Merge work:/home/bk/mysql-4.0

into mysql.sashanet.com:/home/sasha/src/bk/mysql-4.0


BitKeeper/etc/logging_ok:
  auto-union

BitKeeper/etc/logging_ok

@@ -4,6 +4,8 @@ monty@hundin.mysql.fi
 monty@work.mysql.com
 mwagner@evoq.mwagner.org
 paul@central.snake.net
+paul@teton.kitebird.com
 sasha@mysql.sashanet.com
-tonu@hundin.mysql.fi
+serg@serg.mysql.com
 tim@threads.polyesthetic.msg
+tonu@hundin.mysql.fi

Docs/Flags/latvia.eps

+%!PS-Adobe-2.0 EPSF-2.0
+%%Creator: pnmtops
+%%Title: latvia.ps
+%%Pages: 1
+%%BoundingBox: 295 365 317 396
+%%EndComments
+/readstring {
+  currentfile exch readhexstring pop
+} bind def
+/rpicstr 32 string def
+/gpicstr 32 string def
+/bpicstr 32 string def
+%%EndProlog
+%%Page: 1 1
+gsave
+295.44 365.64 translate
+21.12 30.72 scale
+0.5 0.5 translate  90 rotate  -0.5 -0.5 translate
+32 22 8
+[ 32 0 0 -22 0 22 ]
+{ rpicstr readstring }
+{ gpicstr readstring }
+{ bpicstr readstring }
+true 3
+colorimage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+grestore
+showpage
+%%Trailer

include/m_string.h

@@ -69,10 +69,6 @@
 # define memmove(d, s, n)	bmove((d), (s), (n)) /* our bmove */
 #endif
 
-#if defined(HAVE_STPCPY) && !defined(HAVE_mit_thread)
-#define strmov(A,B) stpcpy((A),(B))
-#endif
-
 /* Unixware 7 */
 #if !defined(HAVE_BFILL)
 # define bfill(A,B,C)           memset((A),(C),(B))
@@ -90,6 +86,13 @@
 extern "C" {
 #endif
 
+#if defined(HAVE_STPCPY) && !defined(HAVE_mit_thread)
+#define strmov(A,B) stpcpy((A),(B))
+#ifndef stpcpy
+extern char *stpcpy(char *, const char *);	/* For AIX with gcc 2.95.3 */
+#endif
+#endif
+
 extern char NEAR _dig_vec[];		/* Declared in int2str() */
 
 #ifdef BAD_STRING_COMPILER

innobase/btr/btr0btr.c

@@ -2238,12 +2238,93 @@ btr_check_node_ptr(
 	return(TRUE);
 }
 
+/****************************************************************
+Checks the size and number of fields in a record based on the definition of
+the index. */
+static
+ibool
+btr_index_rec_validate(
+/*====================*/
+				/* out: TRUE if ok */
+	rec_t*		rec,	/* in: index record */
+	dict_index_t*	index)	/* in: index */
+{
+	dtype_t* type;
+	byte*	data;
+	ulint	len;
+	ulint	n;
+	ulint	i;
+	
+	n = dict_index_get_n_fields(index);
+
+	if (rec_get_n_fields(rec) != n) {
+		fprintf(stderr, "Record has %lu fields, should have %lu\n",
+				rec_get_n_fields(rec), n);
+
+		return(FALSE);
+	}
+
+	for (i = 0; i < n; i++) {
+		data = rec_get_nth_field(rec, i, &len);
+
+		type = dict_index_get_nth_type(index, i);
+		
+		if (len != UNIV_SQL_NULL && dtype_is_fixed_size(type)
+		    && len != dtype_get_fixed_size(type)) {
+			fprintf(stderr,
+			"Record field %lu len is %lu, should be %lu\n",
+				i, len, dtype_get_fixed_size(type));
+
+			return(FALSE);
+		}
+	}
+
+	return(TRUE);			
+}
+
+/****************************************************************
+Checks the size and number of fields in records based on the definition of
+the index. */
+static
+ibool
+btr_index_page_validate(
+/*====================*/
+				/* out: TRUE if ok */
+	page_t*		page,	/* in: index page */
+	dict_index_t*	index)	/* in: index */
+{
+	rec_t*		rec;
+	page_cur_t 	cur;
+	ibool		ret	= TRUE;
+	
+	page_cur_set_before_first(page, &cur);
+	page_cur_move_to_next(&cur);
+
+	for (;;) {
+		rec = (&cur)->rec;
+
+		if (page_cur_is_after_last(&cur)) {
+			break;
+		}
+
+		if (!btr_index_rec_validate(rec, index)) {
+
+			ret = FALSE;
+		}
+
+		page_cur_move_to_next(&cur);
+	}
+
+	return(ret);	
+}
+
 /****************************************************************
 Validates index tree level. */
 static
-void
+ibool
 btr_validate_level(
 /*===============*/
+				/* out: TRUE if ok */
 	dict_tree_t*	tree,	/* in: index tree */
 	ulint		level)	/* in: level number */
 {
@@ -2260,6 +2341,8 @@ btr_validate_level(
 	page_cur_t	cursor;
 	mem_heap_t*	heap;
 	dtuple_t*	node_ptr_tuple;
+	ibool		ret	= TRUE;
+	dict_index_t*	index;
 	
 	mtr_start(&mtr);
 
@@ -2278,12 +2361,30 @@ btr_validate_level(
 		page = btr_node_ptr_get_child(node_ptr, &mtr);
 	}
 
+	index = UT_LIST_GET_FIRST(tree->tree_indexes);
+	
 	/* Now we are on the desired level */
 loop:
 	mtr_x_lock(dict_tree_get_lock(tree), &mtr);
 
-	/* Check ordering of records */
-	page_validate(page, UT_LIST_GET_FIRST(tree->tree_indexes));
+	/* Check ordering etc. of records */
+
+	if (!page_validate(page, index)) {
+		fprintf(stderr, "Error in page %lu in index %s\n",
+			buf_frame_get_page_no(page), index->name);
+
+		ret = FALSE;
+	}
+
+	if (level == 0) {
+		if (!btr_index_page_validate(page, index)) {
+ 			fprintf(stderr,
+			"Error in page %lu in index %s\n",
+				buf_frame_get_page_no(page), index->name);
+
+			ret = FALSE;
+		}
+	}
 	
 	ut_a(btr_page_get_level(page, &mtr) == level);
 
@@ -2374,14 +2475,17 @@ loop:
 
 		goto loop;
 	}
+
+	return(ret);
 }
 
 /******************************************************************
 Checks the consistency of an index tree. */
 
-void
+ibool
 btr_validate_tree(
 /*==============*/
+				/* out: TRUE if ok */
 	dict_tree_t*	tree)	/* in: tree */
 {
 	mtr_t	mtr;
@@ -2397,8 +2501,15 @@ btr_validate_tree(
 
 	for (i = 0; i <= n; i++) {
 		
-		btr_validate_level(tree, n - i);
+		if (!btr_validate_level(tree, n - i)) {
+
+			mtr_commit(&mtr);
+
+			return(FALSE);
+		}
 	}
 
 	mtr_commit(&mtr);
+
+	return(TRUE);
 }

innobase/btr/btr0cur.c

@@ -163,9 +163,14 @@ btr_cur_search_to_nth_level(
 				BTR_INSERT and BTR_ESTIMATE;
 				cursor->left_page is used to store a pointer
 				to the left neighbor page, in the cases
-				BTR_SEARCH_PREV and BTR_MODIFY_PREV */
+				BTR_SEARCH_PREV and BTR_MODIFY_PREV;
+				NOTE that if has_search_latch
+				is != 0, we maybe do not have a latch set
+				on the cursor page, we assume
+				the caller uses his search latch
+				to protect the record! */
 	btr_cur_t*	cursor, /* in/out: tree cursor; the cursor page is
-				s- or x-latched */
+				   s- or x-latched, but see also above! */
 	ulint		has_search_latch,/* in: info on the latch mode the
 				caller currently has on btr_search_latch:
 				RW_S_LATCH, or 0 */

innobase/btr/btr0sea.c

@@ -601,7 +601,12 @@ btr_search_guess_on_hash(
 	btr_search_t*	info,		/* in: index search info */
 	dtuple_t*	tuple,		/* in: logical record */
 	ulint		mode,		/* in: PAGE_CUR_L, ... */
-	ulint		latch_mode, 	/* in: BTR_SEARCH_LEAF, ... */
+	ulint		latch_mode, 	/* in: BTR_SEARCH_LEAF, ...;
+					NOTE that only if has_search_latch
+					is 0, we will have a latch set on
+					the cursor page, otherwise we assume
+					the caller uses his search latch
+					to protect the record! */
 	btr_cur_t*	cursor, 	/* out: tree cursor */
 	ulint		has_search_latch,/* in: latch mode the caller
 					currently has on btr_search_latch:
@@ -722,7 +727,9 @@ btr_search_guess_on_hash(
 	}
 	
 	if (!success) {
+		if (!has_search_latch) {
 		          btr_leaf_page_release(page, latch_mode, mtr);
+		}
 
 		goto failure;
 	}

innobase/include/btr0btr.h

@@ -376,9 +376,10 @@ btr_print_tree(
 /******************************************************************
 Checks the consistency of an index tree. */
 
-void
+ibool
 btr_validate_tree(
 /*==============*/
+				/* out: TRUE if ok */
 	dict_tree_t*	tree);	/* in: tree */
 
 #define BTR_N_LEAF_PAGES 	1

innobase/include/btr0cur.h

@@ -98,12 +98,18 @@ btr_cur_search_to_nth_level(
 				the previous page of the record! Inserts
 				should always be made using PAGE_CUR_LE to
 				search the position! */
-	ulint		latch_mode, /* in: BTR_SEARCH_LEAF, ...;
+	ulint		latch_mode, /* in: BTR_SEARCH_LEAF, ..., ORed with
+				BTR_INSERT and BTR_ESTIMATE;
 				cursor->left_page is used to store a pointer
 				to the left neighbor page, in the cases
-				BTR_SEARCH_PREV and BTR_MODIFY_PREV */
-	btr_cur_t*	cursor, /* out: tree cursor; the cursor page is s- or
-				x-latched */
+				BTR_SEARCH_PREV and BTR_MODIFY_PREV;
+				NOTE that if has_search_latch
+				is != 0, we maybe do not have a latch set
+				on the cursor page, we assume
+				the caller uses his search latch
+				to protect the record! */
+	btr_cur_t*	cursor, /* in/out: tree cursor; the cursor page is
+				s- or x-latched, but see also above! */
 	ulint		has_search_latch,/* in: latch mode the caller
 				currently has on btr_search_latch:
 				RW_S_LATCH, or 0 */

innobase/include/btr0pcur.h

@@ -87,7 +87,11 @@ btr_pcur_open_with_no_init(
 				PAGE_CUR_LE, not PAGE_CUR_GE, as the latter
 				may end up on the previous page of the
 				record! */
-	ulint		latch_mode,/* in: BTR_SEARCH_LEAF, ... */
+	ulint		latch_mode,/* in: BTR_SEARCH_LEAF, ...;
+				NOTE that if has_search_latch != 0 then
+				we maybe do not acquire a latch on the cursor
+				page, but assume that the caller uses his
+				btr search latch to protect the record! */
 	btr_pcur_t*	cursor, /* in: memory buffer for persistent cursor */
 	ulint		has_search_latch,/* in: latch mode the caller
 				currently has on btr_search_latch:

innobase/include/btr0pcur.ic

@@ -492,7 +492,11 @@ btr_pcur_open_with_no_init(
 				PAGE_CUR_LE, not PAGE_CUR_GE, as the latter
 				may end up on the previous page of the
 				record! */
-	ulint		latch_mode,/* in: BTR_SEARCH_LEAF, ... */
+	ulint		latch_mode,/* in: BTR_SEARCH_LEAF, ...;
+				NOTE that if has_search_latch != 0 then
+				we maybe do not acquire a latch on the cursor
+				page, but assume that the caller uses his
+				btr search latch to protect the record! */
 	btr_pcur_t*	cursor, /* in: memory buffer for persistent cursor */
 	ulint		has_search_latch,/* in: latch mode the caller
 				currently has on btr_search_latch:

innobase/include/row0mysql.h

@@ -229,6 +229,15 @@ row_rename_table_for_mysql(
 	char*	old_name,	/* in: old table name */
 	char*	new_name,	/* in: new table name */
 	trx_t*	trx);		/* in: transaction handle */
+/*************************************************************************
+Checks a table for corruption. */
+
+ulint
+row_check_table_for_mysql(
+/*======================*/
+					/* out: DB_ERROR or DB_SUCCESS */
+	row_prebuilt_t*	prebuilt);	/* in: prebuilt struct in MySQL
+					handle */
 
