With help from several people (thanks, people ;-), I was able to figure out how to get Hibernate to play nicely with Ant. Eric Burke has convinced me that the problem is on the Hibernate side; you can read the whole story at

http://pyre.third-bit.com/blog/archives/000110.html
http://www.ericburke.com/blog/2004/10/s ... oblem.html
http://pyre.third-bit.com/blog/archives/000112.html
http://pyre.third-bit.com/blog/archives/000114.html

Thanks,
Greg

You know what kills me? People who balk and balk and balk about whose fault something is. In their attempts to determine this "blame" they invariably find workable solution(s). Yet, even though all the products being mentioned here are open source, I'm pretty willing to bet not one of these "blamers" submited a single patch to help fix it. Instead, they'd just rather blog and seem swoon by "academic woe".

Welcome to the next open source movement, I guess...

I totally agree with steve and you know what kills me too? gvwilson is not going to help anybody else here... just will ask and ask and ask and blame and blame about 0.002% part of free tools....

Hope all your codes will be bugfree!!!!

Uhm, if a solid personality like Steve says this, there is something going on. But I agree, I don't like that whole approach either.

Greg, this is not how it works. We have a bug tracking system and if you have verified a problem (I mean really verified, we get tons of bogus reports), please post it there, not on your blog. If you want to help more, post a patch.

Greg sent Gavin and me a private email months ago and asked if we would review some of the work he has done for his students, using Hibernate. I agreed to spend some hours on it, once it is finished and my time allows it.

I didn't expect to be involved in finding trivial bugs and fix Ant problems, so I didn't react to anything else that happened (some more private emails).

It is time to bring this back on normal track: complete and verified bug reports will be considered, we help on the forum if possible. Nothing else.

Wow. Wonder how all those apps I wrote using Hibernate actually worked then, considering I did not even have a hibernate.properties file, just the hibernate.cfg.xml? Just lucky I guess?

Yes, thats my lame attempt at sarcasm :)

So you want a hint at what you are doing wrong? Maybe check out the javadocs for the Configuration class; possibly the various configure() methods in particular.

Or I guess you could just continue to blog more blame and criticism of other people's hard (and freely given) work based on ignorance rather than educate and expand yourself. And someone said you are a professor? Wow!!

Hi everyone. Thanks for your replies; in answer to a few points you've raised:

* I'm blogging because I think it's important for my students to see that grownups have these problems too---right now, many of them are feeling that it's their fault, that they must be stupid because they can't get this stuff to work right away, etc.

* A bug report is on its way: I wanted to make sure I understood what was going on before filing it.

* If you think _these_ posts are grumpy, you should hear what I told the Eclipse developers about their UI... ;-)

Seriously, though, it has taken most of my students 30-40 hours to get their first trivial app running using Hibernate, Tapestry, Tomcat, Eclipse, Ant, and JUnit. How does this compare to other people's experiences? I know it takes a lot less time the second (third, fourth...) time around, but when we try to teach this stuff at the undergrad level, each student only goes around once, and so will always spend the bulk of her/his time on the steepest part of the learning curve. What can I do/should I have done four weeks ago to help them ramp up more quickly?

You mention that you are trying to get a "trivial app" up and running. So what is the scope of this course? Because then you go on to mention Tapestry and Hibernate, each of which could be looked at for a semester by themselves to understand all their intracacies. And what's more, each of them solve fairly complex problems but may themselves seem "overly compex" if you do not understand the problems they help you solve/avoid.

I've had the luxury in my career that my experiences evolved along the same lines as language movements themselves. So starting of with COBOL, moving into pure RDBMS programming, next into C++ and Delphi and C/S programming, and finally into Java. Essentially, this allowed me the vantage point of seeing how "things used to be done" and more importantly why they are done differently now. Even within my Java experience, I've done a lot of the low-level stuff that really make you appreciate things like Hibernate and Tapestry that are probably simply lost on those that have not had to implement systems at those lower levels (and maintain them!). Hibernate mappings, for example, assume a fair understanding of relational modeling concept; Tapestry requires understanding of the various state transitions within a request/response cycle.

So, again, I think it comes back to the question of the scope for this given course. If its meant as an introduction to Java, I'm not so certain that jumping right into things like Hibernate and Tapestry is necessarily the right way to go.

Now if its meant as an intermediate-type course, and time permits, I would think the more beneficial approach would be to build a simple web app using Eclipse, Ant, and Tomcat using direct JDBC and JSP; then some assignements that highlight the instrinsic failures/pitfalls of this approach along with introductions of Hibernate and Tapestry to see how such tools help solve these issues.

As for ways to decrease the "ramp up time", that kind of depends. All I've ever seen from you or Eric is a few blogs about Hibernate and Ant interaction (the "non-reading" of config xml is a user error :) Was there something else I am missing?

Hi Steve; thanks for your post. You're right, I should clarify. Our long-term goal is to build a simple clone of SourceForge using popular Java web app tools (Hibernate, Tapestry, Tomcat, Eclipse, Ant, JUnit, Checkstyle, etc.). The goals are:

- to allow instructors to manage undergrad group programming projects more easily

- to give students a feel for grown-up working practices (projects in industry aren't managed by mailing tarballs around; they're managed using version control and bug tracking systems, so student projects should be too)

- to create a "Minix for the 21st Century", i.e. a working example that can be taken apart in courses to show students how real systems work.

Some students made a start on the system over the summer (see http://pyre.third-bit.com/hippo); this term's students (http://pyre.third-bit.com/neon) are carrying on with their work, or will be once they can find their way around our existing code.

That's the big picture. Right now, I'm trying to create a rational re-implementation of core portions of the data model (users, user groups, projects, and authorization) to give next term's students as a starting point. I'd be happy to make the code available to anyone who wants to look at it---assuming anyone here is still talking to me ;-)

(gavin and christian shudder :) I'd suggest using Maven rather than Ant especially if you are looking for a more sophisticated way to manage projects and artifacts.

I see your already using Subversion...

I'm kind of confused. B4 you said you did not want to just "Hand students a working self-contained starting point". But then you say that this is exactly what you do do (this semesters students start where last semesters students left off). And besides, thats how real world development happens; people come and go...

Sorry, don't have time to look over your code; too busy creating the next generation of open source :) Now if you wanted to contract with JBoss for some production or development support, then my day job would entail me taking a look at it :)