[
https://issues.apache.org/jira/browse/DERBY-3842?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12689031#action_12689031
]
Tiago R. Espinha commented on DERBY-3842:
-----------------------------------------
I have done some modifications to this test, but now on a second read of the
comments I'm not so sure they were good changes.
I see we have a setUp and a tearDown, but for JDBC tests can't/shouldn't we
just use the decorateSQL to create and populate the table?
Additionally, I see that Mamta suggested that we drop the foo table after the
fixture testOrderWithMultipleLevel. On the actual patch by Junjie the table is
being dropped at the end of each fixture but I'm not sure about this.
On the original test it seems to me that the table is created and populated
*once* and then it is queried on several tests, having only the function,
procedure and the table dropped at the end of the whole set.
Can't this influence the outcome and the purpose of this test? And maybe even
make it slower unnecessarily?
Finally, I will make it so that we also look at the actual data, since that is
actually done on the original test. Although the relevance of this is still
debatable I believe. If the cursor isn't «held», then the row number will also
be different than expected since this isn't quite a row number but just a
regular numeric key that is initially inserted in the database.
> Convert
> "org.apache.derbyTesting.functionTests.tests.store.holdCursorExternalSortJDBC30.sql"
> to junit.
> ------------------------------------------------------------------------------------------------------
>
> Key: DERBY-3842
> URL: https://issues.apache.org/jira/browse/DERBY-3842
> Project: Derby
> Issue Type: Test
> Components: Test
> Reporter: Junjie Peng
> Assignee: Tiago R. Espinha
> Attachments: derby-3842-1.patch, derby-3842-1.stat
>
>
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