 /* A struct describing a place for an individual column in the MySQL
 row format which is presented to the table handler in ha_innobase.
@@ -281,7 +290,8 @@ struct row_prebuilt_struct {
 					is set to TRUE */
 	dict_index_t*	index;		/* current index for a search, if any */
 	ulint		template_type;	/* ROW_MYSQL_WHOLE_ROW, 
-					ROW_MYSQL_REC_FIELDS or
+					ROW_MYSQL_REC_FIELDS,
+					ROW_MYSQL_DUMMY_TEMPLATE, or
 					ROW_MYSQL_NO_TEMPLATE */
 	ulint		n_template;	/* number of elements in the
 					template */
@@ -359,6 +369,8 @@ struct row_prebuilt_struct {
 #define ROW_MYSQL_WHOLE_ROW	0
 #define ROW_MYSQL_REC_FIELDS	1
 #define ROW_MYSQL_NO_TEMPLATE	2
+#define ROW_MYSQL_DUMMY_TEMPLATE 3	/* dummy template used in
+					row_scan_and_check_index */
 
 #ifndef UNIV_NONINL
 #include "row0mysql.ic"

innobase/include/trx0trx.h

@@ -24,6 +24,13 @@ saving CPU time. The kernel mutex contention is increased, however. */
 
 extern ulint	trx_n_mysql_transactions;
 
+/************************************************************************
+Releases the search latch if trx has reserved it. */
+
+void
+trx_search_latch_release_if_reserved(
+/*=================================*/
+        trx_t*     trx); /* in: transaction */
 /********************************************************************
 Retrieves the error_info field from a trx. */
 
@@ -282,6 +289,13 @@ struct trx_struct{
 	ulint		n_mysql_tables_in_use; /* number of Innobase tables
 					used in the processing of the current
 					SQL statement in MySQL */
+        ulint           mysql_n_tables_locked;
+                                        /* how many tables the current SQL
+					statement uses, except those
+					in consistent read */
+        ibool           has_search_latch;
+			                /* TRUE if this trx has latched the
+			                search system latch in S-mode */
         ibool           ignore_duplicates_in_insert;
                                         /* in an insert roll back only insert
                                         of the latest row in case

innobase/page/page0page.c

@@ -1199,8 +1199,16 @@ page_rec_validate(
 	n_owned = rec_get_n_owned(rec);
 	heap_no = rec_get_heap_no(rec);
 
-	ut_a(n_owned <= PAGE_DIR_SLOT_MAX_N_OWNED);
-	ut_a(heap_no < page_header_get_field(page, PAGE_N_HEAP));
+	if (!(n_owned <= PAGE_DIR_SLOT_MAX_N_OWNED)) {
+		fprintf(stderr, "Dir slot n owned too big %lu\n", n_owned);
+		return(FALSE);
+	}
+
+	if (!(heap_no < page_header_get_field(page, PAGE_N_HEAP))) {
+		fprintf(stderr, "Heap no too big %lu %lu\n", heap_no,
+				page_header_get_field(page, PAGE_N_HEAP));
+		return(FALSE);
+	}
 	
 	return(TRUE);
 }
@@ -1216,19 +1224,20 @@ page_validate(
 	dict_index_t*	index)	/* in: data dictionary index containing
 				the page record type definition */
 {
+	page_dir_slot_t* slot;
 	mem_heap_t*	heap;
+	page_cur_t 	cur;
 	byte*		buf;
 	ulint		i;
 	ulint		count;
 	ulint		own_count;
 	ulint		slot_no;
 	ulint		data_size;
-	page_cur_t 	cur;
 	rec_t*		rec;
 	rec_t*		old_rec	= NULL;
-	page_dir_slot_t* slot;
 	ulint		offs;
 	ulint		n_slots;
+	ibool		ret	= FALSE;
 	
 	heap = mem_heap_create(UNIV_PAGE_SIZE);
 	
@@ -1244,9 +1253,17 @@ page_validate(
 	overlap. */
 
 	n_slots = page_dir_get_n_slots(page);
-	ut_ad(page_header_get_ptr(page, PAGE_HEAP_TOP) <=
+
+	if (!(page_header_get_ptr(page, PAGE_HEAP_TOP) <=
+			page_dir_get_nth_slot(page, n_slots - 1))) {
+		fprintf(stderr,
+       	"Record heap and dir overlap on a page in index %s, %lu, %lu\n",
+       		index->name, page_header_get_ptr(page, PAGE_HEAP_TOP),
        		page_dir_get_nth_slot(page, n_slots - 1));
 
+       		goto func_exit;
+       	}
+
 	/* Validate the record list in a loop checking also that
 	it is consistent with the directory. */
 	count = 0;
@@ -1259,11 +1276,20 @@ page_validate(
 
 	for (;;) {
 		rec = (&cur)->rec;
-		page_rec_validate(rec);
+
+		if (!page_rec_validate(rec)) {
+			goto func_exit;
+		}
 		
 		/* Check that the records are in the ascending order */
 		if ((count >= 2) && (!page_cur_is_after_last(&cur))) {
-			ut_a(1 == cmp_rec_rec(rec, old_rec, index));
+			if (!(1 == cmp_rec_rec(rec, old_rec, index))) {
+				fprintf(stderr,
+				"Records in wrong order in index %s\n",
+				index->name);
+
+				goto func_exit;
+			}
 		}
 
 		if ((rec != page_get_supremum_rec(page))
@@ -1275,16 +1301,38 @@ page_validate(
 		offs = rec_get_start(rec) - page;
 		
 		for (i = 0; i < rec_get_size(rec); i++) {
-			ut_a(buf[offs + i] == 0); /* No other record may
-						  overlap this */
+			if (!buf[offs + i] == 0) {
+				/* No other record may overlap this */
+
+				fprintf(stderr,
+				"Record overlaps another in index %s \n",
+				index->name);
+
+				goto func_exit;
+			}
+				
 			buf[offs + i] = 1;
 		}
 		
 		if (rec_get_n_owned(rec) != 0) {
 			/* This is a record pointed to by a dir slot */
-			ut_a(rec_get_n_owned(rec) == own_count);
+			if (rec_get_n_owned(rec) != own_count) {
+				fprintf(stderr,
+				"Wrong owned count %lu, %lu, in index %s\n",
+				rec_get_n_owned(rec), own_count,
+				index->name);
+
+				goto func_exit;
+			}
+
+			if (page_dir_slot_get_rec(slot) != rec) {
+				fprintf(stderr,
+				"Dir slot does not point to right rec in %s\n",
+				index->name);
+
+				goto func_exit;
+			}
 			
-			ut_a(page_dir_slot_get_rec(slot) == rec);
 			page_dir_slot_check(slot);
 			
 			own_count = 0;
@@ -1298,44 +1346,88 @@ page_validate(
 			break;
 		}
 
+		if (rec_get_next_offs(rec) < FIL_PAGE_DATA
+				|| rec_get_next_offs(rec) >= UNIV_PAGE_SIZE) {
+			fprintf(stderr,
+			  "Next record offset wrong %lu in index %s\n",
+			  rec_get_next_offs(rec), index->name);
+
+			goto func_exit;
+		}
+
 		count++;		
 		page_cur_move_to_next(&cur);
 		own_count++;
 		old_rec = rec;
 	}
 	
-	ut_a(rec_get_n_owned(rec) != 0);
-	ut_a(slot_no == n_slots - 1);
-	ut_a(page_header_get_field(page, PAGE_N_RECS) + 2 == count + 1);
+	if (rec_get_n_owned(rec) == 0) {
+		fprintf(stderr, "n owned is zero in index %s\n", index->name);
+
+		goto func_exit;
+	}
+		
+	if (slot_no != n_slots - 1) {
+		fprintf(stderr, "n slots wrong %lu %lu in index %s\n",
+			slot_no, n_slots - 1, index->name);
+		goto func_exit;
+	}		
+
+	if (page_header_get_field(page, PAGE_N_RECS) + 2 != count + 1) {
+		fprintf(stderr, "n recs wrong %lu %lu in index %s\n",
+		page_header_get_field(page, PAGE_N_RECS) + 2,  count + 1,
+		index->name);
+
+		goto func_exit;
+	}
 
 	if (data_size != page_get_data_size(page)) {
-		printf("Summed data size %lu, returned by func %lu\n",
+		fprintf(stderr, "Summed data size %lu, returned by func %lu\n",
 			data_size, page_get_data_size(page));
-		ut_error;
+		goto func_exit;
 	}
 
 	/* Check then the free list */
 	rec = page_header_get_ptr(page, PAGE_FREE);
 
 	while (rec != NULL) {
-		page_rec_validate(rec);
+		if (!page_rec_validate(rec)) {
+
+			goto func_exit;
+		}
 		
 		count++;	
 		offs = rec_get_start(rec) - page;
 		
 		for (i = 0; i < rec_get_size(rec); i++) {
-			ut_a(buf[offs + i] == 0);
+
+			if (buf[offs + i] != 0) {
+				fprintf(stderr,
+	                "Record overlaps another in free list, index %s \n",
+				index->name);
+
+				goto func_exit;
+			}
+				
 			buf[offs + i] = 1;
 		}
 		
 		rec = page_rec_get_next(rec);
 	}
 	
-	ut_a(page_header_get_field(page, PAGE_N_HEAP) == count + 1);
+	if (page_header_get_field(page, PAGE_N_HEAP) != count + 1) {
+
+		fprintf(stderr, "N heap is wrong %lu %lu in index %s\n",
+		page_header_get_field(page, PAGE_N_HEAP), count + 1,
+		index->name);
+	}
 
+	ret = TRUE;	
+
+func_exit:
 	mem_heap_free(heap);
 
