Bug#15874001 CREATE INDEX ON A UTF8 CHAR COLUMN FAILS WITH ROW_FORMAT=REDUNDANT

CHAR(n) in ROW_FORMAT=REDUNDANT tables is always fixed-length
(n*mbmaxlen bytes), but in the temporary file it is variable-length
(n*mbminlen to n*mbmaxlen bytes) for variable-length character sets,
such as UTF-8.

The temporary file format used during index creation and online ALTER
TABLE is based on ROW_FORMAT=COMPACT. Thus, it should use the
variable-length encoding even if the base table is in
ROW_FORMAT=REDUNDNAT.

dtype_get_fixed_size_low(): Replace an assertion-like check with a
debug assertion.

rec_init_offsets_comp_ordinary(), rec_convert_dtuple_to_rec_comp():
Make this an inline function.  Replace 'ulint extra' with 'bool temp'.

rec_get_converted_size_comp_prefix_low(): Renamed from
rec_get_converted_size_comp_prefix(), and made inline. Add the
parameter 'bool temp'. If temp=true, do not add REC_N_NEW_EXTRA_BYTES.

rec_get_converted_size_comp_prefix(): Remove the comment about
dict_table_is_comp(). This function is only to be called for other
than ROW_FORMAT=REDUNDANT records.

rec_get_converted_size_temp(): New function for computing temporary
file record size. Omit REC_N_NEW_EXTRA_BYTES from the sizes.

rec_init_offsets_temp(), rec_convert_dtuple_to_temp(): New functions,
for operating on temporary file records.

rb:1559 approved by Jimmy Yang

storage/innodb_plugin/ChangeLog

+2012-11-15	The InnoDB Team
+
+	* include/data0type.ic, include/rem0rec.h,
+	rem/rem0rec.c, row/row0merge.c:
+	Fix Bug#15874001 CREATE INDEX ON A UTF8 CHAR COLUMN FAILS WITH
+	ROW_FORMAT=REDUNDANT
+
 2012-10-18	The InnoDB Team
 
 	* row/row0sel.c:

storage/innodb_plugin/include/data0type.ic

 /*****************************************************************************
 
-Copyright (c) 1996, 2010, Innobase Oy. All Rights Reserved.
+Copyright (c) 1996, 2012, Oracle and/or its affiliates. All Rights Reserved.
 
 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 the Free Software
@@ -439,34 +439,16 @@ dtype_get_fixed_size_low(
 		} else if (!comp) {
 			return(len);
 		} else {
-			/* We play it safe here and ask MySQL for
-			mbminlen and mbmaxlen.	Although
-			mbminlen and mbmaxlen are
-			initialized if and only if prtype
-			is (in one of the 3 functions in this file),
-			it could be that none of these functions
-			has been called. */
-
+#ifdef UNIV_DEBUG
 			ulint	i_mbminlen, i_mbmaxlen;
 
 			innobase_get_cset_width(
 				dtype_get_charset_coll(prtype),
 				&i_mbminlen, &i_mbmaxlen);
 
-			if (UNIV_UNLIKELY(mbminlen != i_mbminlen)
-			    || UNIV_UNLIKELY(mbmaxlen != i_mbmaxlen)) {
-
-				ut_print_timestamp(stderr);
-				fprintf(stderr, "  InnoDB: "
-					"mbminlen=%lu, "
-					"mbmaxlen=%lu, "
-					"type->mbminlen=%lu, "
-					"type->mbmaxlen=%lu\n",
-					(ulong) i_mbminlen,
-					(ulong) i_mbmaxlen,
-					(ulong) mbminlen,
-					(ulong) mbmaxlen);
-			}
+			ut_ad(mbminlen == i_mbminlen);
+			ut_ad(mbmaxlen == i_mbmaxlen);
+#endif /* UNIV_DEBUG */
 			if (mbminlen == mbmaxlen) {
 				return(len);
 			}

storage/innodb_plugin/include/rem0rec.h

@@ -361,24 +361,6 @@ rec_get_offsets_func(
 #define rec_get_offsets(rec,index,offsets,n,heap)	\
 	rec_get_offsets_func(rec,index,offsets,n,heap,__FILE__,__LINE__)
 
