branches/innodb+: Merge revisions 2862:2867 from branches/zip:

------------------------------------------------------------------------
  r2866 | marko | 2008-10-23 23:25:43 +0300 (Thu, 23 Oct 2008) | 4 lines

  branches/zip: ibuf_delete_rec(): When the cursor to the insert buffer record
  cannot be restored, do not complain if the tablespace does not exist.
  This fixes Issue #88.
  ------------------------------------------------------------------------
  r2867 | marko | 2008-10-24 10:24:17 +0300 (Fri, 24 Oct 2008) | 2 lines

  branches/zip: ChangeLog: Document r2763, r2794, r2683, r2799, r2809, r2866.
  ------------------------------------------------------------------------

ChangeLog

@@ -1,3 +1,13 @@
+2008-10-23	The InnoDB Team
+
+	* ibuf/ibuf0ibuf.c:
+
+	ibuf_delete_rec(): When the cursor to the insert buffer record
+	cannot be restored, do not complain if the tablespace does not
+	exist, because the insert buffer record may have been discarded by
+	some other thread.  This bug has existed in MySQL/InnoDB since
+	version 4.1, when innodb_file_per_table was implemented.
+
 2008-10-22	The InnoDB Team
 
 	* dict/dict0dict.c, dict/dict0mem.c, handler/ha_innodb.cc,
@@ -15,6 +25,64 @@
 	Fix Bug#40224 New AUTOINC changes mask reporting of deadlock/timeout
 	errors
 
+2008-10-16	The InnoDB Team
+
+	* dict/dict0dict.c,
+	mysql-test/innodb-index.result, mysql-test/innodb-index.test:
+	Skip the undo log size check when creating REDUNDANT and COMPACT
+	tables.  In ROW_FORMAT=DYNAMIC and ROW_FORMAT=COMPRESSED, column
+	prefix indexes require that prefixes of externally stored columns
+	be written to the undo log. This may make the undo log record
+	bigger than the record on the B-tree page.  The maximum size of an
+	undo log record is the page size.  That must be checked for, in
+	dict_index_add_to_cache().  However, this restriction must not
+	be enforced on REDUNDANT or COMPACT tables.
+
+2008-10-15	The InnoDB Team
+
+	* btr/btr0cur.c, include/btr0cur.h, row/row0ext.c,
+	row/row0sel.c, row/row0upd.c:
+	When the server crashes while freeing an externally stored column
+	of a compressed table, the BTR_EXTERN_LEN field in the BLOB
+	pointer will be written as 0.  Tolerate this in the functions that
+	deal with externally stored columns.  This fixes problems after
+	crash recovery, in the rollback of incomplete transactions, and in
+	the purge of delete-marked records.
+
+2008-10-15	The InnoDB Team
+
+	* btr/btr0btr.c, include/page0zip.h, page/page0zip.c, include/univ.i:
+	When a B-tree node of a compressed table is split or merged, the
+	compression may fail.  In this case, the entire compressed page
+	will be copied and the excess records will be deleted.  However,
+	page_zip_copy(), now renamed to page_zip_copy_recs(), copied too
+	many fields in the page header, overwriting PAGE_BTR_SEG_LEAF and
+	PAGE_BTR_SEG_TOP when splitting the B-tree root.  This caused
+	corruption of compressed tables.  Furthermore, the lock table and
+	the adaptive hash index would be corrupted, because we forgot to
+	update them when invoking page_zip_copy_recs().
+
+	Introduce the symbol UNIV_ZIP_DEBUG for triggering the copying of
+	compressed pages more often, for debugging purposes.
+
+2008-10-10	The InnoDB Team
+
+	* handler/handler0alter.cc, include/row0merge.h, row/row0merge.c,
+	row/row0mysql.c:
+	Fix some locking issues, mainly in fast index creation.  The
+	InnoDB data dictionary cache should be latched whenever a
+	transaction is holding locks on any data dictionary tables.
+	Otherwise, lock waits or deadlocks could occur.  Furthermore, the
+	data dictionary transaction must be committed (and the locks
+	released) before the data dictionary latch is released.
+
+	ha_innobase::add_index(): Lock the data dictionary before renaming
+	or dropping the created indexes, because neither operation will
+	commit the data dictionary transaction.
+
+	ha_innobase::final_drop_index(): Commit the transactions before
+	unlocking the data dictionary.
+
 2008-10-09	The InnoDB Team
 
 	* buf/buf0lru.c:

ibuf/ibuf0ibuf.c

@@ -3658,6 +3658,13 @@ ibuf_delete_rec(
 	success = btr_pcur_restore_position(BTR_MODIFY_TREE, pcur, mtr);
 
 	if (!success) {
+		if (fil_space_get_flags(space) == ULINT_UNDEFINED) {
+			/* The tablespace has been dropped.  It is possible
+			that another thread has deleted the insert buffer
+			entry.  Do not complain. */
+			goto func_exit;
+		}
+
 		fprintf(stderr,
 			"InnoDB: ERROR: Submit the output to"
 			" http://bugs.mysql.com\n"
@@ -3684,11 +3691,7 @@ ibuf_delete_rec(
 		fprintf(stderr, "InnoDB: ibuf tree ok\n");
 		fflush(stderr);
 
-		btr_pcur_close(pcur);
-
-		mutex_exit(&ibuf_mutex);
-
-		return(TRUE);
+		goto func_exit;
 	}
 
 	root = ibuf_tree_root_get(mtr);
@@ -3699,13 +3702,12 @@ ibuf_delete_rec(
 
 #ifdef UNIV_IBUF_COUNT_DEBUG
 	ibuf_count_set(space, page_no, ibuf_count_get(space, page_no) - 1);
-#else
-	UT_NOT_USED(space);
 #endif
 	ibuf_size_update(root, mtr);
 
 	btr_pcur_commit_specify_mtr(pcur, mtr);
 
+func_exit:
 	btr_pcur_close(pcur);
 
 	mutex_exit(&ibuf_mutex);