Peter,
I just finished reviewing all uses of prepareStatement, and spotted the
dangling statement in AbstractJdbcUsersRepository.getUserByName(), too. I
was just finishing up a reply when your message came.
This is a bug, and it really ought to be fixed. The change is
well-understood, used in all other areas, and easily tested. We know that
we have a number of Oracle users, and this is a resource leak for them. I
just checked my private tree against the CVS, and I don't have any other
differences that would be termed a serious or potentially wide-spread
defect. We ought to vote on whether we would prefer to support the Release
with or without the fix.
By the way, Christie, good catch! :-)
--- Noel
-----Original Message-----
From: Peter M. Goldstein [mailto:[EMAIL PROTECTED]]
Sent: Sunday, November 24, 2002 14:00
To: 'James Users List'
Subject: RE: James-Oracle Database-Outlook
Noel and Christie,
I thought we had nailed this one too. We did indeed add code that
closed almost all the PreparedStatements. Except upon examination of
the code it seems we missed one. In AbstractJdbcUsersRepository's
getUserByName method we don't properly close either the result set or
the prepared statement. I'm not sure how we missed that one.
Christie, we are now past code freeze and I'm reluctant to suggest we
break it for a problem like this. So I'd suggest one of two options.
i) If you feel comfortable doing so, I'll send you the patched source
code and you can build your own .sar file for deployment or ii) I can
build you a .sar file to use. Then, after 2.1 is out the door, we can
patch this and a couple of other minor issues that have arisen since
code freeze and deliver a minor patch release. Would either of these
options work for you?
--Peter
--
To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]>
For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
