Many files:

  InnoDB true VARCHAR


sql/ha_innodb.h:
  InnoDB true VARCHAR
sql/ha_innodb.cc:
  InnoDB true VARCHAR
innobase/include/data0type.h:
  InnoDB true VARCHAR
innobase/include/que0que.h:
  InnoDB true VARCHAR
innobase/include/row0mysql.h:
  InnoDB true VARCHAR
innobase/include/data0type.ic:
  InnoDB true VARCHAR
innobase/include/row0mysql.ic:
  InnoDB true VARCHAR
innobase/row/row0ins.c:
  InnoDB true VARCHAR
innobase/row/row0mysql.c:
  InnoDB true VARCHAR
innobase/row/row0sel.c:
  InnoDB true VARCHAR
innobase/trx/trx0trx.c:
  InnoDB true VARCHAR

11 files changed, 654 insertions, 391 deletions

diff --git a/innobase/include/data0type.h b/innobase/include/data0type.h
index 174665ca1fa..b5120e22041 100644
--- a/innobase/include/data0type.h
+++ b/innobase/include/data0type.h
@@ -24,7 +24,11 @@ extern dtype_t* 	dtype_binary;
 /*-------------------------------------------*/
 /* The 'MAIN TYPE' of a column */
 #define	DATA_VARCHAR	1	/* character varying of the
-				latin1_swedish_ci charset-collation */
+				latin1_swedish_ci charset-collation; note
+				that the MySQL format for this, DATA_BINARY,
+				DATA_VARMYSQL, is also affected by whether the
+				'precise type' contains
+				DATA_MYSQL_TRUE_VARCHAR */
 #define DATA_CHAR	2	/* fixed length character of the
 				latin1_swedish_ci charset-collation */
 #define DATA_FIXBINARY	3	/* binary string of fixed length */
@@ -102,6 +106,8 @@ columns, and for them the precise type is usually not used at all.
 
 #define DATA_MYSQL_TYPE_MASK 255 /* AND with this mask to extract the MySQL
 				 type from the precise type */
+#define DATA_MYSQL_TRUE_VARCHAR 15 /* MySQL type code for the >= 5.0.3
+				   format true VARCHAR */
 
 /* Precise data types for system columns and the length of those columns;
 NOTE: the values must run from 0 up in the order given! All codes must
@@ -134,6 +140,10 @@ be less than 256 */
 				In earlier versions this was set for some
 				BLOB columns.
 */
+#define	DATA_LONG_TRUE_VARCHAR 4096	/* this is ORed to the precise data
+				type when the column is true VARCHAR where
+				MySQL uses 2 bytes to store the data len;
+				for shorter VARCHARs MySQL uses only 1 byte */
 /*-------------------------------------------*/
 
 /* This many bytes we need to store the type information affecting the
@@ -145,6 +155,15 @@ store the charset-collation number; one byte is left unused, though */
 #define DATA_NEW_ORDER_NULL_TYPE_BUF_SIZE	6
 
 /*************************************************************************
+Gets the MySQL type code from a dtype. */
+UNIV_INLINE
+ulint
+dtype_get_mysql_type(
+/*=================*/
+				/* out: MySQL type code; this is NOT an InnoDB
+				type code! */
+	dtype_t*	type);	/* in: type struct */
+/*************************************************************************
 Determine how many bytes the first n characters of the given string occupy.
 If the string is shorter than n characters, returns the number of bytes
 the characters in the string occupy. */
diff --git a/innobase/include/data0type.ic b/innobase/include/data0type.ic
index e63dde98974..bf04e1c9b27 100644
--- a/innobase/include/data0type.ic
+++ b/innobase/include/data0type.ic
@@ -33,6 +33,19 @@ dtype_get_charset_coll(
 }
 
 /*************************************************************************
+Gets the MySQL type code from a dtype. */
+UNIV_INLINE
+ulint
+dtype_get_mysql_type(
+/*=================*/
+				/* out: MySQL type code; this is NOT an InnoDB
+				type code! */
+	dtype_t*	type)	/* in: type struct */
+{
+	return(type->prtype & 0xFFUL);
+}
+
+/*************************************************************************
 Sets the mbminlen and mbmaxlen members of a data type structure. */
 UNIV_INLINE
 void
diff --git a/innobase/include/que0que.h b/innobase/include/que0que.h
index 298ec494750..4113e52d425 100644
--- a/innobase/include/que0que.h
+++ b/innobase/include/que0que.h
@@ -359,7 +359,8 @@ struct que_thr_struct{
 					the control came */
 	ulint		resource;	/* resource usage of the query thread
 					thus far */
-  ulint   lock_state;	/* lock state of thread (table or row) */
+  	ulint   	lock_state;	/* lock state of thread (table or
+					row) */
 };
 
 #define QUE_THR_MAGIC_N		8476583
diff --git a/innobase/include/row0mysql.h b/innobase/include/row0mysql.h
index 48a9d9bc941..e44d689b88b 100644
--- a/innobase/include/row0mysql.h
+++ b/innobase/include/row0mysql.h
@@ -22,36 +22,6 @@ Created 9/17/2000 Heikki Tuuri
 typedef struct row_prebuilt_struct row_prebuilt_t;
 
 /***********************************************************************
-Stores a variable-length field (like VARCHAR) length to dest, in the
-MySQL format. */
-UNIV_INLINE
-byte*
-row_mysql_store_var_len(
-/*====================*/
-			/* out: dest + 2 */
-	byte*	dest,	/* in: where to store */
-	ulint	len);	/* in: length, must fit in two bytes */
-/***********************************************************************
-Reads a MySQL format variable-length field (like VARCHAR) length and
-returns pointer to the field data. */
-UNIV_INLINE
-byte*
-row_mysql_read_var_ref(
-/*===================*/
-			/* out: field + 2 */
-	ulint*	len,	/* out: variable-length field length */
-	byte*	field);	/* in: field */
-/***********************************************************************
-Reads a MySQL format variable-length field (like VARCHAR) length and
-returns pointer to the field data. */
-
-byte*
-row_mysql_read_var_ref_noninline(
-/*=============================*/
-			/* out: field + 2 */
-	ulint*	len,	/* out: variable-length field length */
-	byte*	field);	/* in: field */
-/***********************************************************************
 Frees the blob heap in prebuilt when no longer needed. */
 
 void
@@ -60,6 +30,30 @@ row_mysql_prebuilt_free_blob_heap(
 	row_prebuilt_t*	prebuilt);	/* in: prebuilt struct of a
 					ha_innobase:: table handle */
 /***********************************************************************
+Stores a >= 5.0.3 format true VARCHAR length to dest, in the MySQL row
+format. */
+
+byte*
+row_mysql_store_true_var_len(
+/*=========================*/
+			/* out: pointer to the data, we skip the 1 or 2 bytes
+			at the start that are used to store the len */
+	byte*	dest,	/* in: where to store */
+	ulint	len,	/* in: length, must fit in two bytes */
+	ulint	lenlen);/* in: storage length of len: either 1 or 2 bytes */
+/***********************************************************************
+Reads a >= 5.0.3 format true VARCHAR length, in the MySQL row format, and
+returns a pointer to the data. */
+
+byte*
+row_mysql_read_true_varchar(
+/*========================*/
+			/* out: pointer to the data, we skip the 1 or 2 bytes
+			at the start that are used to store the len */
+	ulint*	len,	/* out: variable-length field length */
+	byte*	field,	/* in: field in the MySQL format */
+	ulint	lenlen);/* in: storage length of len: either 1 or 2 bytes */
+/***********************************************************************
 Stores a reference to a BLOB in the MySQL format. */
 