-	return(TRUE);			  
+	return(ret);			  
 }
 
 /*******************************************************************

innobase/rem/rem0cmp.c

@@ -177,7 +177,9 @@ cmp_whole_field(
 				(int)(type->prtype & ~DATA_NOT_NULL),
 				a, a_length, b, b_length));
 	default:
-		assert(0);
+	        fprintf(stderr,
+			"InnoDB: unknown type number %lu\n", data_type);
+	        ut_a(0);
 	}
 
 	return(0);

innobase/row/row0mysql.c

@@ -1129,3 +1129,146 @@ funct_exit:
 	
 	return((int) err);
 }
+
+/*************************************************************************
+Checks that the index contains entries in an ascending order, unique
+constraint is not broken, and calculates the number of index entries
+in the read view of the current transaction. */
+static
+ibool
+row_scan_and_check_index(
+/*=====================*/
+					/* out: TRUE if ok */
+	row_prebuilt_t*	prebuilt,	/* in: prebuilt struct in MySQL */
+	dict_index_t*	index,		/* in: index */
+	ulint*		n_rows)		/* out: number of entries seen in the
+					current consistent read */
+{
+	mem_heap_t*	heap;
+	dtuple_t*	prev_entry = NULL;
+	ulint		matched_fields;
+	ulint		matched_bytes;
+	byte*		buf;
+	ulint		ret;
+	rec_t*		rec;
+	ibool		is_ok	= TRUE;
+	int		cmp;
+	
+	*n_rows = 0;
+	
+	buf = mem_alloc(UNIV_PAGE_SIZE);
+	heap = mem_heap_create(100);
+	
+	/* Make a dummy template in prebuilt, which we will use
+	in scanning the index entries */
+
+	prebuilt->index = index;
+	prebuilt->sql_stat_start = TRUE;
+	prebuilt->template_type = ROW_MYSQL_DUMMY_TEMPLATE;
+	prebuilt->n_template = 0;
+	prebuilt->need_to_access_clustered = FALSE;
+
+ 	dtuple_set_n_fields(prebuilt->search_tuple, 0);
+
+	prebuilt->select_lock_type = LOCK_NONE;
+
+	ret = row_search_for_mysql(buf, PAGE_CUR_G, prebuilt, 0, 0);
+loop:
+	if (ret != DB_SUCCESS) {
+
+		mem_free(buf);
+		mem_heap_free(heap);
+
+		return(is_ok);
+	}
+
+	*n_rows = *n_rows + 1;
+	
+	/* row_search... returns the index record in buf, record origin offset
+	within buf stored in the first 4 bytes, because we have built a dummy
+	template */
+	
+	rec = buf + mach_read_from_4(buf);
+	
+	if (prev_entry != NULL) {
+		matched_fields = 0;
+		matched_bytes = 0;
+	
+		cmp = cmp_dtuple_rec_with_match(prev_entry, rec,
+						&matched_fields,
+						&matched_bytes);
+		if (cmp > 0) {
+			fprintf(stderr,
+			"Error: index records in a wrong order in index %s\n",
+			index->name);
+
+			is_ok = FALSE;
+		} else if ((index->type & DICT_UNIQUE)
+			   && matched_fields >=
+			   dict_index_get_n_ordering_defined_by_user(index)) {
+			fprintf(stderr,
+			"Error: duplicate key in index %s\n",
+			index->name);
+
+			is_ok = FALSE;			   	
+		}
+	}
+
+	mem_heap_empty(heap);
+	
+	prev_entry = row_rec_to_index_entry(ROW_COPY_DATA, index, rec, heap);
+
+	ret = row_search_for_mysql(buf, PAGE_CUR_G, prebuilt, 0, ROW_SEL_NEXT);	
+
+	goto loop;	
+}
+
+/*************************************************************************
+Checks a table for corruption. */
+
+ulint
+row_check_table_for_mysql(
+/*======================*/
+					/* out: DB_ERROR or DB_SUCCESS */
+	row_prebuilt_t*	prebuilt)	/* in: prebuilt struct in MySQL
+					handle */
+{
+	dict_table_t*	table	= prebuilt->table;
+	dict_index_t*	index;
+	ulint		n_rows;
+	ulint		n_rows_in_table;
+	ulint		ret 	= DB_SUCCESS;
+	
+	index = dict_table_get_first_index(table);
+
+	while (index != NULL) {
+      /*        fprintf(stderr, "Validating index %s\n", index->name); */
+	
+		if (!btr_validate_tree(index->tree)) {
+			ret = DB_ERROR;
+		} else {
+			if (!row_scan_and_check_index(prebuilt,
+							index, &n_rows)) {
+				ret = DB_ERROR;
+			}
+
+			/* fprintf(stderr, "%lu entries in index %s\n", n_rows,
+			  index->name); */
+
+			if (index == dict_table_get_first_index(table)) {
+				n_rows_in_table = n_rows;
+			} else if (n_rows != n_rows_in_table) {
+
+				ret = DB_ERROR;
+ 
+				fprintf(stderr,
+		"Error: index %s contains %lu entries, should be %lu\n",
+					index->name, n_rows, n_rows_in_table);
+			}
+		}
+
+		index = dict_table_get_next_index(index);
+	}
+
+	return(ret);
+}

innobase/row/row0sel.c

@@ -2341,6 +2341,65 @@ row_sel_push_cache_row_for_mysql(
 	prebuilt->n_fetch_cached++;
 }
 	
+/*************************************************************************
+Tries to do a shortcut to fetch a clustered index record with a unique key,
+using the hash index if possible (not always). We assume that the search
+mode is PAGE_CUR_GE, it is a consistent read, trx has already a read view,
+btr search latch has been locked in S-mode. */
+static
+ulint
+row_sel_try_search_shortcut_for_mysql(
+/*==================================*/
+				/* out: SEL_FOUND, SEL_EXHAUSTED, SEL_RETRY */
+	rec_t**		out_rec,/* out: record if found */
+	row_prebuilt_t*	prebuilt,/* in: prebuilt struct */
+	mtr_t*		mtr)	/* in: started mtr */
+{
+	dict_index_t*	index		= prebuilt->index;
+	dtuple_t*	search_tuple	= prebuilt->search_tuple;
+	btr_pcur_t*	pcur		= prebuilt->pcur;
+	trx_t*		trx		= prebuilt->trx;
+	rec_t*		rec;
+	
+	ut_ad(index->type & DICT_CLUSTERED);
+
+	btr_pcur_open_with_no_init(index, search_tuple, PAGE_CUR_GE,
+					BTR_SEARCH_LEAF, pcur,
+					RW_S_LATCH, mtr);
+	rec = btr_pcur_get_rec(pcur);
+	
+	if (!page_rec_is_user_rec(rec)) {
+
+		return(SEL_RETRY);
+	}
+
+	/* As the cursor is now placed on a user record after a search with
+	the mode PAGE_CUR_GE, the up_match field in the cursor tells how many
+	fields in the user record matched to the search tuple */ 
+
+	if (btr_pcur_get_up_match(pcur) < dtuple_get_n_fields(search_tuple)) {
+
+		return(SEL_EXHAUSTED);
+	}
+
+	/* This is a non-locking consistent read: if necessary, fetch
+	a previous version of the record */
+			
+	if (!lock_clust_rec_cons_read_sees(rec, index, trx->read_view)) {
+
+		return(SEL_RETRY);
+	}
+
+	if (rec_get_deleted_flag(rec)) {
+
+		return(SEL_EXHAUSTED);
+	}
+
+	*out_rec = rec;
+	
+	return(SEL_FOUND);
+}
+
 /************************************************************************
 Searches for rows in the database. This is used in the interface to
 MySQL. This function opens a cursor, and also implements fetch next
@@ -2387,6 +2446,7 @@ row_search_for_mysql(
 	ibool		cons_read_requires_clust_rec;
 	ibool		was_lock_wait;
 	ulint		ret;
+	ulint		shortcut;
 	ibool		unique_search_from_clust_index	= FALSE;
 	ibool		mtr_has_extra_clust_latch 	= FALSE;
 	ibool		moves_up 			= FALSE;
@@ -2452,6 +2512,8 @@ row_search_for_mysql(
 		mode = pcur->search_mode;
 	}
 
+	mtr_start(&mtr);
+
 	if (match_mode == ROW_SEL_EXACT && index->type & DICT_UNIQUE
 		&& index->type & DICT_CLUSTERED
 		&& dtuple_get_n_fields(search_tuple)
@@ -2464,6 +2526,8 @@ row_search_for_mysql(
 			restore cursor position, and must return
  			immediately */
 
+ 			mtr_commit(&mtr);
+
 			return(DB_RECORD_NOT_FOUND);
 		}
 
@@ -2472,6 +2536,49 @@ row_search_for_mysql(
 		mode = PAGE_CUR_GE;
 
 		unique_search_from_clust_index = TRUE;
+
+		if (trx->mysql_n_tables_locked == 0
+					&& !prebuilt->sql_stat_start) {
+
+			/* This is a SELECT query done as a consistent read,
+			and the read view has already been allocated:
+			let us try a search shortcut through the hash
+			index */
+			
+			if (!trx->has_search_latch) {
+				rw_lock_s_lock(&btr_search_latch);
+				trx->has_search_latch = TRUE;
+
+			} else if (btr_search_latch.writer_is_wait_ex) {
+			        /* There is an x-latch request waiting:
+				release the s-latch for a moment to reduce
+				starvation */
+
+			        rw_lock_s_unlock(&btr_search_latch);
+			        rw_lock_s_lock(&btr_search_latch);
+			}
+
+			shortcut = row_sel_try_search_shortcut_for_mysql(&rec,
+							prebuilt, &mtr);
+			if (shortcut == SEL_FOUND) {
+				row_sel_store_mysql_rec(buf, prebuilt, rec);
+	
+ 				mtr_commit(&mtr);
+
+				return(DB_SUCCESS);
+			
+			} else if (shortcut == SEL_EXHAUSTED) {
+
+ 				mtr_commit(&mtr);
+
+				return(DB_RECORD_NOT_FOUND);
+			}
+		}
+	}
+	
+	if (trx->has_search_latch) {
+		rw_lock_s_unlock(&btr_search_latch);
+		trx->has_search_latch = FALSE;
 	}			
 
 	/* Note that if the search mode was GE or G, then the cursor
@@ -2486,8 +2593,6 @@ row_search_for_mysql(
 		moves_up = TRUE;
 	}
 
-	mtr_start(&mtr);
-
 	thr = que_fork_get_first_thr(prebuilt->sel_graph);
 
 	que_thr_move_to_run_state_for_mysql(thr, trx);
@@ -2711,7 +2816,9 @@ rec_loop:
 	if (prebuilt->n_rows_fetched >= MYSQL_FETCH_CACHE_THRESHOLD
 			&& !prebuilt->templ_contains_blob
 			&& prebuilt->select_lock_type == LOCK_NONE
-			&& !prebuilt->clust_index_was_generated) {
+			&& !prebuilt->clust_index_was_generated
+	                && prebuilt->template_type
+	                                 != ROW_MYSQL_DUMMY_TEMPLATE) {
 
