Bidirectional relationships do not yet work.
There are two workarounds that have worked for me:
Manually add (and remove!) both sides of the relationship. So your
Marketplace.addStorefront() method would look like this:
void addStorefront(Storefront front)
{
I frequently hear this mantra repeated, and while it is largely a good
idea, I have difficulty seeing why it should be adhered to dogmatically.
There is value in being able to define side-effects of setters and
getters or minimal bean-specific business logic. As a trivial example I
offer my
Why not just put the superclass initialization in ejbCreate() and
require all the subclasses to call super.ejbCreate()?
Jeff
-Original Message-
From: Randahl Fink Isaksen [mailto:[EMAIL PROTECTED]]
Sent: Monday, February 19, 2001 4:18 PM
To: Orion-Interest
Subject: Best practices: How
From: Joseph B. Ottinger [mailto:[EMAIL PROTECTED]]
Why do you say "pity?" (I'm assuming you don't mean "piety" here.) Why
should it be open source? Do you think you can apply patches
faster than
the Orion team? (I don't think I could, nor do I think you
For me, the value of source is not
Kaffe is the Open Source implementation of Java that is shipped with
most flavors of Linux. It looks like the Kaffe version of the java
command is higher up in the path than the Sun JDK java command, thus you
are running the wrong JVM.
I suggest "rpm --erase kaffe". That should solve your
I'm not sure the same lessons apply; while still a small company by BEA
standards, Unify has a lot of mouths to feed. Other than 1999, it looks
like all of their years have been unprofitable, and 2000 was
dramatically so. They've been financing this with equity issues, but at
the current
Dude, do you really think a Swedish company with a handful of employees
is going to be able to field a worldwide army of training professionals?
My guess is that the entire world population of Orion experts is reading
this right now; depending on where you are, a nicely phrased request and
an
Dude, do you really think a Swedish company with a handful of employees
is going to be able to field a worldwide army of training
professionals?
It just occurred to me that this sounds rather bad... I was trying to
imply that it would be unlikely for such training professionals to be
sent all
That's an interesting question. The investor story is not good - they
recently terminated their CEO/President and CFO, restated a year of
earnings (which put them in the red), and reduced their heacount from
132 to 106. Their stock, which was trading ~$20 this time last year, is
now at $0.31.
:
OpenEJB is the brainchild of Richard Monson-Haefel, author of
Enterprise JavaBeans, 2nd Edition (O'Reilly 2000). The core development
team for OpenEJB is Mr. Monson-Haefel and David Blevins.
-Original Message-
From: Jeff Schnitzer [ mailto:[EMAIL PROTECTED
different approaches, so I
fully expect the answer to be "it depends" :-) What do various app
servers do? The j2ee spec doesn't seem to prescribe an answer to this.
Thanks,
Jeff Schnitzer
[EMAIL PROTECTED] mailto:[EMAIL PROTECTED]
===
much more complicted buildfile
("similarity"), which shows one ejb-jar, a main web-war, a test web-war,
two application clients, and a lot of rmi compiling (some using IIOP,
which requires a nightly build version of ant).
Enjoy :-)
Jeff Schnitzer
[EMAIL PROTECTED]
ServerTest.zip
?x
Are you sure that the JDBC driver isn't finding its way into the
classpath? Did you try the console remotely from a machine that you
know doesn't have the JDBC driver?
You may already know this, but just in case: Almost all of the orion
tools (including orionconsole.jar) are simply empty jar
I'm fully prepared to believe that my understanding of how this works is
wrong, but if so it raises quite a few questions for me.
Connection, Statement, ResultSet, etc are just interfaces; something
must implement them. Normally they are implemented by the JDBC driver,
thus allowing the client
From looking at the source, counter.jar only writes to the database for
every ten keys generated. With pessimistic concurrency, loads do not
occur for each use.
Here's the relevant code from CounterEJB.java:
public String ejbCreate(String name)
throws CreateException
{
ject: Re: Session EJB Accessibility
I believe Orion uses pessimistic concurrency control.
However, I don't
think optimistic concurrency control would allow dirty reads -- I
understand
a dirty read to imply a transaction seeing the uncommitted state of
another
transaction.
Vidur
Jeff Schnitzer
I'm confused by your comments; does it need to manage state, or doesn't
it? I'm assuming it does, otherwise you would just use a stateless
session bean.
Here's some fodder for conversation:
I don't think there is an EJB facility which will help you. SLSBs are
pooled and can timeout, SFSBs
of view, I just disagree with it.