-/******************************************************//**
-Determine the offset to each field in a leaf-page record
-in ROW_FORMAT=COMPACT.  This is a special case of
-rec_init_offsets() and rec_get_offsets_func(). */
-UNIV_INTERN
-void
-rec_init_offsets_comp_ordinary(
-/*===========================*/
-	const rec_t*		rec,	/*!< in: physical record in
-					ROW_FORMAT=COMPACT */
-	ulint			extra,	/*!< in: number of bytes to reserve
-					between the record header and
-					the data payload
-					(usually REC_N_NEW_EXTRA_BYTES) */
-	const dict_index_t*	index,	/*!< in: record descriptor */
-	ulint*			offsets);/*!< in/out: array of offsets;
-					in: n=rec_offs_n_fields(offsets) */
-
 /******************************************************//**
 The following function determines the offsets to each field
 in the record.  It can reuse a previously allocated array. */
@@ -639,8 +621,48 @@ rec_copy(
 /*=====*/
 	void*		buf,	/*!< in: buffer */
 	const rec_t*	rec,	/*!< in: physical record */
-	const ulint*	offsets);/*!< in: array returned by rec_get_offsets() */
+	const ulint*	offsets)/*!< in: array returned by rec_get_offsets() */
+	__attribute__((nonnull));
 #ifndef UNIV_HOTBACKUP
+/**********************************************************//**
+Determines the size of a data tuple prefix in a temporary file.
+@return	total size */
+UNIV_INTERN
+ulint
+rec_get_converted_size_temp(
+/*========================*/
+	const dict_index_t*	index,	/*!< in: record descriptor */
+	const dfield_t*		fields,	/*!< in: array of data fields */
+	ulint			n_fields,/*!< in: number of data fields */
+	ulint*			extra)	/*!< out: extra size */
+	__attribute__((warn_unused_result, nonnull));
+
+/******************************************************//**
+Determine the offset to each field in temporary file.
+@see rec_convert_dtuple_to_temp() */
+UNIV_INTERN
+void
+rec_init_offsets_temp(
+/*==================*/
+	const rec_t*		rec,	/*!< in: temporary file record */
+	const dict_index_t*	index,	/*!< in: record descriptor */
+	ulint*			offsets)/*!< in/out: array of offsets;
+					in: n=rec_offs_n_fields(offsets) */
+	__attribute__((nonnull));
+
+/*********************************************************//**
+Builds a temporary file record out of a data tuple.
+@see rec_init_offsets_temp() */
+UNIV_INTERN
+void
+rec_convert_dtuple_to_temp(
+/*=======================*/
+	rec_t*			rec,		/*!< out: record */
+	const dict_index_t*	index,		/*!< in: record descriptor */
+	const dfield_t*		fields,		/*!< in: array of data fields */
+	ulint			n_fields)	/*!< in: number of fields */
+	__attribute__((nonnull));
+
 /**************************************************************//**
 Copies the first n fields of a physical record to a new physical record in
 a buffer.
@@ -675,21 +697,6 @@ rec_fold(
 	__attribute__((pure));
 #endif /* !UNIV_HOTBACKUP */
 /*********************************************************//**
-Builds a ROW_FORMAT=COMPACT record out of a data tuple. */
-UNIV_INTERN
-void
-rec_convert_dtuple_to_rec_comp(
-/*===========================*/
-	rec_t*			rec,	/*!< in: origin of record */
-	ulint			extra,	/*!< in: number of bytes to
-					reserve between the record
-					header and the data payload
-					(normally REC_N_NEW_EXTRA_BYTES) */
-	const dict_index_t*	index,	/*!< in: record descriptor */
-	ulint			status,	/*!< in: status bits of the record */
-	const dfield_t*		fields,	/*!< in: array of data fields */
-	ulint			n_fields);/*!< in: number of data fields */
-/*********************************************************//**
 Builds a physical record out of a data tuple and
 stores it into the given buffer.
 @return	pointer to the origin of physical record */
@@ -722,10 +729,7 @@ UNIV_INTERN
 ulint
 rec_get_converted_size_comp_prefix(
 /*===============================*/
-	const dict_index_t*	index,	/*!< in: record descriptor;
-					dict_table_is_comp() is
-					assumed to hold, even if
-					it does not */
+	const dict_index_t*	index,	/*!< in: record descriptor */
 	const dfield_t*		fields,	/*!< in: array of data fields */
 	ulint			n_fields,/*!< in: number of data fields */
 	ulint*			extra);	/*!< out: extra size */