 		/* Inside an update, for example, we do not cache rows,
 		since we may use the cursor position to do the actual
@@ -2725,8 +2832,14 @@ rec_loop:
 		}
 
 		goto next_rec;
+	} else {
+		if (prebuilt->template_type == ROW_MYSQL_DUMMY_TEMPLATE) {
+			ut_memcpy(buf + 4, rec - rec_get_extra_size(rec),
+						rec_get_size(rec));
+			mach_write_to_4(buf, rec_get_extra_size(rec) + 4);
 		} else {
 			row_sel_store_mysql_rec(buf, prebuilt, rec);
+		}
 
 		if (prebuilt->clust_index_was_generated) {
 			row_sel_store_row_id_to_prebuilt(prebuilt, index_rec,

innobase/trx/trx0trx.c

@@ -22,6 +22,7 @@ Created 3/26/1996 Heikki Tuuri
 #include "read0read.h"
 #include "srv0srv.h"
 #include "thr0loc.h"
+#include "btr0sea.h"
 
 /* Dummy session used currently in MySQL interface */
 sess_t*		trx_dummy_sess = NULL;
@@ -63,6 +64,7 @@ trx_create(
 	trx->dict_operation = FALSE;
 
 	trx->n_mysql_tables_in_use = 0;
+	trx->mysql_n_tables_locked = 0;
 
 	trx->ignore_duplicates_in_insert = FALSE;
 
@@ -96,6 +98,8 @@ trx_create(
 	trx->lock_heap = mem_heap_create_in_buffer(256);
 	UT_LIST_INIT(trx->trx_locks);
 
+	trx->has_search_latch = FALSE;
+
 	trx->read_view_heap = mem_heap_create(256);
 	trx->read_view = NULL;
 
@@ -132,6 +136,21 @@ trx_allocate_for_mysql(void)
 	return(trx);
 }
 
+/************************************************************************
+Releases the search latch if trx has reserved it. */
+
+void
+trx_search_latch_release_if_reserved(
+/*=================================*/
+        trx_t*     trx) /* in: transaction */
+{
+  if (trx->has_search_latch) {
+    rw_lock_s_unlock(&btr_search_latch);
+
+    trx->has_search_latch = FALSE;
+  }
+}
+
 /************************************************************************
 Frees a transaction object. */
 
@@ -149,6 +168,7 @@ trx_free(
 	ut_a(trx->update_undo == NULL); 
 
 	ut_a(trx->n_mysql_tables_in_use == 0);
+	ut_a(trx->mysql_n_tables_locked == 0);
 	
 	if (trx->undo_no_arr) {
 		trx_undo_arr_free(trx->undo_no_arr);
@@ -160,6 +180,8 @@ trx_free(
 	ut_a(trx->wait_lock == NULL);
 	ut_a(UT_LIST_GET_LEN(trx->wait_thrs) == 0);
 
+	ut_a(!trx->has_search_latch);
+
 	if (trx->lock_heap) {
 		mem_heap_free(trx->lock_heap);
 	}

mysql-test/r/innodb.result

@@ -144,7 +144,7 @@ test.t1	optimize	error	The handler for the table doesn't support check/repair
 a
 2
 Table	Op	Msg_type	Msg_text
-test.t1	check	error	The handler for the table doesn't support check/repair
+test.t1	check	status	OK
 a	b
 2	testing
 Table	Op	Msg_type	Msg_text

mysql-test/t/select.test

@@ -1609,7 +1609,7 @@ select t2.fld1,count(*) from t2,t3 where t2.fld1=158402 and t3.name=t2.fld3 grou
 #
 
 select sum(Period)/count(*) from t1;
-select companynr,count(price) as "count",sum(price) as "sum" ,sum(price)/count(price)-avg(price) as "diff",(0+count(price))*companynr as func from t3 group by companynr;
+select companynr,count(price) as "count",sum(price) as "sum" ,abs(sum(price)/count(price)-avg(price)) as "diff",(0+count(price))*companynr as func from t3 group by companynr;
 select companynr,sum(price)/count(price) as avg from t3 group by companynr having avg > 70000000 order by avg;
 
 #

mysys/getvar.c

@@ -101,7 +101,7 @@ my_bool set_changeable_var(my_string str,CHANGEABLE_VAR *vars)
       }
       if (num < (longlong) found->min_value)
 	num=(longlong) found->min_value;
-      else if (num > (longlong) (ulong) found->max_value)
+      else if (num > 0 && (ulonglong) num > (ulonglong) (ulong) found->max_value)
 	num=(longlong) (ulong) found->max_value;
       num=((num- (longlong) found->sub_size) / (ulonglong) found->block_size);
       (*found->varptr)= (long) (num*(ulonglong) found->block_size);

mysys/tree.c

@@ -266,6 +266,8 @@ int tree_delete(TREE *tree, void *key)
   }
   if (remove_colour == BLACK)
     rb_delete_fixup(tree,parent);
+  if (tree->free)
+    (*tree->free)(ELEMENT_KEY(tree,element));
   my_free((gptr) element,MYF(0));
   tree->elements_in_tree--;
   return 0;

sql-bench/Comments/postgres.benchmark

-# This file describes how to run MySQL benchmarks with Postgres
+# This file describes how to run MySQL benchmarks with PostgreSQL
 #
+# WARNING:
+#
+# Don't run the --fast test on a PostgreSQL 7.1.1 database on
+# which you have any critical data; During one of our test runs
+# PostgreSQL got a corrupted database and all data was destroyed!
+# (When we tried to restart postmaster, It died with a
+# 'no such file or directory' error and never recovered from that!
+#
+# WARNING
 
 # The test was run on a Intel Xeon 2x 550 Mzh machine with 1G memory,
 # 9G hard disk.  The OS is Suse 6.4, with Linux 2.2.14 compiled with SMP
@@ -8,49 +17,52 @@
 # on the same machine. No other cpu intensive process was used during
 # the benchmark.
 
-#
-#
-# First, install postgresql-7.0.2.tar.gz
-#
+# First, install postgresql-7.1.1.tar.gz
 
-#
-# Start by adding the following lines to your ~/.bash_profile or
+# Adding the following lines to your ~/.bash_profile or
 # corresponding file. If you are using csh, use ´setenv´.
 #
 
-export POSTGRES_INCLUDE=/usr/local/pgsql/include
-export POSTGRES_LIB=/usr/local/pgsql/lib
+export POSTGRES_INCLUDE=/usr/local/pg/include
+export POSTGRES_LIB=/usr/local/pg/lib
 
-PATH=$PATH:/usr/local/pgsql/bin
-MANPATH=$MANPATH:/usr/local/pgsql/man
+PATH=$PATH:/usr/local/pg/bin
+MANPATH=$MANPATH:/usr/local/pg/man
 
 #
 # Add the following line to /etc/ld.so.conf:
 #
 
-/usr/local/pgsql/lib
-and run ldconfig.
+/usr/local/pg/lib
 
-#
-# untar the postgres source distribution and cd to src/
-# run the following commands:
-#
+# and run:
+
+ldconfig
 
-./configure
+# untar the postgres source distribution,  cd to postgresql-*
+# and run the following commands:
+
+CFLAGS=-O3 ./configure
 gmake
 gmake install
 
-mkdir /usr/local/pgsql/data
-chown postgres /usr/local/pgsql/data
+mkdir /usr/local/pg/data
+chown postgres /usr/local/pg/data
 su - postgres
-/usr/local/pgsql/bin/initdb -D /usr/local/pgsql/data
-su postgres -c "/usr/local/pgsql/bin/postmaster -o -F -D /usr/local/pgsql/data" &
-su postgres -c "/usr/local/pgsql/bin/createdb test"
+/usr/local/pg/bin/initdb -D /usr/local/pg/data
+/usr/local/pg/bin/postmaster -o -F -D /usr/local/pg/data &
+/usr/local/pg/bin/createdb test
+exit
 
 #
-# Second, install packages DBD-Pg-0.95.tar.gz and DBI-1.14.tar.gz,
+# Second, install packages DBD-Pg-1.00.tar.gz and DBI-1.14.tar.gz,
 # available from http://www.perl.com/CPAN/
-#
+
+export POSTGRES_LIB=/usr/local/pg/lib/
+export POSTGRES_INCLUDE=/usr/local/pg/include/postgresql
+perl Makefile.PL
+make
+make install
 
 #
 # Now we run the test that can be found in the sql-bench directory in the
@@ -59,17 +71,20 @@ su postgres -c "/usr/local/pgsql/bin/createdb test"
 # We did run two tests:
 # The standard test
 
-run-all-tests --comment="Intel Xeon, 2x550 Mhz, 1G, pg started with -o -F" --user=postgres --server=pg --cmp=mysql
+run-all-tests --comment="Intel Xeon, 2x550 Mhz, 512M, pg started with -o -F" --user=postgres --server=pg --cmp=mysql
 
 # and a test where we do a vacuum() after each update.
 # (The time for vacuum() is counted in the book-keeping() column)
 
-run-all-tests --comment="Intel Xeon, 2x550 Mhz, 1G, pg started with -o -F" --user=postgres --server=pg --cmp=mysql --fast
+run-all-tests --comment="Intel Xeon, 2x550 Mhz, 512M, pg started with -o -F" --user=postgres --server=pg --cmp=mysql --fast
 
 # If you want to store the results in a output/RUN-xxx file, you should
 # repeate the benchmark with the extra option --log --use-old-result
 # This will create a the RUN file based of the previous results
-#
 
-run-all-tests --comment="Intel Xeon, 2x550 Mhz, 1G, pg started with -o -F" --user=postgres --server=pg --cmp=mysql --log --use-old-result
-run-all-tests --comment="Intel Xeon, 2x550 Mhz, 1G, pg started with -o -F" --user=postgres --server=pg --cmp=mysql --fast --log --use-old-result
+run-all-tests --comment="Intel Xeon, 2x550 Mhz, 512M, pg started with -o -F" --user=postgres --server=pg --cmp=mysql --log --use-old-result
+run-all-tests --comment="Intel Xeon, 2x550 Mhz, 512MG, pg started with -o -F" --user=postgres --server=pg --cmp=mysql --fast --log --use-old-result
+
+# Between running the different tests we dropped and recreated the PostgreSQL
+# database to ensure that PostgreSQL should get a clean start,
+# independent of the previous runs.