 void
@@ -83,24 +77,40 @@ row_mysql_read_blob_ref(
 	ulint	col_len);	/* in: BLOB reference length (not BLOB
 				length) */
 /******************************************************************
-Stores a non-SQL-NULL field given in the MySQL format in the Innobase
-format. */
-UNIV_INLINE
-void
+Stores a non-SQL-NULL field given in the MySQL format in the InnoDB format.
+The counterpart of this function is row_sel_field_store_in_mysql_format() in
+row0sel.c. */
+
+byte*
 row_mysql_store_col_in_innobase_format(
 /*===================================*/
-	dfield_t*	dfield,		/* in/out: dfield */
-	byte*		buf,		/* in/out: buffer for the converted
-					value */
+					/* out: up to which byte we used
+					buf in the conversion */
+	dfield_t*	dfield,		/* in/out: dfield where dtype
+					information must be already set when
+					this function is called! */
+	byte*		buf,		/* in/out: buffer for a converted
+					integer value; this must be at least
+					col_len long then! */
+	ibool		row_format_col,	/* TRUE if the mysql_data is from
+					a MySQL row, FALSE if from a MySQL
+					key value;
+					in MySQL, a true VARCHAR storage
+					format differs in a row and in a
+					key value: in a key value the length
+					is always stored in 2 bytes! */
 	byte*		mysql_data,	/* in: MySQL column value, not
 					SQL NULL; NOTE that dfield may also
 					get a pointer to mysql_data,
 					therefore do not discard this as long
 					as dfield is used! */
-	ulint		col_len,	/* in: MySQL column length */
-	ulint		type,		/* in: data type */
-	bool		comp,		/* in: TRUE=compact format */
-	ulint		is_unsigned);	/* in: != 0 if unsigned integer type */
+	ulint		col_len,	/* in: MySQL column length; NOTE that
+					this is the storage length of the
+					column in the MySQL format row, not
+					necessarily the length of the actual
+					payload data; if the column is a true
+					VARCHAR then this is irrelevant */
+	ibool		comp);		/* in: TRUE = compact format */
 /********************************************************************
 Handles user errors and lock waits detected by the database engine. */
 
@@ -457,6 +467,16 @@ struct mysql_row_templ_struct {
 					zero if column cannot be NULL */
 	ulint	type;			/* column type in Innobase mtype
 					numbers DATA_CHAR... */
+	ulint	mysql_type;		/* MySQL type code; this is always
+					< 256 */
+	ulint	mysql_length_bytes;	/* if mysql_type
+					== DATA_MYSQL_TRUE_VARCHAR, this tells
+					whether we should use 1 or 2 bytes to
+					store the MySQL true VARCHAR data
+					length at the start of row in the MySQL
+					format (NOTE that the MySQL key value
+					format always uses 2 bytes for the data
+					len) */ 
 	ulint	charset;		/* MySQL charset-collation code
 					of the column, or zero */
 	ulint	mbminlen;		/* minimum length of a char, in bytes,
diff --git a/innobase/include/row0mysql.ic b/innobase/include/row0mysql.ic
index 910546e298c..aa8a70d8761 100644
--- a/innobase/include/row0mysql.ic
+++ b/innobase/include/row0mysql.ic
@@ -5,149 +5,3 @@ MySQL interface for Innobase
 
 Created 1/23/2001 Heikki Tuuri
 *******************************************************/
-
-/***********************************************************************
-Stores a variable-length field (like VARCHAR) length to dest, in the
-MySQL format. No real var implemented in MySQL yet! */
-UNIV_INLINE
-byte*
-row_mysql_store_var_len(
-/*====================*/
-			/* out: dest + 2 */
-	byte*	dest,	/* in: where to store */
-	ulint	len __attribute__((unused)))  /* in: length, must fit in two
-                                                 bytes */
-{
-	ut_ad(len < 256 * 256);
-/*	
-	mach_write_to_2_little_endian(dest, len);
-
-	return(dest + 2);
-*/
-	return(dest);	/* No real var implemented in MySQL yet! */
-}
-
-/***********************************************************************
-Reads a MySQL format variable-length field (like VARCHAR) length and
-returns pointer to the field data. No real var implemented in MySQL yet! */
-UNIV_INLINE
-byte*
-row_mysql_read_var_ref(
-/*===================*/
-			/* out: field + 2 */
-	ulint*	len,	/* out: variable-length field length; does not work
-			yet! */
-	byte*	field)	/* in: field */
-{
-/*	
-	*len = mach_read_from_2_little_endian(field);
-
-	return(field + 2);
-*/
-	UT_NOT_USED(len);
-
-	return(field);	/* No real var implemented in MySQL yet! */
-}
-
-/******************************************************************
-Stores a non-SQL-NULL field given in the MySQL format in the Innobase
-format. */
-UNIV_INLINE
-void
-row_mysql_store_col_in_innobase_format(
-/*===================================*/
-	dfield_t*	dfield,		/* in/out: dfield */
-	byte*		buf,		/* in/out: buffer for the converted
-					value; this must be at least col_len
-					long! */
-	byte*		mysql_data,	/* in: MySQL column value, not
-					SQL NULL; NOTE that dfield may also
-					get a pointer to mysql_data,
-					therefore do not discard this as long
-					as dfield is used! */
-	ulint		col_len,	/* in: MySQL column length */
-	ulint		type,		/* in: data type */
-	bool		comp,		/* in: TRUE=compact format */
-	ulint		is_unsigned)	/* in: != 0 if unsigned integer type */
-{
-	byte*	ptr 	= mysql_data;
-
-	if (type == DATA_INT) {
-		/* Store integer data in Innobase in a big-endian format,
-		sign bit negated */
-
-		ptr = buf + col_len;
-
-		for (;;) {
-			ptr--;
-			*ptr = *mysql_data;
-			if (ptr == buf) {
-				break;
-			}
-			mysql_data++;
-		}
-
-		if (!is_unsigned) {
-			*ptr = (byte) (*ptr ^ 128);
-		}
-	} else if (type == DATA_VARCHAR || type == DATA_VARMYSQL
-						|| type == DATA_BINARY) {
-		/* Remove trailing spaces. */
-
-		/* Handle UCS2 strings differently. */
-		ulint	mbminlen	= dtype_get_mbminlen(
-						dfield_get_type(dfield));
-		ptr = row_mysql_read_var_ref(&col_len, mysql_data);
-		if (mbminlen == 2) {
-			/* space=0x0020 */
-			/* Trim "half-chars", just in case. */
-			col_len &= ~1;
-
-			while (col_len >= 2 && ptr[col_len - 2] == 0x00
-					&& ptr[col_len - 1] == 0x20) {
-				col_len -= 2;
-			}
-		} else {
-			ut_a(mbminlen == 1);
-			/* space=0x20 */
-			while (col_len > 0 && ptr[col_len - 1] == 0x20) {
-				col_len--;
-			}
-		}
-	} else if (comp && type == DATA_MYSQL
-			&& dtype_get_mbminlen(dfield_get_type(dfield)) == 1
-			&& dtype_get_mbmaxlen(dfield_get_type(dfield)) > 1) {
-		/* We assume that this CHAR field is encoded in a
-		variable-length character set where spaces have
-		1:1 correspondence to 0x20 bytes, such as UTF-8.
-
-		Consider a CHAR(n) field, a field of n characters.
-		It will contain between n*mbminlen and n*mbmaxlen bytes.
-		We will try to truncate it to n bytes by stripping
-		space padding.  If the field contains single-byte
-		characters only, it will be truncated to n characters.
-		Consider a CHAR(5) field containing the string ".a   "
-		where "." denotes a 3-byte character represented by
-		the bytes "$%&".  After our stripping, the string will
-		be stored as "$%&a " (5 bytes).  The string ".abc "
-		will be stored as "$%&abc" (6 bytes).
-
-		The space padding will be restored in row0sel.c, function
-		row_sel_field_store_in_mysql_format(). */
-
-		ulint		n_chars;
-		dtype_t*	dtype = dfield_get_type(dfield);
-
-		ut_a(!(dtype_get_len(dtype) % dtype_get_mbmaxlen(dtype)));
-		n_chars = dtype_get_len(dtype) / dtype_get_mbmaxlen(dtype);
-
-		/* Strip space padding. */
-		while (col_len > n_chars && ptr[col_len - 1] == 0x20) {
-			col_len--;
-		}
-	} else if (type == DATA_BLOB) {
-		ptr = row_mysql_read_blob_ref(&col_len, mysql_data, col_len);
-	}
-
-	dfield_set_data(dfield, ptr, col_len);
-}
diff --git a/innobase/row/row0ins.c b/innobase/row/row0ins.c
index fdbbe993ff0..303fe5749bc 100644
--- a/innobase/row/row0ins.c
+++ b/innobase/row/row0ins.c
@@ -521,6 +521,10 @@ row_ins_cascade_calc_update_vec(
 
