> You keep claiming that this problem can be solved, but have yet to
> demonstrate a working test case.

Hmmm.. not sure where that comes from since no such claim was made.

"Using pure jdbc api's, the size of the LOB will be limited to 4-5k when
using the 'thin' driver"

My claim is that the thin driver fails to load LOBs greater than 4K under
JBoss only. The "pure" APIs are Oracle's implementation contained in the
driver. Under WebLogic there was no jDriver in use: that is the driver
WebLogic provides. What I used in WebLogic was
oracle.jdbc.driver.OracleDriver and also in a stand-alone test. There are
several ways of loading and storing LOBs and that was my point: they all
work using the thin driver except when in JBoss. There is no 4K limit
specified by JDBC and there is no such limit specified by Oracle and that
seems to hold water in that it works everywhere except in JBoss.

So you see, there is no way I can post a working example when in fact I am
in complete agreement that no matter what you do it fails if you are using
the "thin" driver. The solution is to use the OCI driver which we also agree
on.  The point of disagreement is on your claim about the 4K limit being
Oracle's problem and my claim that it's a JBoss issue. If you were to
provide an example that backs your assertion outside of JBoss that would be
convincing. Then again, maybe I goofed up my previous test when this issue
first came up. Let me throw one together tomorrow morning and if it fails
with the thin driver outside of JBoss I will assume my previous tests were
in error and log a TAR with Oracle. Let me retract my "FALSE" and subtitute
"wait a minute, something doesn't fit here".

----- Original Message -----
From: "Stephen Coy" <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Sunday, February 16, 2003 8:50 PM
Subject: Re: [JBoss-user] JMS persistence manager for Oracle8


> On Monday, February 17, 2003, at 02:51  PM, Rod Macpherson wrote:
>
> > I still hold there is no
> > 4K limit using JDBC calls.
>
> I have tested this in the past using straight java(x).sql.* calls and
> was unable to get it to work unless the oci driver was used. I suspect
> that weblogic is using the Oracle APIs, because they don't have to
> worry about source distribution and enabling Joe Programmer to build it.
>
> You keep claiming that this problem can be solved, but have yet to
> demonstrate a working test case.
>
> This:
> > PreparedStatement statement = connnection.prepareStatement(
> > "UPDATE BAR SET FOO = ? WHERE BAR_ID = 1");
> > statement.setBinaryStream(1, stream, file.length());
> > statement.executeUpdate();
>
> is exactly what JBoss CMP is doing now, and it exhibits the 4k problem.
>
> Steve Coy
>
>
>
> -------------------------------------------------------
> This sf.net email is sponsored by:ThinkGeek
> Welcome to geek heaven.
> http://thinkgeek.com/sf
> _______________________________________________
> JBoss-user mailing list
> [EMAIL PROTECTED]
> https://lists.sourceforge.net/lists/listinfo/jboss-user



-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
JBoss-user mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/jboss-user

Reply via email to