sql-bench/Results/ATIS-pg-Linux_2.4.0_64GB_SMP_i686-cmp-mysql,pg

+Testing server 'PostgreSQL version ???' at 2001-06-03  4:40:22
+
+ATIS table test
+
+Creating tables
+Time for create_table (28):  0 wallclock secs ( 0.01 usr  0.00 sys +  0.00 cusr  0.00 csys =  0.01 CPU)
+
+Inserting data
+Time to insert (9768):  8 wallclock secs ( 2.97 usr  0.28 sys +  0.00 cusr  0.00 csys =  3.25 CPU)
+
+Retrieving data
+Time for select_simple_join (500):  3 wallclock secs ( 0.74 usr  0.03 sys +  0.00 cusr  0.00 csys =  0.77 CPU)
+Time for select_join (100):  4 wallclock secs ( 0.52 usr  0.13 sys +  0.00 cusr  0.00 csys =  0.65 CPU)
+Time for select_key_prefix_join (100): 11 wallclock secs ( 4.30 usr  0.16 sys +  0.00 cusr  0.00 csys =  4.46 CPU)
+Time for select_distinct (800): 22 wallclock secs ( 1.95 usr  0.18 sys +  0.00 cusr  0.00 csys =  2.13 CPU)
+Time for select_group (2600): 52 wallclock secs ( 1.43 usr  0.19 sys +  0.00 cusr  0.00 csys =  1.62 CPU)
+
+Removing tables
+Time to drop_table (28):  1 wallclock secs ( 0.01 usr  0.00 sys +  0.00 cusr  0.00 csys =  0.01 CPU)
+Total time: 101 wallclock secs (11.93 usr  0.98 sys +  0.00 cusr  0.00 csys = 12.91 CPU)

sql-bench/Results/RUN-pg-Linux_2.4.0_64GB_SMP_i686-cmp-mysql,pg

+Benchmark DBD suite: 2.13
+Date of test:        2001-06-03 19:30:53
+Running tests on:    Linux 2.4.0-64GB-SMP i686
+Arguments:           
+Comments:            Intel Xeon, 2x550 Mhz, 512M, pg started with -o -F
+Limits from:         mysql,pg
+Server version:      PostgreSQL version 7.1.1
+
+ATIS: Total time: 101 wallclock secs (11.93 usr  0.98 sys +  0.00 cusr  0.00 csys = 12.91 CPU)
+alter-table: Total time: 52 wallclock secs ( 0.49 usr  0.14 sys +  0.00 cusr  0.00 csys =  0.63 CPU)
+big-tables: Total time: 1324 wallclock secs ( 8.10 usr  0.69 sys +  0.00 cusr  0.00 csys =  8.79 CPU)
+connect: Total time: 555 wallclock secs (51.66 usr 14.01 sys +  0.00 cusr  0.00 csys = 65.67 CPU)
+create: Total time: 10008 wallclock secs (27.97 usr  5.83 sys +  0.00 cusr  0.00 csys = 33.80 CPU)
+insert: Estimated total time: 199360 wallclock secs (879.85 usr 202.59 sys +  0.00 cusr  0.00 csys = 1082.45 CPU)
+select: Estimated total time: 13197 wallclock secs (68.30 usr  8.18 sys +  0.00 cusr  0.00 csys = 76.48 CPU)
+wisconsin: Total time: 52 wallclock secs (12.40 usr  2.23 sys +  0.00 cusr  0.00 csys = 14.63 CPU)
+
+All 8 test executed successfully
+Tests with estimated time have a + at end of line
+
+Totals per operation:
+Operation             seconds     usr     sys     cpu   tests
+alter_table_add                       49.00    0.26    0.06    0.32     992 
+connect                              143.00    8.01    1.89    9.90   10000 
+connect+select_1_row                 195.00   10.94    2.31   13.25   10000 
+connect+select_simple                157.00   10.42    2.41   12.83   10000 
+count                                131.00    0.03    0.00    0.03     100 
+count_distinct                       132.00    0.31    0.06    0.37    1000 
+count_distinct_2                     213.00    0.37    0.03    0.40    1000 
+count_distinct_big                   266.00    7.91    0.25    8.16     120 
+count_distinct_group                 384.00    1.07    0.08    1.15    1000 
+count_distinct_group_on_key          488.00    0.41    0.03    0.44    1000 
+count_distinct_group_on_key_parts    383.00    1.10    0.07    1.17    1000 
+count_distinct_key_prefix            179.00    0.28    0.07    0.35    1000 
+count_group_on_key_parts             331.00    1.13    0.06    1.19    1000 
+count_on_key                        1850.00   15.78    1.99   17.77   50100 +
+create+drop                         3280.00   10.74    1.89   12.63   10000 
+create_MANY_tables                   160.00    3.67    1.35    5.02    5000 
+create_index                           1.00    0.00    0.00    0.00       8 
+create_key+drop                     5781.00   10.70    1.53   12.23   10000 
+create_table                           1.00    0.01    0.00    0.01      31 
+delete_all                          2478.00    0.01    0.00    0.01      12 
+delete_all_many_keys                  94.00    0.05    0.00    0.05       1 
+delete_big                             0.00    0.01    0.00    0.01       1 
+delete_big_many_keys                  93.00    0.05    0.00    0.05     128 
+delete_key                            85.00    3.19    0.48    3.67   10000 
+drop_index                             0.00    0.01    0.00    0.01       8 
+drop_table                             1.00    0.01    0.00    0.01      28 
+drop_table_when_MANY_tables          772.00    1.29    0.28    1.57    5000 
+insert                               353.00  104.09   24.32  128.41  350768 
+insert_duplicates                    120.00   30.53   10.61   41.14  100000 
+insert_key                           804.00   47.08   47.06   94.14  100000 
+insert_many_fields                   528.00    1.12    0.21    1.33    2000 
+insert_select_1_key                   86.00    0.00    0.00    0.00       1 
+insert_select_2_keys                 196.00    0.00    0.00    0.00       1 
+min_max                               60.00    0.02    0.00    0.02      60 
+min_max_on_key                      10543.00   25.38    4.37   29.75   85000 ++
+order_by_big                         148.00   21.11    0.72   21.83      10 
+order_by_big_key                     145.00   24.01    1.27   25.28      10 
+order_by_big_key2                    132.00   21.28    0.64   21.92      10 
+order_by_big_key_desc                145.00   23.93    1.27   25.20      10 
+order_by_big_key_diff                138.00   21.30    0.56   21.86      10 
+order_by_big_key_prefix              133.00   21.16    0.80   21.96      10 
+order_by_key2_diff                     7.00    1.94    0.03    1.97     500 
+order_by_key_prefix                    4.00    1.04    0.08    1.12     500 
+order_by_range                         4.00    1.13    0.06    1.19     500 
+outer_join                          2539.00    0.00    0.01    0.01      10 
+outer_join_found                    2515.00    0.00    0.00    0.00      10 
+outer_join_not_found                124666.00    0.00    0.00    0.00     500 +
+outer_join_on_key                   2307.00    0.00    0.00    0.00      10 
+select_1_row                           6.00    2.25    0.46    2.71   10000 
+select_2_rows                          7.00    2.77    0.38    3.15   10000 
+select_big                            93.00   33.23    9.79   43.02   10080 
+select_column+column                   8.00    2.78    0.41    3.19   10000 
+select_diff_key                        0.00    0.21    0.02    0.23     500 
+select_distinct                       22.00    1.95    0.18    2.13     800 
+select_group                         326.00    1.47    0.20    1.67    2711 
+select_group_when_MANY_tables         15.00    1.57    0.78    2.35    5000 
+select_join                            4.00    0.52    0.13    0.65     100 
+select_key                           243.00   68.03    8.10   76.13  200000 
+select_key2                          208.00   66.48    8.68   75.16  200000 
+select_key2_return_key               200.00   66.41    7.77   74.18  200000 
+select_key2_return_prim              204.00   64.75    7.90   72.65  200000 
+select_key_prefix                    208.00   66.62    8.81   75.43  200000 
+select_key_prefix_join                11.00    4.30    0.16    4.46     100 
+select_key_return_key                239.00   66.86    8.37   75.23  200000 
+select_many_fields                   795.00    6.97    0.48    7.45    2000 
+select_query_cache                  2549.00    3.25    0.52    3.77   10000 
+select_query_cache2                 2547.00    3.04    0.53    3.57   10000 
+select_range                         465.00   10.41    0.63   11.04     410 
+select_range_key2                   20341.00    4.22    0.52    4.74   25010 ++
+select_range_prefix                 20344.00    6.32    1.04    7.36   25010 ++
+select_simple                          5.00    2.73    0.30    3.03   10000 
+select_simple_join                     3.00    0.74    0.03    0.77     500 
+update_big                          6046.00    0.01    0.00    0.01      10 
+update_of_key                        136.00   16.21   11.85   28.06   50000 
+update_of_key_big                    320.00    0.16    0.09    0.25     501 
+update_of_primary_key_many_keys     5365.00    0.16    0.03    0.19     256 
+update_with_key                      518.00   89.50   33.03  122.53  300000 
+update_with_key_prefix               186.00   30.32   15.83   46.15  100000 
+wisc_benchmark                        16.00    3.30    0.65    3.95     114 
+TOTALS                              224650.00 1060.42  234.52 1294.94 2551551 ++++++++

sql-bench/Results/alter-table-pg-Linux_2.4.0_64GB_SMP_i686-cmp-mysql,pg

+Testing server 'PostgreSQL version ???' at 2001-06-03  4:42:04
+
+Testing of ALTER TABLE
+Testing with 1000 columns and 1000 rows in 20 steps
+Insert data into the table
+Time for insert (1000) 1 wallclock secs ( 0.21 usr  0.08 sys +  0.00 cusr  0.00 csys =  0.29 CPU)
+
+Time for alter_table_add (992): 49 wallclock secs ( 0.26 usr  0.06 sys +  0.00 cusr  0.00 csys =  0.32 CPU)
+
+Time for create_index (8):  1 wallclock secs ( 0.00 usr  0.00 sys +  0.00 cusr  0.00 csys =  0.00 CPU)
+
+Time for drop_index (8):  0 wallclock secs ( 0.01 usr  0.00 sys +  0.00 cusr  0.00 csys =  0.01 CPU)
+
+Total time: 52 wallclock secs ( 0.49 usr  0.14 sys +  0.00 cusr  0.00 csys =  0.63 CPU)

sql-bench/Results/big-tables-pg-Linux_2.4.0_64GB_SMP_i686-cmp-mysql,pg

+Testing server 'PostgreSQL version ???' at 2001-06-03  4:42:56
+
+Testing of some unusual tables
+All tests are done 1000 times with 1000 fields
+
+Testing table with 1000 fields
+Testing select * from table with 1 record
+Time to select_many_fields(1000): 338 wallclock secs ( 3.28 usr  0.22 sys +  0.00 cusr  0.00 csys =  3.50 CPU)
+
+Testing select all_fields from table with 1 record
+Time to select_many_fields(1000): 457 wallclock secs ( 3.69 usr  0.26 sys +  0.00 cusr  0.00 csys =  3.95 CPU)
+
+Testing insert VALUES()
+Time to insert_many_fields(1000): 229 wallclock secs ( 0.40 usr  0.07 sys +  0.00 cusr  0.00 csys =  0.47 CPU)
+
+Testing insert (all_fields) VALUES()
+Time to insert_many_fields(1000): 299 wallclock secs ( 0.72 usr  0.14 sys +  0.00 cusr  0.00 csys =  0.86 CPU)
+
+Total time: 1324 wallclock secs ( 8.10 usr  0.69 sys +  0.00 cusr  0.00 csys =  8.79 CPU)

