Mailing-List: contact derby-commits-help@db.apache.org; run by ezmlm
Precedence: bulk
Reply-To: "Derby Development" <derby-dev@db.apache.org>
Content-Type: text/plain; charset="utf-8"
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Subject: svn propchange: r1182570 - svn:log
Date: Wed, 12 Oct 2011 23:44:07 -0000
To: derby-commits@db.apache.org
From: mamta@apache.org
Message-Id: <20111012234407.E31C323889DE@eris.apache.org>

Author: mamta
Revision: 1182570
Modified property: svn:log

Modified: svn:log at Wed Oct 12 23:44:07 2011
------------------------------------------------------------------------------
--- svn:log (original)
+++ svn:log Wed Oct 12 23:44:07 2011
@@ -1,16 +1,8 @@
 DERBY-3823 NullPointerException in stress.multi test
 
-Adding a test case showing that in case of a network server, an open resulset's metadata can get changed underneath it but it is
-	not reflected in the metadata. The test creates a table with one of column as varchar(5). It inserts 1000 rows and then
-	opens a reulset on that table with varchar column as one of the columns. The test verifies that the reulset's metadata
-	at this point shows the length of the column as 5. Next, while the resulset is still open, the tests does an ALTER TABLE
-	to increase the varchar column's length to 8. In case of embedded mode, this fails because of the open resulset. In case 
-	of network server, because of prefetching of rows, the ALTER TABLE is allowed but when the test gets the resulset's 
-	metadata again and checks the length of varchar column, it still shows the length to be 5 rather than 8. 
+Adding a test case showing that in case of a network server, an open resulset's metadata can get changed underneath it but it is not reflected in the metadata. The test creates a table with one of column as varchar(5). It inserts 1000 rows and then opens a reulset on that table with varchar column as one of the columns. The test verifies that the reulset's metadata at this point shows the length of the column as 5. Next, while the resulset is still open, the tests does an ALTER TABLE to increase the varchar column's length to 8. In case of embedded mode, this fails because of the open resulset. In case of network server, because of prefetching of rows, the ALTER TABLE is allowed but when the test gets the resulset's metadata again and checks the length of varchar column, it still shows the length to be 5 rather than 8. 
 
 There are couple other jiras related to network server prefetching, namely, DERBY-3839 and DERBY-4373.
 
-Once DERBY-3823 is fixed, we should see the change in metadata reflected in resultset's metadata. A fix for DERBY-3823 will
-        cause the following the test added here to fail. Right now, the new test accepts the incorrect metadata length obtained 
-	through the resultset's metadata after ALTER TABLE has been performed in network server mode.
+Once DERBY-3823 is fixed, we should see the change in metadata reflected in resultset's metadata. A fix for DERBY-3823 will cause the following the test added here to fail. Right now, the new test accepts the incorrect metadata length obtained through the resultset's metadata after ALTER TABLE has been performed in network server mode.