Further fixes for 2390: ucs2 alignment

mysql-test/r/ctype_ucs.result

@@ -324,3 +324,59 @@ HEX(_ucs2 0x0123456789ABCDE)
 SELECT HEX(_ucs2 0x0123456789ABCDEF);
 HEX(_ucs2 0x0123456789ABCDEF)
 0123456789ABCDEF
+SELECT hex(cast(0xAA as char character set ucs2));
+hex(cast(0xAA as char character set ucs2))
+00AA
+SELECT hex(convert(0xAA using ucs2));
+hex(convert(0xAA using ucs2))
+00AA
+CREATE TABLE t1 (a char(10) character set ucs2);
+INSERT INTO t1 VALUES (0xA),(0xAA),(0xAAA),(0xAAAA),(0xAAAAA);
+SELECT HEX(a) FROM t1;
+HEX(a)
+000A
+00AA
+0AAA
+AAAA
+000AAAAA
+DROP TABLE t1;
+CREATE TABLE t1 (a varchar(10) character set ucs2);
+INSERT INTO t1 VALUES (0xA),(0xAA),(0xAAA),(0xAAAA),(0xAAAAA);
+SELECT HEX(a) FROM t1;
+HEX(a)
+000A
+00AA
+0AAA
+AAAA
+000AAAAA
+DROP TABLE t1;
+CREATE TABLE t1 (a text character set ucs2);
+INSERT INTO t1 VALUES (0xA),(0xAA),(0xAAA),(0xAAAA),(0xAAAAA);
+SELECT HEX(a) FROM t1;
+HEX(a)
+000A
+00AA
+0AAA
+AAAA
+000AAAAA
+DROP TABLE t1;
+CREATE TABLE t1 (a mediumtext character set ucs2);
+INSERT INTO t1 VALUES (0xA),(0xAA),(0xAAA),(0xAAAA),(0xAAAAA);
+SELECT HEX(a) FROM t1;
+HEX(a)
+000A
+00AA
+0AAA
+AAAA
+000AAAAA
+DROP TABLE t1;
+CREATE TABLE t1 (a longtext character set ucs2);
+INSERT INTO t1 VALUES (0xA),(0xAA),(0xAAA),(0xAAAA),(0xAAAAA);
+SELECT HEX(a) FROM t1;
+HEX(a)
+000A
+00AA
+0AAA
+AAAA
+000AAAAA
+DROP TABLE t1;

mysql-test/t/ctype_ucs.test

@@ -200,7 +200,7 @@ DROP TABLE t1;
 
 
 # Bug #2390
-# Check alignment
+# Check alignment for constants
 #
 SELECT HEX(_ucs2 0x0);
 SELECT HEX(_ucs2 0x01);
@@ -218,3 +218,39 @@ SELECT HEX(_ucs2 0x0123456789ABC);
 SELECT HEX(_ucs2 0x0123456789ABCD);
 SELECT HEX(_ucs2 0x0123456789ABCDE);
 SELECT HEX(_ucs2 0x0123456789ABCDEF);
+
+#
+# Check alignment for from-binary-conversion with CAST and CONVERT
+#
+SELECT hex(cast(0xAA as char character set ucs2));
+SELECT hex(convert(0xAA using ucs2));
+
+#
+# Check alignment for string types
+#
+CREATE TABLE t1 (a char(10) character set ucs2);
+INSERT INTO t1 VALUES (0xA),(0xAA),(0xAAA),(0xAAAA),(0xAAAAA);
+SELECT HEX(a) FROM t1;
+DROP TABLE t1;
+
+CREATE TABLE t1 (a varchar(10) character set ucs2);
+INSERT INTO t1 VALUES (0xA),(0xAA),(0xAAA),(0xAAAA),(0xAAAAA);
+SELECT HEX(a) FROM t1;
+DROP TABLE t1;
+
+CREATE TABLE t1 (a text character set ucs2);
+INSERT INTO t1 VALUES (0xA),(0xAA),(0xAAA),(0xAAAA),(0xAAAAA);
+SELECT HEX(a) FROM t1;
+DROP TABLE t1;
+
+CREATE TABLE t1 (a mediumtext character set ucs2);
+INSERT INTO t1 VALUES (0xA),(0xAA),(0xAAA),(0xAAAA),(0xAAAAA);
+SELECT HEX(a) FROM t1;
+DROP TABLE t1;
+
+CREATE TABLE t1 (a longtext character set ucs2);
+INSERT INTO t1 VALUES (0xA),(0xAA),(0xAAA),(0xAAAA),(0xAAAAA);
+SELECT HEX(a) FROM t1;
+DROP TABLE t1;
+
+-- the same should be also done with enum and set

sql/field.cc

@@ -4055,18 +4055,13 @@ void Field_datetime::sql_type(String &res) const
 
 	/* Copy a string and fill with space */
 
-static bool use_conversion(CHARSET_INFO *cs1, CHARSET_INFO *cs2)
-{
-  return (cs1 != &my_charset_bin) && (cs2 != &my_charset_bin) && (cs1!=cs2);
-}
-
 int Field_string::store(const char *from,uint length,CHARSET_INFO *cs)
 {
   int error= 0;
   char buff[80];
   String tmpstr(buff,sizeof(buff), &my_charset_bin);
   /* Convert character set if nesessary */
-  if (use_conversion(cs, field_charset))
+  if (String::needs_conversion(from, length, cs, field_charset))
   { 
     tmpstr.copy(from, length, cs, field_charset);
     from= tmpstr.ptr();
@@ -4254,7 +4249,7 @@ int Field_varstring::store(const char *from,uint length,CHARSET_INFO *cs)
   char buff[80];
   String tmpstr(buff,sizeof(buff), &my_charset_bin);
   /* Convert character set if nesessary */
-  if (use_conversion(cs, field_charset))
+  if (String::needs_conversion(from, length, cs, field_charset))
   { 
     tmpstr.copy(from, length, cs, field_charset);
     from= tmpstr.ptr();
@@ -4572,7 +4567,8 @@ int Field_blob::store(const char *from,uint length,CHARSET_INFO *cs)
     String tmpstr(buff,sizeof(buff), &my_charset_bin);
 