sql-bench/Results/connect-pg-Linux_2.4.0_64GB_SMP_i686-cmp-mysql,pg

+Testing server 'PostgreSQL version ???' at 2001-06-03  5:05:01
+
+Testing the speed of connecting to the server and sending of data
+All tests are done 10000 times
+
+Testing connection/disconnect
+Time to connect (10000): 143 wallclock secs ( 8.01 usr  1.89 sys +  0.00 cusr  0.00 csys =  9.90 CPU)
+
+Test connect/simple select/disconnect
+Time for connect+select_simple (10000): 157 wallclock secs (10.42 usr  2.41 sys +  0.00 cusr  0.00 csys = 12.83 CPU)
+
+Test simple select
+Time for select_simple (10000):  5 wallclock secs ( 2.73 usr  0.30 sys +  0.00 cusr  0.00 csys =  3.03 CPU)
+
+Testing connect/select 1 row from table/disconnect
+Time to connect+select_1_row (10000): 195 wallclock secs (10.94 usr  2.31 sys +  0.00 cusr  0.00 csys = 13.25 CPU)
+
+Testing select 1 row from table
+Time to select_1_row (10000):  6 wallclock secs ( 2.25 usr  0.46 sys +  0.00 cusr  0.00 csys =  2.71 CPU)
+
+Testing select 2 rows from table
+Time to select_2_rows (10000):  7 wallclock secs ( 2.77 usr  0.38 sys +  0.00 cusr  0.00 csys =  3.15 CPU)
+
+Test select with aritmetic (+)
+Time for select_column+column (10000):  8 wallclock secs ( 2.78 usr  0.41 sys +  0.00 cusr  0.00 csys =  3.19 CPU)
+
+Testing retrieval of big records (65000 bytes)
+Time to select_big (10000): 34 wallclock secs (11.75 usr  5.84 sys +  0.00 cusr  0.00 csys = 17.59 CPU)
+
+Total time: 555 wallclock secs (51.66 usr 14.01 sys +  0.00 cusr  0.00 csys = 65.67 CPU)

sql-bench/Results/create-pg-Linux_2.4.0_64GB_SMP_i686-cmp-mysql,pg

+Testing server 'PostgreSQL version ???' at 2001-06-03  5:14:17
+
+Testing the speed of creating and droping tables
+Testing with 5000 tables and 10000 loop count
+
+Testing create of tables
+Time for create_MANY_tables (5000): 160 wallclock secs ( 3.67 usr  1.35 sys +  0.00 cusr  0.00 csys =  5.02 CPU)
+
+Accessing tables
+Time to select_group_when_MANY_tables (5000): 15 wallclock secs ( 1.57 usr  0.78 sys +  0.00 cusr  0.00 csys =  2.35 CPU)
+
+Testing drop
+Time for drop_table_when_MANY_tables (5000): 772 wallclock secs ( 1.29 usr  0.28 sys +  0.00 cusr  0.00 csys =  1.57 CPU)
+
+Testing create+drop
+Time for create+drop (10000): 3280 wallclock secs (10.74 usr  1.89 sys +  0.00 cusr  0.00 csys = 12.63 CPU)
+Time for create_key+drop (10000): 5781 wallclock secs (10.70 usr  1.53 sys +  0.00 cusr  0.00 csys = 12.23 CPU)
+Total time: 10008 wallclock secs (27.97 usr  5.83 sys +  0.00 cusr  0.00 csys = 33.80 CPU)

sql-bench/Results/insert-pg-Linux_2.4.0_64GB_SMP_i686-cmp-mysql,pg

+Testing server 'PostgreSQL version ???' at 2001-06-03  8:01:05
+
+Testing the speed of inserting data into 1 table and do some selects on it.
+The tests are done with a table that has 100000 rows.
+
+Generating random keys
+Creating tables
+Inserting 100000 rows in order
+Inserting 100000 rows in reverse order
+Inserting 100000 rows in random order
+Time for insert (300000): 302 wallclock secs (89.07 usr 22.07 sys +  0.00 cusr  0.00 csys = 111.14 CPU)
+
+Testing insert of duplicates
+Time for insert_duplicates (100000): 120 wallclock secs (30.53 usr 10.61 sys +  0.00 cusr  0.00 csys = 41.14 CPU)
+
+Retrieving data from the table
+Time for select_big (10:3000000): 58 wallclock secs (21.31 usr  3.95 sys +  0.00 cusr  0.00 csys = 25.26 CPU)
+Time for order_by_big_key (10:3000000): 145 wallclock secs (24.01 usr  1.27 sys +  0.00 cusr  0.00 csys = 25.28 CPU)
+Time for order_by_big_key_desc (10:3000000): 145 wallclock secs (23.93 usr  1.27 sys +  0.00 cusr  0.00 csys = 25.20 CPU)
+Time for order_by_big_key_prefix (10:3000000): 133 wallclock secs (21.16 usr  0.80 sys +  0.00 cusr  0.00 csys = 21.96 CPU)
+Time for order_by_big_key2 (10:3000000): 132 wallclock secs (21.28 usr  0.64 sys +  0.00 cusr  0.00 csys = 21.92 CPU)
+Time for order_by_big_key_diff (10:3000000): 138 wallclock secs (21.30 usr  0.56 sys +  0.00 cusr  0.00 csys = 21.86 CPU)
+Time for order_by_big (10:3000000): 148 wallclock secs (21.11 usr  0.72 sys +  0.00 cusr  0.00 csys = 21.83 CPU)
+Time for order_by_range (500:125750):  4 wallclock secs ( 1.13 usr  0.06 sys +  0.00 cusr  0.00 csys =  1.19 CPU)
+Time for order_by_key_prefix (500:125750):  4 wallclock secs ( 1.04 usr  0.08 sys +  0.00 cusr  0.00 csys =  1.12 CPU)
+Time for order_by_key2_diff (500:250500):  7 wallclock secs ( 1.94 usr  0.03 sys +  0.00 cusr  0.00 csys =  1.97 CPU)
+Time for select_diff_key (500:1000):  0 wallclock secs ( 0.21 usr  0.02 sys +  0.00 cusr  0.00 csys =  0.23 CPU)
+Note: Query took longer then time-limit: 600
+Estimating end time based on:
+210 queries in 210 loops of 5010 loops took 616 seconds
+Estimated time for select_range_prefix (5010:1764): 14696 wallclock secs ( 2.62 usr  0.48 sys +  0.00 cusr  0.00 csys =  3.10 CPU)
+Note: Query took longer then time-limit: 600
+Estimating end time based on:
+210 queries in 210 loops of 5010 loops took 615 seconds
+Estimated time for select_range_key2 (5010:1764): 14672 wallclock secs ( 1.67 usr  0.24 sys +  0.00 cusr  0.00 csys =  1.91 CPU)
+Time for select_key_prefix (200000): 208 wallclock secs (66.62 usr  8.81 sys +  0.00 cusr  0.00 csys = 75.43 CPU)
+Time for select_key (200000): 243 wallclock secs (68.03 usr  8.10 sys +  0.00 cusr  0.00 csys = 76.13 CPU)
+Time for select_key_return_key (200000): 239 wallclock secs (66.86 usr  8.37 sys +  0.00 cusr  0.00 csys = 75.23 CPU)
+Time for select_key2 (200000): 208 wallclock secs (66.48 usr  8.68 sys +  0.00 cusr  0.00 csys = 75.16 CPU)
+Time for select_key2_return_key (200000): 200 wallclock secs (66.41 usr  7.77 sys +  0.00 cusr  0.00 csys = 74.18 CPU)
+Time for select_key2_return_prim (200000): 204 wallclock secs (64.75 usr  7.90 sys +  0.00 cusr  0.00 csys = 72.65 CPU)
+
+Test of compares with simple ranges
+Note: Query took longer then time-limit: 600
+Estimating end time based on:
+2160 queries in 54 loops of 500 loops took 610 seconds
+Estimated time for select_range_prefix (20000:4698): 5648 wallclock secs ( 3.70 usr  0.56 sys +  0.00 cusr  0.00 csys =  4.26 CPU)
+Note: Query took longer then time-limit: 600
+Estimating end time based on:
+2120 queries in 53 loops of 500 loops took 601 seconds
+Estimated time for select_range_key2 (20000:4611): 5669 wallclock secs ( 2.55 usr  0.28 sys +  0.00 cusr  0.00 csys =  2.83 CPU)
+Time for select_group (111): 274 wallclock secs ( 0.04 usr  0.01 sys +  0.00 cusr  0.00 csys =  0.05 CPU)
+Note: Query took longer then time-limit: 600
+Estimating end time based on:
+1320 queries in 220 loops of 2500 loops took 601 seconds
+Estimated time for min_max_on_key (15000): 6829 wallclock secs ( 5.23 usr  0.91 sys +  0.00 cusr  0.00 csys =  6.14 CPU)
+Time for min_max (60): 60 wallclock secs ( 0.02 usr  0.00 sys +  0.00 cusr  0.00 csys =  0.02 CPU)
+Time for count_on_key (100): 116 wallclock secs ( 0.02 usr  0.00 sys +  0.00 cusr  0.00 csys =  0.02 CPU)
+Time for count (100): 131 wallclock secs ( 0.03 usr  0.00 sys +  0.00 cusr  0.00 csys =  0.03 CPU)
+Time for count_distinct_big (20): 201 wallclock secs ( 0.03 usr  0.00 sys +  0.00 cusr  0.00 csys =  0.03 CPU)
+
+Testing update of keys with functions
+Time for update_of_key (50000):  136 wallclock secs (16.21 usr 11.85 sys +  0.00 cusr  0.00 csys = 28.06 CPU)
+Time for update_of_key_big (501): 320 wallclock secs ( 0.16 usr  0.09 sys +  0.00 cusr  0.00 csys =  0.25 CPU)
+
+Testing update with key
+Time for update_with_key (300000):  518 wallclock secs (89.50 usr 33.03 sys +  0.00 cusr  0.00 csys = 122.53 CPU)
+Time for update_with_key_prefix (100000):  186 wallclock secs (30.32 usr 15.83 sys +  0.00 cusr  0.00 csys = 46.15 CPU)
+
+Testing update of all rows
+Time for update_big (10):  6046 wallclock secs ( 0.01 usr  0.00 sys +  0.00 cusr  0.00 csys =  0.01 CPU)
+
+Testing left outer join
+Time for outer_join_on_key (10:10):  2307 wallclock secs ( 0.00 usr  0.00 sys +  0.00 cusr  0.00 csys =  0.00 CPU)
+Time for outer_join (10:10):  2539 wallclock secs ( 0.00 usr  0.01 sys +  0.00 cusr  0.00 csys =  0.01 CPU)
+Time for outer_join_found (10:10):  2515 wallclock secs ( 0.00 usr  0.00 sys +  0.00 cusr  0.00 csys =  0.00 CPU)
+Note: Query took longer then time-limit: 600
+Estimating end time based on:
+3 queries in 3 loops of 500 loops took 748 seconds
+Estimated time for outer_join_not_found (500:500):  124666 wallclock secs ( 0.00 usr  0.00 sys +  0.00 cusr  0.00 csys =  0.00 CPU)
+
+Testing INSERT INTO ... SELECT
+Time for insert_select_1_key (1):  86 wallclock secs ( 0.00 usr  0.00 sys +  0.00 cusr  0.00 csys =  0.00 CPU)
+Time for insert_select_2_keys (1):  196 wallclock secs ( 0.00 usr  0.00 sys +  0.00 cusr  0.00 csys =  0.00 CPU)
+Time for drop table(2): 22 wallclock secs ( 0.00 usr  0.00 sys +  0.00 cusr  0.00 csys =  0.00 CPU)
+
+Testing delete
+Time for delete_key (10000): 85 wallclock secs ( 3.19 usr  0.48 sys +  0.00 cusr  0.00 csys =  3.67 CPU)
+Time for delete_all (12): 2478 wallclock secs ( 0.01 usr  0.00 sys +  0.00 cusr  0.00 csys =  0.01 CPU)
+
+Insert into table with 16 keys and with a primary key with 16 parts
+Time for insert_key (100000): 804 wallclock secs (47.08 usr 47.06 sys +  0.00 cusr  0.00 csys = 94.14 CPU)
+
+Testing update of keys
+Time for update_of_primary_key_many_keys (256): 5365 wallclock secs ( 0.16 usr  0.03 sys +  0.00 cusr  0.00 csys =  0.19 CPU)
+
+Deleting rows from the table
+Time for delete_big_many_keys (128): 93 wallclock secs ( 0.05 usr  0.00 sys +  0.00 cusr  0.00 csys =  0.05 CPU)
+
+Deleting everything from table
+Time for delete_all_many_keys (1): 94 wallclock secs ( 0.05 usr  0.00 sys +  0.00 cusr  0.00 csys =  0.05 CPU)
+
+Estimated total time: 199360 wallclock secs (879.85 usr 202.59 sys +  0.00 cusr  0.00 csys = 1082.45 CPU)

