Briggs wrote:
Well, here is a very simple test case. It's not junit or anything, but
if you look at the code, by calling the main method it will:
1) Start up a network server
2) Create a test table
3) Insert 20K duplicate records
4) Attempt to iterate though them, and fail.
If you wish to test the reading portion (after you've created a
database), you'll need to comment out the code that does so. Sorry,
haven't the time to get too fancy with the test case. But, you can
see that it will fail. Plus, when you run out of memory, don't forget
to kill the process or you'll have a db running on 2222.
If you are running unix, you'll need to change the connection string
to point to a valid dir (other than c:/tmp).
Is this good enough?
Oh, if you want, you can also just create the database and insert the
data, start a network server, then connect via ij. Then just do a
"select * from test_case.test_lob"; you will eventually get an index
out of bounds exception because the server caused an error.
Yes, the code is ugly, but I didn't think connection handling and
transaction management was important.
On Feb 8, 2008 1:51 PM, Briggs <[EMAIL PROTECTED]> wrote:
Working on a test case now...
On Feb 8, 2008 1:29 PM, Kathey Marsden <[EMAIL PROTECTED]> wrote:
Øystein Grøvlen wrote:
This looks like DERBY-3354. It is a bug, and as far as I can tell the
cause has been identified, so it should probably be possible to fix
this. For client/server you will probably also need the fix for
DERBY-2892.
Since he said his case works fine in embedded mode, I don't think it is
DERBY-3354. Briggs, could you post a reproducible test case so we can
see if this is a new issue.
Kathey
--
"Conscious decisions by conscious minds are what make reality real"
Running against 10.3.2.1 with IBM JDK 1.5 I am not seeing the OOM
exception. Should I be running with a smaller heap? I also had to add
this line to the program to avoid a no suitable driver error.
Class.forName("org.apache.derby.jdbc.ClientDriver").newInstance();
The output ends with:
....
Read: 19997 records.
Read: 19998 records.
Read: 19999 records.