 				fixed_size = dtype_get_fixed_size(type);
 
+				/* TODO: pad in UCS-2 with 0x0020.
+				TODO: How does the special truncation of
+				UTF-8 CHAR cols affect this? */
+
 				if (fixed_size
 				    && ufield->new_val.len != UNIV_SQL_NULL
 				    && ufield->new_val.len < fixed_size) {
diff --git a/innobase/row/row0mysql.c b/innobase/row/row0mysql.c
index 2b2b2d83002..b13ba056d85 100644
--- a/innobase/row/row0mysql.c
+++ b/innobase/row/row0mysql.c
@@ -106,20 +106,6 @@ row_mysql_delay_if_needed(void)
 }
 
 /***********************************************************************
-Reads a MySQL format variable-length field (like VARCHAR) length and
-returns pointer to the field data. */
-
-byte*
-row_mysql_read_var_ref_noninline(
-/*=============================*/
-			/* out: field + 2 */
-	ulint*	len,	/* out: variable-length field length */
-	byte*	field)	/* in: field */
-{
-	return(row_mysql_read_var_ref(len, field));
-}
-
-/***********************************************************************
 Frees the blob heap in prebuilt when no longer needed. */
 
 void
@@ -133,6 +119,61 @@ row_mysql_prebuilt_free_blob_heap(
 }
 
 /***********************************************************************
+Stores a >= 5.0.3 format true VARCHAR length to dest, in the MySQL row
+format. */
+
+byte*
+row_mysql_store_true_var_len(
+/*=========================*/
+			/* out: pointer to the data, we skip the 1 or 2 bytes
+			at the start that are used to store the len */
+	byte*	dest,	/* in: where to store */
+	ulint	len,	/* in: length, must fit in two bytes */
+	ulint	lenlen)	/* in: storage length of len: either 1 or 2 bytes */
+{
+	if (lenlen == 2) {
+		ut_a(len < 256 * 256);
+
+		mach_write_to_2_little_endian(dest, len);
+
+		return(dest + 2);
+	}
+
+	ut_a(lenlen == 1);
+	ut_a(len < 256);
+
+	mach_write_to_1(dest, len);
+
+	return(dest + 1);
+}
+
+/***********************************************************************
+Reads a >= 5.0.3 format true VARCHAR length, in the MySQL row format, and
+returns a pointer to the data. */
+
+byte*
+row_mysql_read_true_varchar(
+/*========================*/
+			/* out: pointer to the data, we skip the 1 or 2 bytes
+			at the start that are used to store the len */
+	ulint*	len,	/* out: variable-length field length */
+	byte*	field,	/* in: field in the MySQL format */
+	ulint	lenlen)	/* in: storage length of len: either 1 or 2 bytes */
+{
+	if (lenlen == 2) {
+		*len = mach_read_from_2_little_endian(field);
+
+		return(field + 2);
+	}
+
+	ut_a(lenlen == 1);
+
+	*len = mach_read_from_1(field);
+
+	return(field + 1);
+}
+
+/***********************************************************************
 Stores a reference to a BLOB in the MySQL format. */
 