sql-bench/Results/select-pg-Linux_2.4.0_64GB_SMP_i686-cmp-mysql,pg

+Testing server 'PostgreSQL version ???' at 2001-06-03 16:37:16
+
+Testing the speed of selecting on keys that consist of many parts
+The test-table has 10000 rows and the test is done with 500 ranges.
+
+Creating table
+Inserting 10000 rows
+Time to insert (10000): 10 wallclock secs ( 2.96 usr  0.39 sys +  0.00 cusr  0.00 csys =  3.35 CPU)
+
+Test if the database has a query cache
+Time for select_query_cache (10000): 2549 wallclock secs ( 3.25 usr  0.52 sys +  0.00 cusr  0.00 csys =  3.77 CPU)
+
+Time for select_query_cache2 (10000): 2547 wallclock secs ( 3.04 usr  0.53 sys +  0.00 cusr  0.00 csys =  3.57 CPU)
+
+Testing big selects on the table
+Time for select_big (70:17207):  1 wallclock secs ( 0.17 usr  0.00 sys +  0.00 cusr  0.00 csys =  0.17 CPU)
+Time for select_range (410:1057904): 465 wallclock secs (10.41 usr  0.63 sys +  0.00 cusr  0.00 csys = 11.04 CPU)
+Note: Query took longer then time-limit: 600
+Estimating end time based on:
+11326 queries in 1618 loops of 10000 loops took 601 seconds
+Estimated time for min_max_on_key (70000): 3714 wallclock secs (20.15 usr  3.46 sys +  0.00 cusr  0.00 csys = 23.61 CPU)
+Note: Query took longer then time-limit: 600
+Estimating end time based on:
+17320 queries in 3464 loops of 10000 loops took 601 seconds
+Estimated time for count_on_key (50000): 1734 wallclock secs (15.76 usr  1.99 sys +  0.00 cusr  0.00 csys = 17.75 CPU)
+
+Time for count_group_on_key_parts (1000:100000): 331 wallclock secs ( 1.13 usr  0.06 sys +  0.00 cusr  0.00 csys =  1.19 CPU)
+Testing count(distinct) on the table
+Time for count_distinct_key_prefix (1000:1000): 179 wallclock secs ( 0.28 usr  0.07 sys +  0.00 cusr  0.00 csys =  0.35 CPU)
+Time for count_distinct (1000:1000): 132 wallclock secs ( 0.31 usr  0.06 sys +  0.00 cusr  0.00 csys =  0.37 CPU)
+Time for count_distinct_2 (1000:1000): 213 wallclock secs ( 0.37 usr  0.03 sys +  0.00 cusr  0.00 csys =  0.40 CPU)
+Time for count_distinct_group_on_key (1000:6000): 488 wallclock secs ( 0.41 usr  0.03 sys +  0.00 cusr  0.00 csys =  0.44 CPU)
+Time for count_distinct_group_on_key_parts (1000:100000): 383 wallclock secs ( 1.10 usr  0.07 sys +  0.00 cusr  0.00 csys =  1.17 CPU)
+Time for count_distinct_group (1000:100000): 384 wallclock secs ( 1.07 usr  0.08 sys +  0.00 cusr  0.00 csys =  1.15 CPU)
+Time for count_distinct_big (100:1000000): 65 wallclock secs ( 7.88 usr  0.25 sys +  0.00 cusr  0.00 csys =  8.13 CPU)
+Estimated total time: 13197 wallclock secs (68.30 usr  8.18 sys +  0.00 cusr  0.00 csys = 76.48 CPU)

sql-bench/Results/wisconsin-pg-Linux_2.4.0_64GB_SMP_i686-cmp-mysql,pg

+Testing server 'PostgreSQL version ???' at 2001-06-03 19:06:27
+
+Wisconsin benchmark test
+
+Time for create_table (3):  1 wallclock secs ( 0.00 usr  0.00 sys +  0.00 cusr  0.00 csys =  0.00 CPU)
+
+Inserting data
+Time to insert (31000): 33 wallclock secs ( 9.09 usr  1.58 sys +  0.00 cusr  0.00 csys = 10.67 CPU)
+Time to delete_big (1):  0 wallclock secs ( 0.01 usr  0.00 sys +  0.00 cusr  0.00 csys =  0.01 CPU)
+
+Running actual benchmark
+Time for wisc_benchmark (114): 16 wallclock secs ( 3.30 usr  0.65 sys +  0.00 cusr  0.00 csys =  3.95 CPU)
+
+Total time: 52 wallclock secs (12.40 usr  2.23 sys +  0.00 cusr  0.00 csys = 14.63 CPU)

sql-bench/bench-init.pl.sh

@@ -31,7 +31,7 @@
 # $server	Object for current server
 # $limits	Hash reference to limits for benchmark
 
-$benchmark_version="2.12";
+$benchmark_version="2.13";
 use Getopt::Long;
 
 require "$pwd/server-cfg" || die "Can't read Configuration file: $!\n";

sql-bench/server-cfg.sh

@@ -122,53 +122,49 @@ sub new
   $self->{'vacuum'}		= 1; # When using with --fast
   $self->{'drop_attr'}		= "";
 
-  $limits{'max_conditions'}	= 9999; # (Actually not a limit)
-  $limits{'max_columns'}	= 2000;	# Max number of columns in table
-  # Windows can't handle that many files in one directory
-  $limits{'max_tables'}		= (($machine || '') =~ "^win") ? 5000 : 65000;
-  $limits{'max_text_size'}	= 65000; # Max size with default buffers.
-  $limits{'query_size'}		= 1000000; # Max size with default buffers.
-  $limits{'max_index'}		= 16; # Max number of keys
-  $limits{'max_index_parts'}	= 16; # Max segments/key
-  $limits{'max_column_name'}	= 64; # max table and column name
-
-  $limits{'join_optimizer'}	= 1; # Can optimize FROM tables
-  $limits{'load_data_infile'}	= 1; # Has load data infile
-  $limits{'lock_tables'}	= 1; # Has lock tables
-  $limits{'functions'}		= 1; # Has simple functions (+/-)
-  $limits{'group_functions'}	= 1; # Have group functions
-  $limits{'group_func_sql_min_str'} = 1; # Can execute MIN() and MAX() on strings
-  $limits{'group_distinct_functions'}= 1; # Have count(distinct)
-  $limits{'select_without_from'}= 1; # Can do 'select 1';
-  $limits{'multi_drop'}		= 1; # Drop table can take many tables
-  $limits{'subqueries'}		= 0; # Doesn't support sub-queries.
-  $limits{'left_outer_join'}	= 1; # Supports left outer joins
-  $limits{'table_wildcard'}	= 1; # Has SELECT table_name.*
-  $limits{'having_with_alias'}  = 1; # Can use aliases in HAVING
-  $limits{'having_with_group'}	= 1; # Can use group functions in HAVING
-  $limits{'like_with_column'}	= 1; # Can use column1 LIKE column2
-  $limits{'order_by_position'}  = 1; # Can use 'ORDER BY 1'
-  $limits{'group_by_position'}  = 1; # Can use 'GROUP BY 1'
-  $limits{'alter_table'}	= 1; # Have ALTER TABLE
+  $limits{'NEG'}		= 1; # Supports -id
   $limits{'alter_add_multi_col'}= 1; #Have ALTER TABLE t add a int,add b int;
+  $limits{'alter_table'}	= 1; # Have ALTER TABLE
   $limits{'alter_table_dropcol'}= 1; # Have ALTER TABLE DROP column
-  $limits{'insert_multi_value'} = 1; # Have INSERT ... values (1,2),(3,4)
-
-  $limits{'group_func_extra_std'} = 1; # Have group function std().
-
-  $limits{'func_odbc_mod'}	= 1; # Have function mod.
+  $limits{'column_alias'}	= 1; # Alias for fields in select statement.
   $limits{'func_extra_%'}	= 1; # Has % as alias for mod()
-  $limits{'func_odbc_floor'}	= 1; # Has func_odbc_floor function
   $limits{'func_extra_if'}	= 1; # Have function if.
-  $limits{'column_alias'}	= 1; # Alias for fields in select statement.
-  $limits{'NEG'}		= 1; # Supports -id
   $limits{'func_extra_in_num'}	= 1; # Has function in
-  $limits{'limit'}		= 1;		# supports the limit attribute
-  $limits{'unique_index'}	= 1; # Unique index works or not
+  $limits{'func_odbc_floor'}	= 1; # Has func_odbc_floor function
+  $limits{'func_odbc_mod'}	= 1; # Have function mod.
+  $limits{'functions'}		= 1; # Has simple functions (+/-)
+  $limits{'group_by_position'}  = 1; # Can use 'GROUP BY 1'
+  $limits{'group_distinct_functions'}= 1; # Have count(distinct)
+  $limits{'group_func_extra_std'} = 1; # Have group function std().
+  $limits{'group_func_sql_min_str'} = 1; # Can execute MIN() and MAX() on strings
+  $limits{'group_functions'}	= 1; # Have group functions
+  $limits{'having_with_alias'}  = 1; # Can use aliases in HAVING
+  $limits{'having_with_group'}	= 1; # Can use group functions in HAVING
+  $limits{'insert_multi_value'} = 1; # Have INSERT ... values (1,2),(3,4)
   $limits{'insert_select'}	= 1;
-  $limits{'working_blobs'}	= 1; # If big varchar/blobs works
+  $limits{'join_optimizer'}	= 1; # Can optimize FROM tables
+  $limits{'left_outer_join'}	= 1; # Supports left outer joins
+  $limits{'like_with_column'}	= 1; # Can use column1 LIKE column2
+  $limits{'limit'}		= 1;		# supports the limit attribute
+  $limits{'load_data_infile'}	= 1; # Has load data infile
+  $limits{'lock_tables'}	= 1; # Has lock tables
+  $limits{'max_column_name'}	= 64; # max table and column name
+  $limits{'max_columns'}	= 2000;	# Max number of columns in table
+  $limits{'max_conditions'}	= 9999; # (Actually not a limit)
+  $limits{'max_index'}		= 16; # Max number of keys
+  $limits{'max_index_parts'}	= 16; # Max segments/key
+  $limits{'max_tables'}		= (($machine || '') =~ "^win") ? 5000 : 65000;
+  $limits{'max_text_size'}	= 1000000; # Good enough for tests
+  $limits{'multi_drop'}		= 1; # Drop table can take many tables
+  $limits{'order_by_position'}  = 1; # Can use 'ORDER BY 1'
   $limits{'order_by_unused'}	= 1;
+  $limits{'query_size'}		= 1000000; # Max size with default buffers.
+  $limits{'select_without_from'}= 1; # Can do 'select 1';
+  $limits{'subqueries'}		= 0; # Doesn't support sub-queries.
+  $limits{'table_wildcard'}	= 1; # Has SELECT table_name.*
+  $limits{'unique_index'}	= 1; # Unique index works or not
   $limits{'working_all_fields'} = 1;
+  $limits{'working_blobs'}	= 1; # If big varchar/blobs works
 