Jeff Schnitzer wrote:
If the client is going to use the JDBC driver, it must be
able to load
the class(es). This means you need to package the driver with the
client application. I'm puzzled by your comments about clients not
needing to care about drivers
From: Gary Shea [mailto:[EMAIL PROTECTED]]
I could use a little help here. My limited understanding of entity
beans suggests that if I create an EB using a particular
key value, as long as I refer only to that same key value there
would only be one instance of the EB. Is that not true because
From: Mark Bernardinis [mailto:[EMAIL PROTECTED]]
It seems like what you want is either a SLSB which never
times out and
is guaranteed to only have one instance in the pool, or a BMP entity
bean with a guarantee of serialized transactions.
This is exactly what I want to do. The only
I've found that it's pretty hard to beat Hypersonic for development.
The only caveat is that you need to run it in "server mode" in order to
get concurrent access to the database; otherwise the database files are
locked by Orion and you can't see what's going on.
As far as Interbase goes, there
Has anyone else noticed that 1.4.5 seems to neglect changes made to the
ejb-jar? It unpacks the ear and deploys changed war files just fine,
and if I change the ejb-jar.xml, it will unpack the ejb jar. But if I
change only the ejb class files, Orion just upacks the ear and sits.
Even restarting
The extra table is definately unnecessary, and hopefully it will go away
in a future version of Orion. Take a look at bug #209 in Bugzilla.
Jeff Schnitzer
[EMAIL PROTECTED]
-Original Message-
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]
Sent: Tuesday, January 23, 2001 12:14 AM
Me too. Not only are the joins expensive, but it's a lot of extra data,
too. My database is already key-heavy without the extra tables.
Jeff
-Original Message-
From: Dumitru Sbenghe [mailto:[EMAIL PROTECTED]]
Sent: Tuesday, January 23, 2001 8:04 AM
To: Orion-Interest
Subject: RE: O/R
like this, I'm sure
the added comments would be helpful :-)
Jeff Schnitzer
[EMAIL PROTECTED]
-Original Message-
From: Daniel Cardin [mailto:[EMAIL PROTECTED]]
Sent: Tuesday, January 23, 2001 4:44 AM
To: Orion-Interest
Subject: RE: E_Roman e-commerce application(Mastering Java Beans)Wiley
You could put jsp:forward page="c?param=1" in the default.jsp.
Jeff
-Original Message-
From: Johnson, Robert [mailto:[EMAIL PROTECTED]]
Sent: Thursday, January 18, 2001 8:52 PM
To: Orion-Interest
Subject: Best way to redirect root "/" URL to servlet with parameters
Is there an easy
I believe some browsers don't issue a referrer header.
You could always store the page history in the http session. At the top
of every JSP (using an @include or custom tag) you could have code like
this:
%
String lastPage = session.getAttribute("lastPage");
String currURL;
Is it really necessary to have your own user manager?
I model my users as entity EJBs. I use the vanilla
DataSourceUserManager mapped to the same table that the entities are
written to. All manipulation of user data is done through the entity
beans. The only trick is that you need to set:
Just a thought, but have you checked the system clocks on the various
machines involved? Maybe the timestamps are fouled.
"reloading" of JSPs works fine for me; my deployment process just copies
the new .ear over the old one.
Jeff
-Original Message-
From: Ozzie Gurkan [mailto:[EMAIL
Have you successfully installed the sample junitee test in the JUnitEE
tutorial?
You shouldn't need to modify the global application.xml. Here's how my
system is set up, given that my ear contains two wars, "similarity-web"
and "similarity-test-web" (which contains my TestServlet and JUnit
this is (yes I'm a little bitter). My modifications
are labeled with my name (Jeff Schnitzer).
This modification may or may not be necessary for you; it's obvious from
reading the mailing list (and the code) that nobody working on the JDBC
driver has any clue what PostgreSQL is supposed to send
From: Darren Pamatat [mailto:[EMAIL PROTECTED]]
I like the layout of the complex-or example more so
than the orion-cmp primer, because it is not required
to jar it up, and classes can just be recompiled, and
run (if only I could run something against it). Also,
what is the difference in the EJB
Out of curiosity, why do you have so much logic in servlets that you
need to test them? Is your page transition logic that complicated?
I have found that by keeping all my business logic in session beans
(which are tested with JUnitEE, http://www.infohazard.org/junitee), my
servlets/JSPs stay
m VHS tape.
There is also an AWT replacement called Pure Java AWT from
www.eteks.com, but I'm scared of it. That their website is in French
only amplifies my terror. Has anyone succesfully used it with JDK1.3?
Gr.
/rant
I feel better now. At least, I feel better pretending that I felt
better
I've attached a buildfile (not the orion-primer buildfile).
I wanted the source to mirror the jar file structure (I think this is
less confusing) and I didn't like the proliferation of dependencies in
the orion-primer example. This is the result.