 void
@@ -191,15 +232,177 @@ row_mysql_read_blob_ref(
 }
 
 /******************************************************************
-Convert a row in the MySQL format to a row in the Innobase format. */
+Stores a non-SQL-NULL field given in the MySQL format in the InnoDB format.
+The counterpart of this function is row_sel_field_store_in_mysql_format() in
+row0sel.c. */
+
+byte*
+row_mysql_store_col_in_innobase_format(
+/*===================================*/
+					/* out: up to which byte we used
+					buf in the conversion */
+	dfield_t*	dfield,		/* in/out: dfield where dtype
+					information must be already set when
+					this function is called! */
+	byte*		buf,		/* in/out: buffer for a converted
+					integer value; this must be at least
+					col_len long then! */
+	ibool		row_format_col,	/* TRUE if the mysql_data is from
+					a MySQL row, FALSE if from a MySQL
+					key value;
+					in MySQL, a true VARCHAR storage
+					format differs in a row and in a
+					key value: in a key value the length
+					is always stored in 2 bytes! */
+	byte*		mysql_data,	/* in: MySQL column value, not
+					SQL NULL; NOTE that dfield may also
+					get a pointer to mysql_data,
+					therefore do not discard this as long
+					as dfield is used! */
+	ulint		col_len,	/* in: MySQL column length; NOTE that
+					this is the storage length of the
+					column in the MySQL format row, not
+					necessarily the length of the actual
+					payload data; if the column is a true
+					VARCHAR then this is irrelevant */
+	ibool		comp)		/* in: TRUE = compact format */
+{
+	byte*		ptr 	= mysql_data;
+	dtype_t*	dtype;
+	ulint		type;
+	ulint		lenlen;
+
+	dtype = dfield_get_type(dfield);
+
+	type = dtype->mtype;
+
+	if (type == DATA_INT) {
+		/* Store integer data in Innobase in a big-endian format,
+		sign bit negated if the data is a signed integer. In MySQL,
+		integers are stored in a little-endian format. */
+
+		ptr = buf + col_len;
+
+		for (;;) {
+			ptr--;
+			*ptr = *mysql_data;
+			if (ptr == buf) {
+				break;
+			}
+			mysql_data++;
+		}
+
+		if (!(dtype->prtype & DATA_UNSIGNED)) {
+
+			*ptr = (byte) (*ptr ^ 128);
+		}
+
+		buf += col_len;
+	} else if ((type == DATA_VARCHAR
+		    || type == DATA_VARMYSQL
+		    || type == DATA_BINARY)) {
+
+		if (dtype_get_mysql_type(dtype) == DATA_MYSQL_TRUE_VARCHAR) {
+			/* The length of the actual data is stored to 1 or 2
+			bytes at the start of the field */
+			
+			if (row_format_col) {
+				if (dtype->prtype & DATA_LONG_TRUE_VARCHAR) {
+					lenlen = 2;
+				} else {
+					lenlen = 1;
+				}
+			} else {
+				/* In a MySQL key value, lenlen is always 2 */
+				lenlen = 2;
+			}
+
+			ptr = row_mysql_read_true_varchar(&col_len, mysql_data,
+								      lenlen);
+		} else {
+			/* Remove trailing spaces from old style VARCHAR
+			columns. */
+
+			/* Handle UCS2 strings differently. */
+			ulint	mbminlen	= dtype_get_mbminlen(dtype);
+
+			ptr = mysql_data;
+
+			if (mbminlen == 2) {
+				/* space=0x0020 */
+				/* Trim "half-chars", just in case. */
+				col_len &= ~1;
+
+				while (col_len >= 2 && ptr[col_len - 2] == 0x00
+						&& ptr[col_len - 1] == 0x20) {
+					col_len -= 2;
+				}
+			} else {
+				ut_a(mbminlen == 1);
+				/* space=0x20 */
+				while (col_len > 0
+						&& ptr[col_len - 1] == 0x20) {
+					col_len--;
+				}
+			}
+		}
+	} else if (comp && type == DATA_MYSQL
+			&& dtype_get_mbminlen(dtype) == 1
+			&& dtype_get_mbmaxlen(dtype) > 1) {
+		/* In some cases we strip trailing spaces from UTF-8 and other
+		multibyte charsets, from FIXED-length CHAR columns, to save
+		space. UTF-8 would otherwise normally use 3 * the string length
+		bytes to store a latin1 string! */
+
+		/* We assume that this CHAR field is encoded in a
+		variable-length character set where spaces have
+		1:1 correspondence to 0x20 bytes, such as UTF-8.
+
+		Consider a CHAR(n) field, a field of n characters.
+		It will contain between n * mbminlen and n * mbmaxlen bytes.
+		We will try to truncate it to n bytes by stripping
+		space padding.  If the field contains single-byte
+		characters only, it will be truncated to n characters.
+		Consider a CHAR(5) field containing the string ".a   "
+		where "." denotes a 3-byte character represented by
+		the bytes "$%&".  After our stripping, the string will
+		be stored as "$%&a " (5 bytes).  The string ".abc "
+		will be stored as "$%&abc" (6 bytes).
+
+		The space padding will be restored in row0sel.c, function
+		row_sel_field_store_in_mysql_format(). */
+
+		ulint		n_chars;
+
+		ut_a(!(dtype_get_len(dtype) % dtype_get_mbmaxlen(dtype)));
+
+		n_chars = dtype_get_len(dtype) / dtype_get_mbmaxlen(dtype);
+
+		/* Strip space padding. */
+		while (col_len > n_chars && ptr[col_len - 1] == 0x20) {
+			col_len--;
+		}
+	} else if (type == DATA_BLOB && row_format_col) {
+
+		ptr = row_mysql_read_blob_ref(&col_len, mysql_data, col_len);
+	}
+
+	dfield_set_data(dfield, ptr, col_len);
+
+	return(buf);
+}
+
+/******************************************************************
+Convert a row in the MySQL format to a row in the Innobase format. Note that
+the function to convert a MySQL format key value to an InnoDB dtuple is
+row_sel_convert_mysql_key_to_innobase() in row0sel.c. */
 static
 void
 row_mysql_convert_row_to_innobase(
 /*==============================*/
 	dtuple_t*	row,		/* in/out: Innobase row where the
 					field type information is already
-					copied there, or will be copied
-					later */
+					copied there! */
 	row_prebuilt_t*	prebuilt,	/* in: prebuilt struct where template
 					must be of type ROW_MYSQL_WHOLE_ROW */
 	byte*		mysql_rec)	/* in: row in the MySQL format;
@@ -236,10 +439,10 @@ row_mysql_convert_row_to_innobase(
 		row_mysql_store_col_in_innobase_format(dfield,
 					prebuilt->ins_upd_rec_buff
 						+ templ->mysql_col_offset,
+					TRUE, /* MySQL row format data */
 					mysql_rec + templ->mysql_col_offset,
 					templ->mysql_col_len,
-					templ->type, prebuilt->table->comp,
-					templ->is_unsigned);
+					prebuilt->table->comp);
 next_column:
 		;
 	} 
@@ -594,7 +797,8 @@ static
 dtuple_t*
 row_get_prebuilt_insert_row(
 /*========================*/
-					/* out: prebuilt dtuple */
+					/* out: prebuilt dtuple; the column
+					type information is also set in it */ 
 	row_prebuilt_t*	prebuilt)	/* in: prebuilt struct in MySQL
 					handle */
 {
@@ -784,6 +988,7 @@ row_unlock_tables_for_mysql(
 	lock_release_tables_off_kernel(trx);
 	mutex_exit(&kernel_mutex);
 }
+
 /*************************************************************************
 Sets a table lock on the table mentioned in prebuilt. */
 
@@ -962,10 +1167,13 @@ run_again:
 
 	if (err != DB_SUCCESS) {
 		que_thr_stop_for_mysql(thr);
-    thr->lock_state= QUE_THR_LOCK_ROW;
+
+/* TODO: what is this? */ thr->lock_state= QUE_THR_LOCK_ROW;
+
 		was_lock_wait = row_mysql_handle_errors(&err, trx, thr,
 								&savept);
-    thr->lock_state= QUE_THR_LOCK_NOLOCK;
+		thr->lock_state= QUE_THR_LOCK_NOLOCK;
+
 		if (was_lock_wait) {
 			goto run_again;
 		}
diff --git a/innobase/row/row0sel.c b/innobase/row/row0sel.c
index 54dfbe997ce..a09e09342e0 100644
--- a/innobase/row/row0sel.c
+++ b/innobase/row/row0sel.c
@@ -2119,10 +2119,10 @@ row_sel_convert_mysql_key_to_innobase(
 				   + 256 * key_ptr[data_offset + 1];
 			data_field_len = data_offset + 2 + field->prefix_len;
 			data_offset += 2;
-			
-			type = DATA_CHAR; /* now that we know the length, we
-					  store the column value like it would
-					  be a fixed char field */
+
+			/* now that we know the length, we store the column
+			value like it would be a fixed char field */
+
 		} else if (field->prefix_len > 0) {
 			/* Looks like MySQL pads unused end bytes in the
 			prefix with space. Therefore, also in UTF-8, it is ok
@@ -2146,11 +2146,12 @@ row_sel_convert_mysql_key_to_innobase(
 		
 		if (!is_null) {
 		        row_mysql_store_col_in_innobase_format(
-					dfield, buf, key_ptr + data_offset,
-					data_len, type,
-					index->table->comp,
-					dfield_get_type(dfield)->prtype
-							& DATA_UNSIGNED);
+					dfield,
+					buf,
+					FALSE, /* MySQL key value format col */
+					key_ptr + data_offset,
+					data_len,
+					index->table->comp);
 			buf += data_len;
 		}
 