   $smds{'time'}			= 1;
   $smds{'q1'} 	= 'b';		# with time not supp by mysql ('')
@@ -569,12 +565,12 @@ sub new
   $self->{'drop_attr'}		= "";
   $self->{"vacuum"}		= 1;
   $limits{'join_optimizer'}	= 1;		# Can optimize FROM tables
-  $limits{'load_data_infile'}	= 0;		# Is this true ?
+  $limits{'load_data_infile'}	= 0;
 
-  $limits{'NEG'}		= 1;		# Can't handle -id
-  $limits{'alter_table'}	= 1;		# alter ??
+  $limits{'NEG'}		= 1;
   $limits{'alter_add_multi_col'}= 0;		# alter_add_multi_col ?
-  $limits{'alter_table_dropcol'}= 0;		# alter_drop_col ?
+  $limits{'alter_table'}	= 1;
+  $limits{'alter_table_dropcol'}= 0;
   $limits{'column_alias'}	= 1;
   $limits{'func_extra_%'}	= 1;
   $limits{'func_extra_if'}	= 0;
@@ -583,33 +579,33 @@ sub new
   $limits{'func_odbc_mod'}	= 1;		# Has %
   $limits{'functions'}		= 1;
   $limits{'group_by_position'}  = 1;
+  $limits{'group_distinct_functions'}= 1; # Have count(distinct)
   $limits{'group_func_extra_std'} = 0;
   $limits{'group_func_sql_min_str'}= 1; # Can execute MIN() and MAX() on strings
   $limits{'group_functions'}	= 1;
-  $limits{'group_distinct_functions'}= 1; # Have count(distinct)
   $limits{'having_with_alias'}  = 0;
   $limits{'having_with_group'}	= 1;
-  $limits{'left_outer_join'}	= 0;
+  $limits{'insert_select'}	= 1;
+  $limits{'left_outer_join'}	= 1;
   $limits{'like_with_column'}	= 1;
   $limits{'lock_tables'}	= 0;		# in ATIS gives this a problem
+  $limits{'max_column_name'} 	= 128;
+  $limits{'max_columns'}	= 1000;		# 500 crashes pg 6.3
+  $limits{'max_conditions'}	= 9999;		# This makes Pg real slow
+  $limits{'max_index'}		= 64;		# Big enough
+  $limits{'max_index_parts'}	= 16;
+  $limits{'max_tables'}		= 5000;		# 10000 crashes pg 7.0.2
+  $limits{'max_text_size'}	= 65000;	# Good enough for test
   $limits{'multi_drop'}		= 1;
   $limits{'order_by_position'}  = 1;
+  $limits{'order_by_unused'}	= 1;
+  $limits{'query_size'}		= 16777216;
   $limits{'select_without_from'}= 1;
   $limits{'subqueries'}		= 1;
   $limits{'table_wildcard'}	= 1;
-  $limits{'max_column_name'} 	= 32;		# Is this true
-  $limits{'max_columns'}	= 1000;		# 500 crashes pg 6.3
-  $limits{'max_tables'}		= 5000;		# 10000 crashes pg 7.0.2
-  $limits{'max_conditions'}	= 30;		# This makes Pg real slow
-  $limits{'max_index'}		= 64;		# Is this true ?
-  $limits{'max_index_parts'}	= 16;		# Is this true ?
-  $limits{'max_text_size'}	= 7000;		# 8000 crashes pg 6.3
-  $limits{'query_size'}		= 16777216;
   $limits{'unique_index'}	= 1; # Unique index works or not
-  $limits{'insert_select'}	= 1;
-  $limits{'working_blobs'}	= 1; # If big varchar/blobs works
-  $limits{'order_by_unused'}	= 1;
   $limits{'working_all_fields'} = 1;
+  $limits{'working_blobs'}	= 1; # If big varchar/blobs works
 
   # the different cases per query ...
   $smds{'q1'} 	= 'b'; # with time
@@ -640,7 +636,7 @@ sub new
 sub version
 {
   my ($version,$dir);
-  foreach $dir ($ENV{'PGDATA'},"/usr/local/pgsql/data", "/my/local/pgsql/")
+  foreach $dir ($ENV{'PGDATA'},"/usr/local/pgsql/data", "/usr/local/pg/data")
   {
     if ($dir && -e "$dir/PG_VERSION")
     {

sql-bench/test-connect.sh

@@ -266,7 +266,7 @@ for ($i=0 ; $i < $opt_loop_count ; $i++)
 }
 
 $end_time=new Benchmark;
-print "Time to select_big ($opt_loop_count): " .
+print "Time to select_big_str ($opt_loop_count): " .
   timestr(timediff($end_time, $loop_time),"all") . "\n\n";
 
 $sth = $dbh->do("drop table bench1" . $server->{'drop_attr'})

sql/ha_gemini.cc

@@ -19,11 +19,7 @@
 #pragma implementation				// gcc: Class implementation
 #endif
 
-#include <string.h>
-
 #include "mysql_priv.h"
-#include "my_pthread.h"
-
 #ifdef HAVE_GEMINI_DB
 #include "ha_gemini.h"
 #include "dbconfig.h"
@@ -33,6 +29,7 @@
 
 #include <m_ctype.h>
 #include <myisampack.h>
+#include <m_string.h>
 #include <assert.h>
 #include <hash.h>
 #include <stdarg.h>

sql/ha_innobase.cc

@@ -2142,6 +2142,7 @@ ha_innobase::external_lock(
 	prebuilt->in_update_remember_pos = TRUE;
 
 	if (lock_type == F_WRLCK) {
+
 		/* If this is a SELECT, then it is in UPDATE TABLE ...
 		or SELECT ... FOR UPDATE */
 		prebuilt->select_lock_type = LOCK_X;
@@ -2153,15 +2154,29 @@ ha_innobase::external_lock(
 		}
 
 		trx->n_mysql_tables_in_use++;
+
+		if (prebuilt->select_lock_type != LOCK_NONE) {
+
+		  trx->mysql_n_tables_locked++;
+		}
 	} else {
 		trx->n_mysql_tables_in_use--;
 
-		if (trx->n_mysql_tables_in_use == 0 &&
-		    !(thd->options
+		if (trx->n_mysql_tables_in_use == 0) {
+
+		  trx->mysql_n_tables_locked = 0;
+
+		  if (trx->has_search_latch) {
+
+		    trx_search_latch_release_if_reserved(trx);
+		  }
+
+		  if (!(thd->options
 			& (OPTION_NOT_AUTO_COMMIT | OPTION_BEGIN))) {
 		    innobase_commit(thd, trx);
 		  }
 		}
+	}
 
 	DBUG_RETURN(error);
 }
@@ -2690,6 +2705,39 @@ ha_innobase::info(
   	DBUG_VOID_RETURN;
 }
 
+/***********************************************************************
+Tries to check that an InnoDB table is not corrupted. If corruption is
+noticed, prints to stderr information about it. In case of corruption
+may also assert a failure and crash the server. */
+
+int
+ha_innobase::check(
+/*===============*/
+					/* out: HA_ADMIN_CORRUPT or
+					HA_ADMIN_OK */
+	THD* 		thd,		/* in: user thread handle */
+	HA_CHECK_OPT* 	check_opt)	/* in: check options, currently
+					ignored */
+{
+	row_prebuilt_t* prebuilt	= (row_prebuilt_t*) innobase_prebuilt;
+	ulint		ret;
+	
+	if (prebuilt->mysql_template == NULL) {
+		/* Build the template; we will use a dummy template
+		in index scans done in checking */
+
+		build_template(prebuilt, NULL, table, ROW_MYSQL_WHOLE_ROW);
+	}
+
+	ret = row_check_table_for_mysql(prebuilt);
+
+	if (ret == DB_SUCCESS) {
+		return(HA_ADMIN_OK);
+	}
+	
+  	return(HA_ADMIN_CORRUPT); 
+}
+
 /*****************************************************************
 Adds information about free space in the InnoDB tablespace to a
 table comment which is printed out when a user calls SHOW TABLE STATUS. */

sql/ha_innobase.h

@@ -142,7 +142,7 @@ class ha_innobase: public handler
 					HA_CREATE_INFO *create_info);
   	int delete_table(const char *name);
 	int rename_table(const char* from, const char* to);
-
+	int check(THD* thd, HA_CHECK_OPT* check_opt);
         char* update_table_comment(const char* comment);
 
   	THR_LOCK_DATA **store_lock(THD *thd, THR_LOCK_DATA **to,

sql/mysqld.cc

@@ -95,17 +95,16 @@ int deny_severity = LOG_WARNING;
 #include <sys/mman.h>
 #endif
 
+#ifdef _AIX41
+int initgroups(const char *,unsigned int);
+#endif
+
 #if defined(__FreeBSD__) && defined(HAVE_IEEEFP_H)
 #include <ieeefp.h>
 #ifdef HAVE_FP_EXCEPT				// Fix type conflict
 typedef fp_except fp_except_t;
 #endif
 
-#ifdef _AIX41
-extern "C" int initgroups(const char *,int);
-#endif
-
-
   /* We can't handle floating point expections with threads, so disable
      this on freebsd
   */

tests/fork_big.pl

@@ -88,6 +88,7 @@ for ($i=0 ; $i < $opt_threads ; $i ++)
 {
   test_select() if (($pid=fork()) == 0); $work{$pid}="select_key";
 }
+test_select_count() if (($pid=fork()) == 0); $work{$pid}="select_count";
 test_delete() if (($pid=fork()) == 0); $work{$pid}="delete";
 test_update() if (($pid=fork()) == 0); $work{$pid}="update";
 test_flush() if (($pid=fork()) == 0); $work{$pid}= "flush";
@@ -213,6 +214,35 @@ sub test_select
   exit(0);
 }
 
+#
+# Do big select count(distinct..) over the table
+# 
+
+sub test_select_count
+{
+  my ($dbh, $i, $j, $count, $loop);
+
+  $dbh = DBI->connect("DBI:mysql:$opt_db:$opt_host",
+		      $opt_user, $opt_password,
+		    { PrintError => 0}) || die $DBI::errstr;
+
+  $count=0;
+  $i=0;
+  while (!test_if_abort($dbh))
+  {
+    for ($j=0 ; $j < $numtables ; $j++)
+    {
+      my ($table)= $testtables[$j]->[0];
+      simple_query($dbh, "select count(distinct marker),count(distinct id),count(distinct info) from $table");
+      $count++;
+    }
+    sleep(20);		# This query is quite slow
+  }
+  $dbh->disconnect; $dbh=0;
+  print "Test_select: Executed $count select count(distinct) queries\n";
+  exit(0);
+}
+
 #
 # Delete 1-5 rows from the first 2 tables.
 # Test ends when the number of rows for table 3 didn't change during