Jeff Schnitzer
[EMAIL PROTECTED]
-Original
a
mangled deployment descriptor, and the app server promptly sees fit to
drop half the tables in the production database. Hey, it worked in the
test system! :-)
Jeff Schnitzer
[EMAIL PROTECTED]
value="Advanced"
% if (course.getLevel() == LEVEL_ADVANCED)
out.print("checked"); %
/
Jeff Schnitzer
[EMAIL PROTECTED]
-Original Message-
From: Kemp Randy-W18971 [mailto:[EMAIL PROTECTED]]
Sent: Friday, December 15, 2000 8:24 AM
An interesting graph can be found here:
-Original Message-
From: Scott Stirling [mailto:[EMAIL PROTECTED]]
Sent: Wednesday, December 13, 2000 9:02 PM
To: Orion-Interest
Subject: RE: www.orionserver.com down again
That's because it was put back up in the interim. I've noticed these
An interesting graph can be found here:
http://uptime.netcraft.com/graph?display=uptimesite=www.orionserver.com
(sorry about the previous chopped message)
Jeff Schnitzer
[EMAIL PROTECTED]
-Original Message-
From: Scott Stirling [mailto:[EMAIL PROTECTED]]
Sent: Wednesday, December 13
to the test case - although
why that would be useful is somewhat of a mystery to me, since IMHO
important logic should all be in EJBs :-)
Jeff Schnitzer
[EMAIL PROTECTED]
-Original Message-
From: Jason Rimmer [mailto:[EMAIL PROTECTED]]
Sent: Tuesday, December 12, 2000 12:13 PM
To: Orion
same bean instance.
Jeff Schnitzer
[EMAIL PROTECTED]
-Original Message-
From: Keith Kwiatek [mailto:[EMAIL PROTECTED]]
Sent: Tuesday, December 12, 2000 4:15 PM
To: Orion-Interest
Subject: Orion JSP form validation and redisplay...
Hello,
Does orion have anything that helps with
Does this mean that all the java packages will be renamed from
com.evermind.* to com.ironflare.* ? :-)
Jeff
From: Karl Avedal
Subject: Re: Anyone heard from evermind?
Date: Thu, 07 Dec 2000 04:15:49 -0800
Hello,
I thought it was time for some explanation on our silence.
Alot of things are
)ejbContext.getEJBObject());
}
catch (java.rmi.RemoteException ex)
{
throw new EJBException(ex);
}
}
Of course you get data duplication in the database, but it works :-)
Jeff Schnitzer
[EMAIL PROTECTED]
-Original Message-
From: Tim Drury [mailto
than likely, though, you have something misspecified in your
deployment descriptor. Make sure you read the comments in Sun's EJB2.0
DTD thoroughly.
If this doesn't help, you can try posting your deployment descriptor.
Good luck,
Jeff Schnitzer
[EMAIL PROTECTED]
-Original Message-
From
-Original Message-
From: James Manning [mailto:[EMAIL PROTECTED]]
Sent: Friday, December 08, 2000 7:08 AM
To: Orion-Interest
Subject: Re: Changing Orion Generated SQL
[Jeff Schnitzer]
Take a look at the documentation for orion-ejb-jar.xml. It's a lot
easier if you have a few samples to look
I have generally found that a ClassCastException (usually nested
three-layers deep in ominous sounding Orion exceptions) usually
indicates a wrong bean (or class) specified in the deployment
descriptor.
For instance, I was tearing my hair out over a ClassCastException I was
getting every time I
Take a look at the documentation for orion-ejb-jar.xml. It's a lot
easier if you have a few samples to look at alongside.
http://www.orionserver.com/docs/orion-ejb-jar.xml.html
What changes you have in mind?
Jeff Schnitzer
[EMAIL PROTECTED]
-Original Message-
From: Aniket V U
There are many other ways of addressing this issue as well.
Good luck!
Jeff Schnitzer
[EMAIL PROTECTED]
e problem anyways).
Another approach might be to do no extra checking and just simply
recreate the stateful session in the error page (if the error is
java.rmi.NoSuchObjectException). H that would be annoying to
the user, so maybe touching the stateful session bean on ever page load
would be the be
r we enter this kind of stuff in Bugzilla or just leave it
alone and wait. Maybe a separate category should be set up for EJB2.0
bugs?
Jeff Schnitzer
[EMAIL PROTECTED]
-Original Message-
From: John D'Ausilio [mailto:[EMAIL PROTECTED]]
Sent: Tuesday, October 24, 2000 8:40 AM
To: Orion-Intere
Hey, there is nothing wrong with MS SQL Server and Win2K. As someone
who is not a professional DBA and has no desire to become one (I like
building things, not maintaining tempermenal third-party software
systems), I find that SQL Server is by far the most pleasant of the big
relational
to
develop our beans for the last several months can see the error in that
statement :-) WebLogic doesn't even have support yet for
container-managed relationships. Geez.