@@ -2225,7 +2226,7 @@ row_sel_store_row_id_to_prebuilt(
 		dict_index_name_print(stderr, prebuilt->trx, index);
 		fprintf(stderr, "\n"
 "InnoDB: Field number %lu, record:\n",
-			(ulong) dict_index_get_sys_col_pos(index, DATA_ROW_ID));
+		    (ulong) dict_index_get_sys_col_pos(index, DATA_ROW_ID));
 		rec_print_new(stderr, index_rec, offsets);
 		putc('\n', stderr);
 		ut_error;
@@ -2235,8 +2236,9 @@ row_sel_store_row_id_to_prebuilt(
 }
 
 /******************************************************************
-Stores a non-SQL-NULL field in the MySQL format. */
-UNIV_INLINE
+Stores a non-SQL-NULL field in the MySQL format. The counterpart of this
+function is row_mysql_store_col_in_innobase_format() in row0mysql.c. */
+static
 void
 row_sel_field_store_in_mysql_format(
 /*================================*/
@@ -2251,6 +2253,8 @@ row_sel_field_store_in_mysql_format(
 	ulint	len)	/* in: length of the data */
 {
 	byte*	ptr;
+	byte*	field_end;
+	byte*	pad_ptr;
 
 	ut_ad(len != UNIV_SQL_NULL);
 
@@ -2274,25 +2278,66 @@ row_sel_field_store_in_mysql_format(
 		}
 
 		ut_ad(templ->mysql_col_len == len);
-	} else if (templ->type == DATA_VARCHAR || templ->type == DATA_VARMYSQL
-					|| templ->type == DATA_BINARY) {
-		/* Store the length of the data to the first two bytes of
-		dest; does not do anything yet because MySQL has
-		no real vars! */
+	} else if (templ->type == DATA_VARCHAR
+	           || templ->type == DATA_VARMYSQL
+		   || templ->type == DATA_BINARY) {
+
+		field_end = dest + templ->mysql_col_len;
+
+		if (templ->mysql_type == DATA_MYSQL_TRUE_VARCHAR) {
+			/* This is a >= 5.0.3 type true VARCHAR. Store the
+			length of the data to the first byte or the first
+			two bytes of dest. */
 		
-		dest = row_mysql_store_var_len(dest, len);
+			dest = row_mysql_store_true_var_len(dest, len,
+						templ->mysql_length_bytes);
+		}
+
+		/* Copy the actual data */
 		ut_memcpy(dest, data, len);
-#if 0
-		/* No real var implemented in MySQL yet! */
-		ut_ad(templ->mysql_col_len >= len + 2);
-#endif
 		
+		/* Pad with trailing spaces. We pad with spaces also the
+		unused end of a >= 5.0.3 true VARCHAR column, just in case
+		MySQL expects its contents to be deterministic. */
+			
+		pad_ptr = dest + len;
+
+		ut_ad(templ->mbminlen <= templ->mbmaxlen);
+
+		/* We handle UCS2 charset strings differently. */
+		if (templ->mbminlen == 2) {
+			/* A space char is two bytes, 0x0020 in UCS2 */
+
+			if (len & 1) {
+				/* A 0x20 has been stripped from the column.
+				Pad it back. */
+				
+				if (pad_ptr < field_end) {
+					*pad_ptr = 0x20;
+					pad_ptr++;
+				}
+			}
+			
+			/* Pad the rest of the string with 0x0020 */
+
+			while (pad_ptr < field_end) {
+				*pad_ptr = 0x00;
+				pad_ptr++;
+				*pad_ptr = 0x20;
+				pad_ptr++;
+			}
+		} else {
+			ut_ad(templ->mbminlen == 1);
+			/* space=0x20 */
+
+			memset(pad_ptr, 0x20, field_end - pad_ptr);
+		}
 	} else if (templ->type == DATA_BLOB) {
 		/* Store a pointer to the BLOB buffer to dest: the BLOB was
 		already copied to the buffer in row_sel_store_mysql_rec */
 
-		row_mysql_store_blob_ref(dest, templ->mysql_col_len,
-							data, len);
+		row_mysql_store_blob_ref(dest, templ->mysql_col_len, data,
+									len);
 	} else if (templ->type == DATA_MYSQL) {
 		memcpy(dest, data, len);
 
@@ -2306,9 +2351,10 @@ row_sel_field_store_in_mysql_format(
 		ut_a(len * templ->mbmaxlen >= templ->mysql_col_len);
 
 		if (templ->mbminlen != templ->mbmaxlen) {
-			/* Pad with spaces.  This undoes the stripping
+			/* Pad with spaces. This undoes the stripping
 			done in row0mysql.ic, function
 			row_mysql_store_col_in_innobase_format(). */
+
 			memset(dest + len, 0x20, templ->mysql_col_len - len);
 		}
 	} else {
@@ -2320,6 +2366,7 @@ row_sel_field_store_in_mysql_format(
 			|| templ->type == DATA_DOUBLE
 			|| templ->type == DATA_DECIMAL);
 		ut_ad(templ->mysql_col_len == len);
+
 		memcpy(dest, data, len);
 	}
 }
@@ -2436,40 +2483,6 @@ row_sel_store_mysql_rec(
 				mysql_rec + templ->mysql_col_offset,
 				templ, data, len);
 
-			if (templ->type == DATA_VARCHAR
-					|| templ->type == DATA_VARMYSQL
-					|| templ->type == DATA_BINARY) {
-				/* Pad with trailing spaces */
-				data = mysql_rec + templ->mysql_col_offset;
-
-				ut_ad(templ->mbminlen <= templ->mbmaxlen);
-				/* Handle UCS2 strings differently. */
-				if (templ->mbminlen == 2) {
-					/* space=0x0020 */
-					ulint	col_len = templ->mysql_col_len;
-
-					ut_a(!(col_len & 1));
-					if (len & 1) {
-						/* A 0x20 has been stripped
-						from the column.
-						Pad it back. */
-						goto pad_0x20;
-					}
-					/* Pad the rest of the string
-					with 0x0020 */
-					while (len < col_len) {
-						data[len++] = 0x00;
-					pad_0x20:
-						data[len++] = 0x20;
-					}
-				} else {
-					ut_ad(templ->mbminlen == 1);
-					/* space=0x20 */
-					memset(data + len, 0x20,
-						templ->mysql_col_len - len);
-				}
-			}
-
 			/* Cleanup */
 			if (extern_field_heap) {
  				mem_heap_free(extern_field_heap);
diff --git a/innobase/trx/trx0trx.c b/innobase/trx/trx0trx.c
index 614058e6860..643f7e164e5 100644
--- a/innobase/trx/trx0trx.c
+++ b/innobase/trx/trx0trx.c
@@ -1958,7 +1958,7 @@ trx_recover_for_mysql(
 
 	ut_print_timestamp(stderr);
 	fprintf(stderr,
-"  InnoDB: %d transactions in prepare state after recovery\n",
+"  InnoDB: %d transactions in prepared state after recovery\n",
 		count);
 
 	return (count);			
diff --git a/sql/ha_innodb.cc b/sql/ha_innodb.cc
index 73d5ac9e94e..7132ab00bb9 100644
--- a/sql/ha_innodb.cc
+++ b/sql/ha_innodb.cc
@@ -1074,6 +1074,8 @@ innobase_init(void)
 
   	DBUG_ENTER("innobase_init");
 
+	ut_a(DATA_MYSQL_TRUE_VARCHAR == (ulint)MYSQL_TYPE_VARCHAR);
+
   	os_innodb_umask = (ulint)my_umask;
 
