Don Schaefer

Greenhorn
+ Follow
since Jan 22, 2002
Merit badge: grant badges
For More
Cows and Likes
Cows
Total received
0
In last 30 days
0
Total given
0
Likes
Total received
0
Received in last 30 days
0
Total given
0
Given in last 30 days
0
Forums and Threads
Scavenger Hunt
expand Ranch Hand Scavenger Hunt
expand Greenhorn Scavenger Hunt

Recent posts by Don Schaefer

Thanks for your response! "JavaScript: the Good Parts" is a book that I use for my work and think it is a fine book. Its discussion on modules is a good one when consider good practices. Likewise I use JSLint from within Textmate!

Another source I follow is the google group - JSMentors.

Thanks,
Don.

Looking (briefly) at the TOC for the book it is not clear if there is a sections on best practices, debugging, code organiztion etc. Lately I have been using JSLint for improving code and would like to know what else people use.

Thanks!
I have done rails work in the past. How much of a learning curve is in store for me as I look into merb? If the learning curve is steep I might consider a more barebones approach like Sinatra, especially for simpler site work.

Thanks,
Don.
14 years ago
Looking at the contents and info for Jeff's book on the manning site, it lists a number of open source projects that can solve parts of a SOA stack. Very kewl. Many folks assume that open source == standards based. This is not always true. My question is this:

How will a collection of open source implementation move me towards a more platform agnostic solution? The book "SOA design patterns" by Thomas Erl re-enforces this goal again and again.

As an aside, I think that this is a very difficult task to attain. The code base by the use of abstraction and interface has a chance at meeting part of this goal but the enterprise processes (monitoring, security, set up tear down, configuration, etc) have a near impossible task...

Thanks,
Don.
14 years ago
The approach we have generally taken for testing web services has been to write the tests on the pojos used by the webservice instead of the services themselves. The impl code merely calls the pojo with the document after applying security and transaction policy. This allows separation between the webservice implementation and our logic. This also allows for much quicker unit testing.

If you want to be a purist about it, testing through the webservice is more of an integration test. For those tests you also have to account for security, transaction policies etc. Much more that a simple junit test. For that testing we build the docs, execute the service and check the results using relative xpath expressions.

Thanks,
Don.
14 years ago
What error do you get? Can you show a stack trace? How is the bean configured related to transactions? Are you using WebSphere's connection pool or Oracle's? Any other details? As a test, can you run invoke the query directly from your servlet?
Weblogic has a multi-protocol listener that distributes requests based on the selected protocol. The primary protocol is t3. I am not sure if this is from the tengah legacy but I believe that is the case.

this url has some information on it (there are others)
Weblogic Documenation
21 years ago
Are you getting any exception when you fail?
I know that the jsp compiler changed from in one of the later 5.1 sps.
Without any additional diagnostic its hard to help.
don.
21 years ago
I don't know the answer but I'd guess that your web.xml does not meet weblogic's expectations. Its looking for an entry in your web.xml and not finding it. That would explain the string out of bounds....
Good luck,
don.
21 years ago
There are a couple of differences although the most significant is in the order of execution.
You need to implement "CustomService" not the T3 interface. CustomService is found in import com.ibm.websphere.runtime.
Unlike in WLS you start classes are fired immediately. Meaning there are no container services available to you when the service is run. If you service needs any j2ee services, it will need to delay until they are available.
JBoss has it done very well. They provide more then one entry point. Basically before all the j2ee services are available and after. I wish IBM and BEA would do the same.
21 years ago
I am confident that this is a two part call. First you must get a context. While getting the context you will perform an authentication. Depending on the identification that you authenticate with you may be able to retrieve information. Depending upon the sub-context you are retrieving you may be able get the data you are looking for in a single call.
I am not sure if I helped....
don.
21 years ago
I'll admit that this is a real reach but could you be picking up a different rmic than you are expecting? Say from another jdk installation?
Like I said, its a reach.
21 years ago
The timer class works well for us as well.
As to start with unix why not invoke your script from rc?
Also some application server products have their own timer implementations exposed that can be useful. Watch out here though. They are proprietary.
What out also with programs with shutdown hooks in the background. You will have to nohup these to prevent the hook from firing prematurely.
You can get a behavior similar to rc with a product like:
http://www.eworksmart.com/JNT/
21 years ago