I sent TheServerSide a little note urging them to issue a correction.
It's more likely to happen if they get many more comme
118) is confusing. It says that the
primary key must be set by the end of ejbCreate(), but that CMR fields
must not be modified until ejbPostCreate(). If a CMR field is the
primary key, we seem to have a catch-22 problem...
Jeff Schnitzer
[EMAIL PROTECTED]
-Original Message-
From: Jim
s tag doesn't seem to be
implemented yet. So I want to give a limited security role to "guest".
Heeelp!
Thanks,
Jeff Schnitzer
[EMAIL PROTECTED] mailto:[EMAIL PROTECTED]
:-)
Good luck,
Jeff
-Original Message-
From: Jim Archer [mailto:[EMAIL PROTECTED]]
Sent: Tuesday, October 17, 2000 1:54 PM
To: Orion-Interest
Cc: Jeff Schnitzer
Subject: RE: Please help! I get null pointer exception after calling
createonan EB with EJB 2.0 CMP Orion 1.3.8...
OK
*much* more useful than the chapter in the Wrox JSP book.
Jeff Schnitzer
[EMAIL PROTECTED] mailto:[EMAIL PROTECTED]
*much* more useful than the chapter in the Wrox JSP book.
Jeff Schnitzer
[EMAIL PROTECTED] mailto:[EMAIL PROTECTED]
Why not test from within Orion?
You might want to look at http://www.infohazard.org/junitee
http://www.infohazard.org/junitee for a testrunner, servlet, and
tutorial for using JUnit with Orion.
If you have any trouble with the site, let me know and I'll mail you a
copy of the zipfile.
Jeff
I have had no problems creating CMP entity beans from session beans with
EJB2.0. Seems to work as advertsied :-)
I suggest posting your code and descriptors; I'll take a look, and the
more examples of EJB2.0 beans floating around the better :-)
Jeff
-Original Message-
From: Jim
It's in Chapter 21 of the EJB2.0 Public Draft. Sun has not released the
DTD in any other form, and isn't planning to until the final version of
the spec. At least, that's what the two architects who are writing the
spec said in the (way too short) chat Sun hosted a couple weeks ago.
Has
The tutorial may be a little rough; comments and suggestions are always
welcome.
Jeff Schnitzer
[EMAIL PROTECTED] mailto:[EMAIL PROTECTED]
that Sun needs to take care of?
For the moment I suppose I can rip the class file out of ejb.jar and add
my own EJBException.
Jeff Schnitzer
[EMAIL PROTECTED] mailto:[EMAIL PROTECTED]
article).
For anyone else interested, here's a link to the JavaWorld article:
http://www.javaworld.com/javaworld/javatips/jw-javatip91.html
Thanks,
Jeff Schnitzer
[EMAIL PROTECTED]
-Original Message-
From: wim veninga [mailto:[EMAIL PROTECTED]]
Sent: Saturday, September 30, 2000 9:24 AM
is heading into
unimplemented space, should I log bugs in Bugzilla or just assume that
the authors are busily working on this very problem and don't need yet
another monkey on their backs? :-)
Is anyone else working with EJB2.0 beans?
Thanks,
Jeff Schnitzer
[EMAIL PROTECTED] mailto:[EMAIL PROTECTED]
?), and the output is
a lot more attractive :-) I would also like to enhance it with any
features that anyone thinks might be useful for unit testing EJBs, so
I'm open to any suggestions!
Later,
Jeff Schnitzer
[EMAIL PROTECTED] mailto:[EMAIL PROTECTED]
-Original Message-
From
://www.infohazard.org/junitee
great, but the link seems to be wrong. could you check?
robert
Enjoy :-)
Jeff Schnitzer
[EMAIL PROTECTED] mailto:[EMAIL PROTECTED]
I apologize if anyone is receiving
Doesn't the use of the Address dependent object in the Person remote
interface violate section 9.4.11 of the EJB2.0 spec? Specifically:
-
The Bean Provider must not expose the dependent object classes or the
persistent Collection classes that are used in container managed
relationships
als.xml implies that it is a user named "anonymous", but this
doesn't seem to work. Is this a bug? Do I win a doughnut?
Jeff Schnitzer
[EMAIL PROTECTED] mailto:[EMAIL PROTECTED]
--- Original Message ---
Hi folks.
I'm struggling with Orion's security model in an attempt to get
he
security-role-mapping of role "users" to group "users". The master
principals.xml has both "anonymous" and "guest" as users which belong to
the "users" group.
Both the atm and news demos work fine. I presume it is because they
explicitly auth
101 - 169 of 169 matches
Mail list logo