 	/* First calculate the default path for innodb_data_home_dir etc.,
@@ -2244,7 +2246,9 @@ innobase_mysql_cmp(
 }
 
 /******************************************************************
-Converts a MySQL type to an InnoDB type. */
+Converts a MySQL type to an InnoDB type. Note that this function returns
+the 'mtype' of InnoDB. InnoDB differentiates between MySQL's old <= 4.1
+VARCHAR and the new true VARCHAR in >= 5.0.3 by the 'prtype'. */
 inline
 ulint
 get_innobase_type_from_mysql_type(
@@ -2259,8 +2263,9 @@ get_innobase_type_from_mysql_type(
 	switch (field->type()) {
 	        /* NOTE that we only allow string types in DATA_MYSQL
 		and DATA_VARMYSQL */
-                case MYSQL_TYPE_VAR_STRING:
-                case MYSQL_TYPE_VARCHAR: if (field->binary()) {
+                case MYSQL_TYPE_VAR_STRING: /* old <= 4.1 VARCHAR */
+                case MYSQL_TYPE_VARCHAR:    /* new >= 5.0.3 true VARCHAR */
+					if (field->binary()) {
 						return(DATA_BINARY);
 					} else if (strcmp(
 						  field->charset()->name,
@@ -2314,6 +2319,35 @@ get_innobase_type_from_mysql_type(
 }
 
 /***********************************************************************
+Writes an unsigned integer value < 64k to 2 bytes, in the little-endian
+storage format. */
+inline
+void
+innobase_write_to_2_little_endian(
+/*==============================*/
+	byte*	buf,	/* in: where to store */
+	ulint	val)	/* in: value to write, must be < 64k */
+{
+	ut_a(val < 256 * 256);
+
+	buf[0] = (byte)(val & 0xFF);
+	buf[1] = (byte)(val / 256);
+}
+
+/***********************************************************************
+Reads an unsigned integer value < 64k from 2 bytes, in the little-endian
+storage format. */
+inline
+uint
+innobase_read_from_2_little_endian(
+/*===============================*/
+			/* out: value */
+	const mysql_byte*	buf)	/* in: from where to read */
+{
+	return((ulint)(buf[0]) + 256 * ((ulint)(buf[1])));
+}
+
+/***********************************************************************
 Stores a key value for a row to a buffer. */
 
 uint
@@ -2352,9 +2386,14 @@ ha_innobase::store_key_val_for_row(
 	3. In a column prefix field, prefix_len next bytes are reserved for
 	data. In a normal field the max field length next bytes are reserved
 	for data. For a VARCHAR(n) the max field length is n. If the stored
-	value is the SQL NULL then these data bytes are set to 0. */	
+	value is the SQL NULL then these data bytes are set to 0.
 
-	/* We have to zero-fill the buffer so that MySQL is able to use a
+	4. We always use a 2 byte length for a true >= 5.0.3 VARCHAR. Note that
+	in the MySQL row format, the length is stored in 1 or 2 bytes,
+	depending on the maximum allowed length. But in the MySQL key value
+	format, the length always takes 2 bytes.
+
+	We have to zero-fill the buffer so that MySQL is able to use a
 	simple memcmp to compare two key values to determine if they are
 	equal. MySQL does this to compare contents of two 'ref' values. */
 
@@ -2377,7 +2416,43 @@ ha_innobase::store_key_val_for_row(
 		field = key_part->field;
 		mysql_type = field->type();
 
-		if (mysql_type == FIELD_TYPE_TINY_BLOB
+		if (mysql_type == MYSQL_TYPE_VARCHAR) {
+						/* >= 5.0.3 true VARCHAR */
+			ulint	lenlen;
+			ulint	len;
+			byte*	data;
+
+			if (is_null) {
+				buff += key_part->length + 2;
+				
+				continue;
+			}
+
+			lenlen = (ulint)
+				(((Field_varstring*)field)->length_bytes);
+
+			data = row_mysql_read_true_varchar(&len, 
+				(byte*) (record
+				+ (ulint)get_field_offset(table, field)),
+				lenlen);
+		
+			/* The length in a key value is always stored in 2
+			bytes */
+
+			row_mysql_store_true_var_len((byte*)buff, len, 2);
+			buff += 2;
+
+			memcpy(buff, data, len);
+
+			/* Note that we always reserve the maximum possible
+			length of the true VARCHAR in the key value, though
+			only len first bytes after the 2 length bytes contain
+			actual data. The rest of the space was reset to zero
+			in the bzero() call above. */
+
+			buff += key_part->length;
+
+		} else if (mysql_type == FIELD_TYPE_TINY_BLOB
 		    || mysql_type == FIELD_TYPE_MEDIUM_BLOB
 		    || mysql_type == FIELD_TYPE_BLOB
 		    || mysql_type == FIELD_TYPE_LONG_BLOB) {
@@ -2385,9 +2460,9 @@ ha_innobase::store_key_val_for_row(
 			ut_a(key_part->key_part_flag & HA_PART_KEY_SEG);
 
 		        if (is_null) {
-				 buff += key_part->length + 2;
+				buff += key_part->length + 2;
 				 
-				 continue;
+				continue;
 			}
 		    
 		        blob_data = row_mysql_read_blob_ref(&blob_len,
@@ -2404,12 +2479,15 @@ ha_innobase::store_key_val_for_row(
 			/* MySQL reserves 2 bytes for the length and the
 			storage of the number is little-endian */
 
-			ut_a(blob_len < 256);
-			*((byte*)buff) = (byte)blob_len;
+			innobase_write_to_2_little_endian(
+					(byte*)buff, (ulint)blob_len);
 			buff += 2;
 
 			memcpy(buff, blob_data, blob_len);
 
+			/* Note that we always reserve the maximum possible
+			length of the BLOB prefix in the key value. */
+
 			buff += key_part->length;
 		} else {
 		        if (is_null) {
@@ -2573,6 +2651,13 @@ build_template(
 
 		templ->mysql_col_len = (ulint) field->pack_length();
 		templ->type = get_innobase_type_from_mysql_type(field);
+		templ->mysql_type = (ulint)field->type();
+
+		if (templ->mysql_type == DATA_MYSQL_TRUE_VARCHAR) {
+			templ->mysql_length_bytes = (ulint)
+				    (((Field_varstring*)field)->length_bytes);
+		}
+	
 		templ->charset = dtype_get_charset_coll_noninline(
 				index->table->cols[i].type.prtype);
 		templ->mbminlen = index->table->cols[i].type.mbminlen;
@@ -2810,54 +2895,6 @@ func_exit:
   	DBUG_RETURN(error);
 }
 