     /* Convert character set if nesessary */
-    if ((was_conversion= use_conversion(cs, field_charset)))
+    if ((was_conversion= String::needs_conversion(from, length,
+						  cs, field_charset)))
     { 
       tmpstr.copy(from, length, cs, field_charset);
       from= tmpstr.ptr();
@@ -5082,7 +5078,7 @@ int Field_enum::store(const char *from,uint length,CHARSET_INFO *cs)
   char buff[80];
   String tmpstr(buff,sizeof(buff), &my_charset_bin);
   /* Convert character set if nesessary */
-  if (use_conversion(cs, field_charset))
+  if (String::needs_conversion(from, length, cs, field_charset))
   { 
     tmpstr.copy(from, length, cs, field_charset);
     from= tmpstr.ptr();
@@ -5263,7 +5259,7 @@ int Field_set::store(const char *from,uint length,CHARSET_INFO *cs)
   String tmpstr(buff,sizeof(buff), &my_charset_bin);
 
   /* Convert character set if nesessary */
-  if (use_conversion(cs, field_charset))
+  if (String::needs_conversion(from, length, cs, field_charset))
   { 
     tmpstr.copy(from, length, cs, field_charset);
     from= tmpstr.ptr();

sql/sql_string.cc

@@ -228,6 +228,32 @@ bool String::copy(const char *str,uint32 arg_length, CHARSET_INFO *cs)
   return FALSE;
 }
 
+
+/*
+  Checks that the source string can be just copied
+  to the destination string without conversion.
+  If either character set conversion or adding leading
+  zeros (e.g. for UCS-2) must be done then return 
+  value is TRUE else FALSE.
+*/
+bool String::needs_conversion(const char *str, uint32 arg_length,
+				     CHARSET_INFO *from_cs,
+				     CHARSET_INFO *to_cs)
+{
+  if (to_cs == &my_charset_bin)
+    return FALSE;
+  if (to_cs == from_cs)
+    return FALSE;
+  if (my_charset_same(from_cs, to_cs))
+    return FALSE;
+  if ((from_cs == &my_charset_bin))
+  {
+    if (!(arg_length % to_cs->mbminlen))
+      return FALSE;
+  }
+  return TRUE;
+}
+
 /*
 ** For real multi-byte, ascii incompatible charactser sets,
 ** like UCS-2, add leading zeros if we have an incomplete character.
@@ -237,7 +263,7 @@ bool String::copy(const char *str,uint32 arg_length, CHARSET_INFO *cs)
 **   SELECT _ucs2 0x00AA
 */
 
-bool String::set_or_copy_aligned(const char *str,uint32 arg_length,
+bool String::copy_aligned(const char *str,uint32 arg_length,
 			  CHARSET_INFO *cs)
 {
   /* How many bytes are in incomplete character */
@@ -245,7 +271,7 @@ bool String::set_or_copy_aligned(const char *str,uint32 arg_length,
 
   if (!offs) /* All characters are complete, just copy */
   {
-    set(str, arg_length, cs);
+    copy(str, arg_length, cs);
     return FALSE;
   }
   
@@ -274,15 +300,35 @@ bool String::set_or_copy_aligned(const char *str,uint32 arg_length,
   return FALSE;
 }
 
+
+bool String::set_or_copy_aligned(const char *str,uint32 arg_length,
+				 CHARSET_INFO *cs)
+{
+  /* How many bytes are in incomplete character */
+  uint32 offs= (arg_length % cs->mbminlen); 
+  
+  if (!offs) /* All characters are complete, just copy */
+  {
+    set(str, arg_length, cs);
+    return FALSE;
+  }
+  return copy_aligned(str, arg_length, cs);
+}
+
 	/* Copy with charset convertion */
 
 bool String::copy(const char *str, uint32 arg_length,
 		  CHARSET_INFO *from_cs, CHARSET_INFO *to_cs)
 {
-  if ((from_cs == &my_charset_bin) || (to_cs == &my_charset_bin))
+  if (!needs_conversion(str, arg_length, from_cs, to_cs))
   {
     return copy(str, arg_length, to_cs);
   }
+  if ((from_cs == &my_charset_bin) && (arg_length % to_cs->mbminlen))
+  {
+    return copy_aligned(str, arg_length, to_cs);
+  }
+  
   uint32 new_length= to_cs->mbmaxlen*arg_length;
   if (alloc(new_length))
     return TRUE;

sql/sql_string.h

@@ -183,6 +183,9 @@ public:
   bool copy();					// Alloc string if not alloced
   bool copy(const String &s);			// Allocate new string
   bool copy(const char *s,uint32 arg_length, CHARSET_INFO *cs);	// Allocate new string
+  static bool needs_conversion(const char *s, uint32 arg_length,
+  			       CHARSET_INFO *cs_from, CHARSET_INFO *cs_to);
+  bool copy_aligned(const char *s, uint32 arg_length, CHARSET_INFO *cs);
   bool set_or_copy_aligned(const char *s, uint32 arg_length, CHARSET_INFO *cs);
   bool copy(const char*s,uint32 arg_length, CHARSET_INFO *csfrom,
 	    CHARSET_INFO *csto);