storage/innodb_plugin/row/row0merge.c

@@ -291,6 +291,7 @@ row_merge_buf_add(
 		const dict_field_t*	ifield;
 		const dict_col_t*	col;
 		ulint			col_no;
+		ulint			fixed_len;
 		const dfield_t*		row_field;
 		ulint			len;
 
@@ -340,9 +341,21 @@ row_merge_buf_add(
 
 		ut_ad(len <= col->len || col->mtype == DATA_BLOB);
 
-		if (ifield->fixed_len) {
-			ut_ad(len == ifield->fixed_len);
+		fixed_len = ifield->fixed_len;
+		if (fixed_len && !dict_table_is_comp(index->table)
+		    && col->mbminlen != col->mbmaxlen) {
+			/* CHAR in ROW_FORMAT=REDUNDANT is always
+			fixed-length, but in the temporary file it is
+			variable-length for variable-length character
+			sets. */
+			fixed_len = 0;
+		}
+
+		if (fixed_len) {
+			ut_ad(len == fixed_len);
 			ut_ad(!dfield_is_ext(field));
+ 			ut_ad(!col->mbmaxlen || len >= col->mbminlen
+			      * (fixed_len / col->mbmaxlen));
 		} else if (dfield_is_ext(field)) {
 			extra_size += 2;
 		} else if (len < 128
@@ -363,12 +376,11 @@ row_merge_buf_add(
 		ulint	size;
 		ulint	extra;
 
-		size = rec_get_converted_size_comp(index,
-						   REC_STATUS_ORDINARY,
-						   entry, n_fields, &extra);
+		size = rec_get_converted_size_temp(
+			index, entry, n_fields, &extra);
 
-		ut_ad(data_size + extra_size + REC_N_NEW_EXTRA_BYTES == size);
-		ut_ad(extra_size + REC_N_NEW_EXTRA_BYTES == extra);
+		ut_ad(data_size + extra_size == size);
+		ut_ad(extra_size == extra);
 	}
 #endif /* UNIV_DEBUG */
 
@@ -572,14 +584,9 @@ row_merge_buf_write(
 		ulint		extra_size;
 		const dfield_t*	entry		= buf->tuples[i];
 
-		size = rec_get_converted_size_comp(index,
-						   REC_STATUS_ORDINARY,
-						   entry, n_fields,
-						   &extra_size);
+		size = rec_get_converted_size_temp(
+			index, entry, n_fields, &extra_size);
 		ut_ad(size >= extra_size);
-		ut_ad(extra_size >= REC_N_NEW_EXTRA_BYTES);
-		extra_size -= REC_N_NEW_EXTRA_BYTES;
-		size -= REC_N_NEW_EXTRA_BYTES;
 
 		/* Encode extra_size + 1 */
 		if (extra_size + 1 < 0x80) {
@@ -592,8 +599,7 @@ row_merge_buf_write(
 
 		ut_ad(b + size < block[1]);
 
-		rec_convert_dtuple_to_rec_comp(b + extra_size, 0, index,
-					       REC_STATUS_ORDINARY,
+		rec_convert_dtuple_to_temp(b + extra_size, index,
 					   entry, n_fields);
 
 		b += size;
@@ -841,7 +847,7 @@ err_exit:
 
 		*mrec = *buf + extra_size;
 
-		rec_init_offsets_comp_ordinary(*mrec, 0, index, offsets);
+		rec_init_offsets_temp(*mrec, index, offsets);
 
 		data_size = rec_offs_data_size(offsets);
 
@@ -860,7 +866,7 @@ err_exit:
 
 	*mrec = b + extra_size;
 
-	rec_init_offsets_comp_ordinary(*mrec, 0, index, offsets);
+	rec_init_offsets_temp(*mrec, index, offsets);
 
 	data_size = rec_offs_data_size(offsets);
 	ut_ad(extra_size + data_size < sizeof *buf);