-/******************************************************************
-Converts field data for storage in an InnoDB update vector. */
-inline
-mysql_byte*
-innobase_convert_and_store_changed_col(
-/*===================================*/
-				/* out: pointer to the end of the converted
-				data in the buffer */
-	upd_field_t*	ufield,	/* in/out: field in the update vector */
-	mysql_byte*	buf,	/* in: buffer we can use in conversion */
-	mysql_byte*	data,	/* in: column data to store */
-	ulint		len,	/* in: data len */
-	ulint		col_type,/* in: data type in InnoDB type numbers */
-	ulint		is_unsigned)/* in: != 0 if an unsigned integer type */
-{
-	uint	i;
-
-	if (len == UNIV_SQL_NULL) {
-		data = NULL;
-        } else if (col_type == DATA_VARCHAR || col_type == DATA_BINARY
-                   || col_type == DATA_VARMYSQL) {
-                /* Remove trailing spaces */
-                while (len > 0 && data[len - 1] == ' ') {
-                        len--;
-                }
-	} else if (col_type == DATA_INT) {
-		/* Store integer data in InnoDB in a big-endian
-		format, sign bit negated, if signed */
-
-		for (i = 0; i < len; i++) {
-			buf[len - 1 - i] = data[i];
-		}
-
-		if (!is_unsigned) {
-			buf[0] = buf[0] ^ 128;
-		}
-
-		data = buf;
-
-		buf += len;
-	}
-
-	ufield->new_val.data = data;
-	ufield->new_val.len = len;
-
-	return(buf);
-}
-
 /**************************************************************************
 Checks which fields have changed in a row and stores information
 of them to an update vector. */
@@ -2878,9 +2915,11 @@ calc_row_difference(
 {
 	mysql_byte*	original_upd_buff = upd_buff;
 	Field*		field;
+	enum_field_types field_mysql_type;
 	uint		n_fields;
 	ulint		o_len;
 	ulint		n_len;
+	ulint		col_pack_len;
 	byte*	        o_ptr;
         byte*	        n_ptr;
         byte*	        buf;
@@ -2888,6 +2927,7 @@ calc_row_difference(
 	ulint		col_type;
 	ulint		is_unsigned;
 	ulint		n_changed = 0;
+	dfield_t	dfield;
 	uint		i;
 
 	n_fields = table->s->fields;
@@ -2907,9 +2947,13 @@ calc_row_difference(
 
 		o_ptr = (byte*) old_row + get_field_offset(table, field);
 		n_ptr = (byte*) new_row + get_field_offset(table, field);
-		o_len = field->pack_length();
-		n_len = field->pack_length();
+		
+		col_pack_len = field->pack_length();
+		o_len = col_pack_len;
+		n_len = col_pack_len;
 
+		field_mysql_type = field->type();
+	
 		col_type = get_innobase_type_from_mysql_type(field);
 		is_unsigned = (ulint) (field->flags & UNSIGNED_FLAG);
 
@@ -2918,14 +2962,29 @@ calc_row_difference(
 		case DATA_BLOB:
 			o_ptr = row_mysql_read_blob_ref(&o_len, o_ptr, o_len);
 			n_ptr = row_mysql_read_blob_ref(&n_len, n_ptr, n_len);
+
 			break;
+
 		case DATA_VARCHAR:
 		case DATA_BINARY:
 		case DATA_VARMYSQL:
-			o_ptr = row_mysql_read_var_ref_noninline(&o_len,
-								o_ptr);
-			n_ptr = row_mysql_read_var_ref_noninline(&n_len,
-								n_ptr);
+			if (field_mysql_type == MYSQL_TYPE_VARCHAR) {
+				/* This is a >= 5.0.3 type true VARCHAR where
+				the real payload data length is stored in
+				1 or 2 bytes */
+			
+				o_ptr = row_mysql_read_true_varchar(
+						&o_len, o_ptr,
+				    (ulint)
+				    (((Field_varstring*)field)->length_bytes));
+								
+				n_ptr = row_mysql_read_true_varchar(
+						&n_len, n_ptr,
+				    (ulint)
+				    (((Field_varstring*)field)->length_bytes));
+			}
+
+			break;
 		default:
 			;
 		}
@@ -2947,12 +3006,29 @@ calc_row_difference(
 			/* The field has changed */
 
 			ufield = uvect->fields + n_changed;
+	
+			/* Let us use a dummy dfield to make the conversion
+			from the MySQL column format to the InnoDB format */
+
+			dfield.type = (prebuilt->table->cols + i)->type;
+
+			if (n_len != UNIV_SQL_NULL) {
+				buf = row_mysql_store_col_in_innobase_format(
+						&dfield,
+						(byte*)buf,
+						TRUE,
+						n_ptr,
+						col_pack_len,
+						prebuilt->table->comp);
+				ufield->new_val.data =
+						dfield_get_data(&dfield);
+				ufield->new_val.len =
+						dfield_get_len(&dfield);
+			} else {
+				ufield->new_val.data = NULL;
+				ufield->new_val.len = UNIV_SQL_NULL;
+			}
 
-			buf = (byte*)
-                          innobase_convert_and_store_changed_col(ufield,
-					  (mysql_byte*)buf,
-					  (mysql_byte*)n_ptr, n_len, col_type,
-						is_unsigned);
 			ufield->exp = NULL;
 			ufield->field_no =
 					(prebuilt->table->cols + i)->clust_pos;
@@ -3701,7 +3777,7 @@ ha_innobase::rnd_pos(
 	}
 
 	if (error) {
-	        DBUG_PRINT("error",("Got error: %ld",error));
+	        DBUG_PRINT("error", ("Got error: %ld", error));
 		DBUG_RETURN(error);
 	}
 
@@ -3709,10 +3785,11 @@ ha_innobase::rnd_pos(
         for the table, and it is == ref_length */
 
 	error = index_read(buf, pos, ref_length, HA_READ_KEY_EXACT);
-	if (error)
-	{
-	  DBUG_PRINT("error",("Got error: %ld",error));
+
+	if (error) {
+		DBUG_PRINT("error", ("Got error: %ld", error));
 	}
+
 	change_active_index(keynr);
 
   	DBUG_RETURN(error);
@@ -3752,12 +3829,11 @@ ha_innobase::position(
 							 ref_length, record);
 	}
 
-	/* Since we do not store len to the buffer 'ref', we must assume
-	that len is always fixed for this table. The following assertion
-	checks this. */
+	/* We assume that the 'ref' value len is always fixed for the same
+	table. */
   
 	if (len != ref_length) {
-	        fprintf(stderr,
+		fprintf(stderr,
 	 "InnoDB: Error: stored ref len is %lu, but table ref len is %lu\n",
 		  (ulong)len, (ulong)ref_length);
 	}
@@ -3788,9 +3864,11 @@ create_table_def(
 	ulint		n_cols;
   	int 		error;
   	ulint		col_type;
+	ulint		col_len;
   	ulint		nulls_allowed;
 	ulint		unsigned_type;
 	ulint		binary_type;
+	ulint		long_true_varchar;
 	ulint		charset_no;
   	ulint		i;
 
@@ -3837,17 +3915,40 @@ create_table_def(
 
 			charset_no = (ulint)field->charset()->number;
 
-			ut_a(charset_no < 256); /* in ut0type.h we assume that
-						the number fits in one byte */
+			ut_a(charset_no < 256); /* in data0type.h we assume
+						that the number fits in one
+						byte */
+		}
+
+		ut_a(field->type() < 256); /* we assume in dtype_form_prtype()
+					   that this fits in one byte */
+		col_len = field->pack_length();
+
+		/* The MySQL pack length contains 1 or 2 bytes length field
+		for a true VARCHAR. Let us subtract that, so that the InnoDB
+		column length in the InnoDB data dictionary is the real
+		maximum byte length of the actual data. */
+	
+		long_true_varchar = 0;
+
+		if (field->type() == MYSQL_TYPE_VARCHAR) {
+			col_len -= ((Field_varstring*)field)->length_bytes;
+
+			if (((Field_varstring*)field)->length_bytes == 2) {
+				long_true_varchar = DATA_LONG_TRUE_VARCHAR;
+			}
 		}
 
