Eclipse 3.2RC1 quite stable

April 18th, 2006

After a personal desaster with M6, the RC1 proved to be quite stable. I know it is idiotic to use always the bleeding edge, but I couldn’t resist (there had been something in the relnotes that looked too tempting to me).

To make a long story short: I wasted nearly whole of last week restarting eclipse and reinstalling JREs (M6 didn’t liked Mustang at all), but with RC1 under 1.5.06 no crash during the weeked using WST 1.5M4. OK, I restarted twice because WST got confused about the state of my Tomcat, but that was OK.

I think. It was the first time I started right out with WST a J2EE application, and hey I was an awesome experience. When using WST on an existing “J2EE project” - that is a project not created with J2EE you’ll miss many features and some things simply don’t work properly (at random, sometimes:-).

I wrote a simple Ajax app with it (actually without any XML, but that seems to be the trend, not?) which was targeted at a PHP deployment. I wanted to learn the SAJAX for the client-side so I decided to do it J2EE-wise where I am 100% sure what I am doing (and that it is also done by the machine - I hate magic, unless I do it myself).

Well, just aside: The PHP implementation took me much longer than developing the Java code!

SAJAX on the other hand is quite nice, when I find some time (so many interesting thing to do, so little time) I’ll contribute something to leverage it from J2EE more easily.

Roller on Glassfish - or the beginning of the end

February 18th, 2006

So, it seems that it works.
I used the JdbcsRealm from the glassfish bug.parade (written by bjb), tweaked a bit antlr, etc and I can show my blogs. But not for long. After 8 or so calls the server hangs - perhaps it has something to do with the encoding handling or whatever.
Deploying on an appserver became more complicated that making C-code going on UNIX. Tomcat/MySQL define the “Windows” here: Everything works (somehow, not as expected) and glassfish reminds my here of the the early UNIX: The aimed for MULTICS, but ran out of money. ClassLoader-problems - library conflicts strange interdependencies between Security-Layers - give me a break. This stuff sucks. Perhaps it is Roller’s fault and I should have tried something other before (I did, and all worked fine). The point is that you have to “buy” a new appserver for a new app, while keeping the old one for the existing ones.
Sounds familiar? Yes this is what we all laughed about about in stories like when Microsoft was an airline, etc.
Same thing with “browsers-based”. I spoiled some weekends to make a simple site look right in different browsers. After getting it right (with some dirty tricks) on Firefox, Opera and IE6&7 I learned that Safari had another understanding (after reading the specs I have to admit that all other browser seem to get it wrong there).
To sum up I’d say that computing actually got fucked up during the last 10 years.
BTW: Does anyone still remember XWindows? If think implemented the todays needs very nicely
PS: I beg for an excuse that I didn’t spice up this post with some links, see next post for the reason of this

Old Apache Java

January 19th, 2006

Apache’s stuff became a classic - especially in the Java-universe. Lot of nice stuff found even its way into Sun’s JVM (BCEL, Xerces), but what about the rest?
Commons is beginning to get on my nerves now. We have - as many people - problem with the memory-leaks and caused by commons-logging, now I stumbled into something in HttpClient/Axis which sets finally some red lights on Apache stuff.
It feels for me that Apache becomes a J#. Perhaps a little bit harsh, but since I also have to keep code-compatability with J# I know what I am talking about when I say that it feels the same.
Apache gave us nice things like their licensing-scheme, the Apache server and many reference implementations of JSRs. But Apache is in the end nothing but a brand. Sometimes it seems that more or less weak ideas try get promoted by the blessing of an Apache-branding.