I’m in a training class all week, leaving little time for blogging about the things I’m learning as I learn them. For my reader, please check back next week, in time for the holiday 😉
Oracle JDeveloper 11g: RTFM
After so many years of finding better answers through Google or Dev2Dev (moment of silence), I rarely use the help menu of applications anymore. Yesterday I was having trouble with one of the JDeveloper wizards. I searched Google, I searched OTN, I posted to discussion lists. No answer. Desperate to get past this issue, I opened the Help menu. I still didn’t find what I was looking for (I solved that by experimentation and guess work), but I did find all of the training material listed right there.
This would have saved me days of navigating around OTN, where I did not find the training I needed that was right in the IDE all the time.
Interesting lesson learned, and the first application of the alternative definition of RTFM I read recently: Read The Fancy Manual
Is There a Prize with That?
I just started answering questions on LinkedIn lately, and got a Best Answer yesterday on a question about Doubt in handling Exceptions in Java*. Before the Internet, everyone had 15 minutes of fame. Now it is more like 15 nanoseconds 🙂
*Requires LinkedIn account to view
WLP Sessions and Porlets in Shared Libraries
Awhile back, I ran into an issue where session data was being lost between requests in a WebLogic Portal project. As it only occurred in a clustered environment, it was obvious that the value was not being persisted, even though persistence was set properly with
<wls:session-descriptor>
<wls:persistent-store-type>replicated_if_clustered</wls:persistent-store-type>
</wls:session-descriptor>
After much frustration, someone on the team ran across an additional setting for the session-descriptor node: sharing-enabled. In the case where a portlet is from a shared library, the persistence does not work unless sharing-enabled is set to true.
<wls:session-descriptor>
<wls:persistent-store-type>replicated_if_clustered</wls:persistent-store-type>
<wls:sharing-enabled>true</wls:sharing-enabled>
</wls:session-descriptor>
In hindsight, this behavior makes some sense, in that most portlets that are distributed as shared libraries by vendors do not require session values. However, in the case where a shared library is created by an application team to reuse portlets across portals where WSRP is not practical, the need to share sessions may arise, and now you know how to fix it.
For other deployment descriptor options, see http://download.oracle.com/docs/cd/E12840_01/wls/docs103/webapp/weblogic_xml.html
Good Thread on White Space in JSPs
While it is tempting to quickly skim an blog entry for the information I’m seeking at the time, a recent find on dealing with white space with JSPs reminded me that there is a darn good reason most blogs have a comment feature.
The thread I’m talking about is Trim Spaces in Your JSP at Raible Designs. Fortunately, my paranoia paid off, and I did not try this setting blindly on the project where I was looking for such an answer (and have since rolled off). I almost did, but decided not to because I knew I would not be around to support the change, and that whoever I told about it may forget it. As it is a performance enhancement and the application had some peformance issues (too many cooks in the presentation layer kitchen IMHO), it was not as easy of a call as it may seem.
The con that comes up in the comments on that post is that it wipes all white space between tags. If I were the only person developing on the project, or if was a small team where I knew for certain that only experience JSP developers were working on the JSPs, I’d still feel safe. This is because if a space has to be in a page between two dynamic parameters, the experienced JSP developer will use an entity code rather than just a blank space.
There is also a comment in the thread about how to cut down white space in a build file. That was actually the type of solution I was seeking at the time (and missed at the time). I would have to experiment with the solution before I suggested it, but it looks like what I wanted.