                       const char *str, const char *strend)
+{
+  uint clen= 0;
+  const unsigned char *b= (const unsigned char *) str;
+  const unsigned char *e= (const unsigned char *) strend;
+  
+  for (clen= 0; b < e; )
+  {
+    if (*b == 0x8E)
+    {
+      clen++;
+      b+= 2;
+    }
+    else if (*b == 0x8F)
+    {
+      clen+= 2;
+      b+= 3;
+    }
+    else if (*b & 0x80)
+    {
+      clen+= 2;
+      b+= 2;
+    }
+    else
+    {
+      clen++;
+      b++;
+    }
+  }
+  return clen;
+}
+
 static int
 my_mb_wc_euc_jp(CHARSET_INFO *cs,my_wc_t *pwc, const uchar *s, const uchar *e)
 {
@@ -8443,7 +8477,7 @@ static MY_CHARSET_HANDLER my_charset_handler=
     my_charpos_mb,
     my_well_formed_len_mb,
     my_lengthsp_8bit,
-    my_numcells_mb,
+    my_numcells_eucjp,
     my_mb_wc_euc_jp,	/* mb_wc       */
     my_wc_mb_euc_jp,	/* wc_mb       */
     my_caseup_str_mb,