db-derby-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "A B (JIRA)" <derby-...@db.apache.org>
Subject [jira] Commented: (DERBY-498) Result set holdability defined inside stored procedures is ignored by server/client
Date Fri, 07 Oct 2005 18:57:49 GMT
    [ http://issues.apache.org/jira/browse/DERBY-498?page=comments#action_12331617 ] 

A B commented on DERBY-498:

I applied the patch to my local codeline and ran the repro as well as the new holdCursorJava.java
test, and it all works correctly.  I also ran the new test cases without the server changes
and the test failed as expected:

< 21, testtable2-one
> EXPECTED:ResultSet is null
< EXPECTED:ResultSet is null
> 11, testtable1-one

So that's good.

I'm not familiar enough with the XA code to say for sure if there are any concerns in that
area, but I did review the changes overall and they look good to me.  My one (very minor)
comment is that the code path for NON-callable statements changes ever-so-slightly with this
patch: in the "addResultSet' method, the value of "newDrdaRs.withHoldCursor" wasn't getting
set before (so far as I can tell); now, it's set to the current DRDAStatement's "withHoldCursor"
value (which is passed in from DRDAStatement.execute()).  I didn't see anything wrong with
this when I looked at the code--and in fact this change seems more "correct" to me--but it
did make me think twice while I was reviewing it, so I thought I'd mention it..  If this is
intentional and okay, then I vote +1 with the hope that someone more familiar with XA can
review this patch, as well...

> Result set holdability defined inside stored procedures is ignored by server/client
> -----------------------------------------------------------------------------------
>          Key: DERBY-498
>          URL: http://issues.apache.org/jira/browse/DERBY-498
>      Project: Derby
>         Type: Bug
>   Components: Network Client, Network Server
>     Versions:,
>     Reporter: A B
>     Assignee: Deepa Remesh
>  Attachments: d498.java, derby-498.diff, derby-498.status
> Assume I have a Java stored procedure that returns one or more result sets, and the holdability
of those result sets is specified as part of the createStatement() method within the procedure
definition (see below for an example).
> If I execute this procedure against Derby embedded, the holdability of each result set
matches that of the statement-specific holdability that is defined within the stored procedure.
 However, if I run the procedure against the Network Server using the Derby client, the holdability
of _all_ result sets is the same, and it is based on the holdability of the statement that
_executed_ the procedure--i.e. the statement-specific holdability that is defined within the
procedure is ignored.
> Ex: If I create a stored procedure that corresponds to the following method:
> public static void p2(ResultSet[] rs1, ResultSet[] rs2,
>     ResultSet[] rs3) throws Exception
> {
>     Connection conn = DriverManager.getConnection(
>         "jdbc:default:connection");
>     Statement st1 = conn.createStatement(
>         ResultSet.TYPE_FORWARD_ONLY,
>         ResultSet.CONCUR_READ_ONLY,
>         ResultSet.HOLD_CURSORS_OVER_COMMIT);
>     rs1[0] = st1.executeQuery("select * from testtable1");
>     Statement st2 = conn.createStatement(
>         ResultSet.TYPE_FORWARD_ONLY,
>         ResultSet.CONCUR_READ_ONLY,
>         ResultSet.CLOSE_CURSORS_AT_COMMIT);
>     rs2[0] = st2.executeQuery("select * from testtable2");
>     Statement st3 = conn.createStatement(
>         ResultSet.TYPE_FORWARD_ONLY,
>         ResultSet.CONCUR_READ_ONLY,
>         ResultSet.HOLD_CURSORS_OVER_COMMIT);
>     rs3[0] = st3.executeQuery("select * from testtable3");
>     return;
>     }
> }
> Then with Derby embedded, if I have a JDBC Statement that executes a call to this procedure,
rs1 and rs3 will behave with HOLD_CURSORS holdability and rs2 will behave with CLOSE_CURSORS
holdability--and that will be the case regardless of the holdability on the Statement that
executed the call.  That seems correct to me.
> But if I do the same thing with Network Server, all of the result sets (rs1, rs2, and
rs3) will have the same holdability as the JDBC Statement that executed the call.  It doesn't
matter what the holdabilities used within the procedure definition are: they will all be over-ridden
by the holdability of the Statement that made the call.

This message is automatically generated by JIRA.
If you think it was sent incorrectly contact one of the administrators:
For more information on JIRA, see:

View raw message