-		dict_mem_table_add_col(table, (char*) field->field_name,
-					col_type, dtype_form_prtype( 
-					(ulint)field->type()
-					| nulls_allowed | unsigned_type
-					| binary_type,
-					+ charset_no),
-					field->pack_length(), 0);
+		dict_mem_table_add_col(table,
+					(char*) field->field_name,
+					col_type,
+					dtype_form_prtype( 
+					    (ulint)field->type()
+					     | nulls_allowed | unsigned_type
+					     | binary_type | long_true_varchar,
+					    charset_no),
+					col_len,
+					0);
 	}
 
 	error = row_create_table_for_mysql(table, trx);
@@ -6125,54 +6226,79 @@ ha_innobase::get_auto_increment()
 	return((ulonglong) nr);
 }
 
+/***********************************************************************
+Compares two 'refs'. A 'ref' is the (internal) primary key value of the row.
+If there is no explicitly declared non-null unique key or a primary key, then
+InnoDB internally uses the row id as the primary key. */
 
 int
 ha_innobase::cmp_ref(
-	const mysql_byte *ref1,
-	const mysql_byte *ref2)
+/*=================*/
+				/* out: < 0 if ref1 < ref2, 0 if equal, else
+				> 0 */
+	const mysql_byte* ref1,	/* in: an (internal) primary key value in the
+				MySQL key value format */
+	const mysql_byte* ref2)	/* in: an (internal) primary key value in the
+				MySQL key value format */
 {
-	row_prebuilt_t*	 prebuilt = (row_prebuilt_t*) innobase_prebuilt;
+	row_prebuilt_t*	prebuilt = (row_prebuilt_t*) innobase_prebuilt;
 	enum_field_types mysql_type;
-	Field*		 field;
-	int result;
-
-	if (prebuilt->clust_index_was_generated)
-		return memcmp(ref1, ref2, DATA_ROW_ID_LEN);
-
-	/* Do type-aware comparison of Primary Key members. PK members
-	are always NOT NULL, so no checks for NULL are performed */
-	KEY_PART_INFO *key_part=
-			table->key_info[table->s->primary_key].key_part;
-	KEY_PART_INFO *key_part_end= 
-	  key_part + table->key_info[table->s->primary_key].key_parts;
+	Field*		field;
+	KEY_PART_INFO*	key_part;
+	KEY_PART_INFO*	key_part_end;
+	uint		len1;
+	uint		len2;
+	int 		result;
+
+	if (prebuilt->clust_index_was_generated) {
+		/* The 'ref' is an InnoDB row id */
+
+		return(memcmp(ref1, ref2, DATA_ROW_ID_LEN));
+	}
+
+	/* Do a type-aware comparison of primary key fields. PK fields
+	are always NOT NULL, so no checks for NULL are performed. */
+
+	key_part = table->key_info[table->s->primary_key].key_part;
+
+	key_part_end = key_part
+			+ table->key_info[table->s->primary_key].key_parts;
+
 	for (; key_part != key_part_end; ++key_part) {
 		field = key_part->field;
 		mysql_type = field->type();
+
 		if (mysql_type == FIELD_TYPE_TINY_BLOB
 		    || mysql_type == FIELD_TYPE_MEDIUM_BLOB
 		    || mysql_type == FIELD_TYPE_BLOB
 		    || mysql_type == FIELD_TYPE_LONG_BLOB) {
 		    
-			ut_a(!ref1[1]);
-			ut_a(!ref2[1]);
-			byte len1= *ref1;
-			byte len2= *ref2;
+			/* In the MySQL key value format, a column prefix of
+			a BLOB is preceded by a 2-byte length field */
+
+			len1 = innobase_read_from_2_little_endian(ref1);
+			len2 = innobase_read_from_2_little_endian(ref2);
+
 			ref1 += 2;
 			ref2 += 2;
-			result =
-			  ((Field_blob*)field)->cmp((const char*)ref1, len1,
+			result = ((Field_blob*)field)->cmp(
+						    (const char*)ref1, len1,
 			                            (const char*)ref2, len2);
 		} else {
-			result = 
-			  field->cmp((const char*)ref1, (const char*)ref2);
+			result = field->cmp((const char*)ref1,
+					    (const char*)ref2);
+		}
+
+		if (result) {
+
+			return(result);
 		}
 
-		if (result)
-			return result;
 		ref1 += key_part->length;
 		ref2 += key_part->length;
 	}
-	return 0;
+
+	return(0);
 }
 
 char*
diff --git a/sql/ha_innodb.h b/sql/ha_innodb.h
index 1c8063b9373..e1ed3a486cf 100644
--- a/sql/ha_innodb.h
+++ b/sql/ha_innodb.h
@@ -1,4 +1,4 @@
-/* Copyright (C) 2000 MySQL AB && Innobase Oy
+/* Copyright (C) 2000-2005 MySQL AB && Innobase Oy
 
    This program is free software; you can redistribute it and/or modify
    it under the terms of the GNU General Public License as published by
@@ -40,9 +40,10 @@ my_bool innobase_query_caching_of_table_permitted(THD* thd, char* full_name,
 /* The class defining a handle to an Innodb table */
 class ha_innobase: public handler
 {
-	void*	innobase_prebuilt;	/* (row_prebuilt_t*) prebuilt
-					struct in Innodb, used to save
-					CPU */
+	void*		innobase_prebuilt;/* (row_prebuilt_t*) prebuilt
+					struct in InnoDB, used to save
+					CPU time with prebuilt data
+					structures*/
 	THD*		user_thd;	/* the thread handle of the user
 					currently using the handle; this is
 					set in external_lock function */
@@ -83,12 +84,12 @@ class ha_innobase: public handler
  public:
   	ha_innobase(TABLE *table): handler(table),
 	  int_table_flags(HA_REC_NOT_IN_SEQ |
-			  HA_NULL_IN_KEY | HA_FAST_KEY_READ |
+			  HA_NULL_IN_KEY |
+			  HA_FAST_KEY_READ |
 			  HA_CAN_INDEX_BLOBS |
 			  HA_CAN_SQL_HANDLER |
 			  HA_NOT_EXACT_COUNT |
 			  HA_PRIMARY_KEY_IN_READ_INDEX |
-                          HA_NO_VARCHAR |
 			  HA_TABLE_SCAN_ON_INDEX),
 	  last_dup_key((uint) -1),
 	  start_of_scan(0),
@@ -108,7 +109,10 @@ class ha_innobase: public handler
  	ulong table_flags() const { return int_table_flags; }
 	ulong index_flags(uint idx, uint part, bool all_parts) const
 	{
-	  return (HA_READ_NEXT | HA_READ_PREV | HA_READ_ORDER | HA_READ_RANGE |
+	  return (HA_READ_NEXT |
+		  HA_READ_PREV |
+		  HA_READ_ORDER |
+		  HA_READ_RANGE |
                   HA_KEYREAD_ONLY);
 	}
   	uint max_supported_keys()          const { return MAX_KEY; }
@@ -163,7 +167,8 @@ class ha_innobase: public handler
 	int start_stmt(THD *thd);
 
   	void position(byte *record);
-  	ha_rows records_in_range(uint inx, key_range *min_key, key_range *max_key);
+  	ha_rows records_in_range(uint inx, key_range *min_key, key_range
+								*max_key);
 	ha_rows estimate_rows_upper_bound();
 
   	int create(const char *name, register TABLE *form,