java.net.UnknownHostException: Invalid address: fe80::1d9:b65a:ed86:7940%11

Not being much of a networking expert I was puzzled about the ‘%11′. Once again superuser.com came to rescue:  http://superuser.com/questions/99746/why-is-there-a-in-the-ipv6-address.

“I think Java became a success despite Sun’s corporate culture. Priorities at Sun changed quite rapidly from time to time. At Oracle it takes longer to make a plan but then with stick with it” (something along the lines of that)

While Henrik Ståhl showed at least some passion and inspiration (neatly hidden underneath the Scandinavian cover) Cameron Purdy seemed to be nearly falling asleep himself flipping through his overloaded Oracle marketing slides. That he brought a colleague on stage who demonstrated the “cloud features” – gosh, how I hate this over-hyped cloud stuff – of the upcoming Glassfish 4.0 to him, instead of the audience, was seriously odd.

Tim Bray might be as high profile at Google Android as Ståhl/Purdy are at Oracle but there was pure passion pouring out of his mouth. Passion for software, passion for code, passion for improvement – and he spoke for himself and much less for Google*. He was vibrant and inspiring. Also, he was very humble and particularly respectful towards the Android competitors. Not a single rant or joke about iPhone or Windows Mobile 7 which he expects to catch up to Apple/Google pretty soon.

Tim Bray @Devoxx 2011:

1. “Information wants to be free.”
2. “The absence of women is the elephant in the living room. It must be discussed.” on the lack of women @Devoxx showing Banksy‘s famous red elephant photo
3. “If you are not working on mobile you are away from the action.”
4. “…wouldn’t you rather feed the poor, cloth the naked, cure the sick?” on why it’d be so much more important to develop apps for people in the third world rather than for us westerns who lead a lifestyle that lacks nothing (mobile Internet access dwarfs “cable-based” access particularly in underdeveloped countries)
5. “The importance of static type checking is proportional to the number of APIs your coding against, but inversely proportional to easyness of unit testing.” or so

* I’m neither an Android developer, and I don’t intend to become one, nor is it the mobile platform of my choice. I’m all Apple

Embedded error: Error performing commands for file transfer
Exit code: 0 - bash: unzip: command not found

for the longest time. I tried all kinds of funny things although I was 100% certain that I do have unzip in path. I was a bit embarresed when I finally realized that this error message was the result of the command executed over SSH on the remote host (Debian Linux) – the one the site was deployed to.

After that it was a piece of cake. Just had to install unzip on Linux using APT: apt-get install unzip

javax.persistence.PersistenceException: [PersistenceUnit: spontacts] Unable to configure EntityManagerFactory
	at org.hibernate.ejb.Ejb3Configuration.configure(Ejb3Configuration.java:265)
	at org.codehaus.mojo.hibernate3.configuration.JPAComponentConfiguration.createConfiguration(JPAComponentConfiguration.java:28)
...
Caused by: java.lang.ArrayStoreException: sun.reflect.annotation.EnumConstantNotPresentExceptionProxy
	at sun.reflect.annotation.AnnotationParser.parseEnumArray(AnnotationParser.java:673)
	at sun.reflect.annotation.AnnotationParser.parseArray(AnnotationParser.java:462)

Once I realized what the actual cause is, the fix was simple. Turns out that the CascadeType is not in the classpath when Maven runs the hbm2ddl goal. Hence you need to add a dependency to JPA to the hibernate3-maven-plugin:

<plugin>
	<groupId>org.codehaus.mojo</groupId>
	<artifactId>hibernate3-maven-plugin</artifactId>
	<dependencies>
		<dependency>
			<groupId>org.hibernate.javax.persistence</groupId>
			<artifactId>hibernate-jpa-2.0-api</artifactId>
			<version>1.0.0.Final</version>
		</dependency>
	</dependencies>
	<executions>
		...
	</executions>
	<configuration>
		...
	</configuration>
</plugin>

A solution to this exception in a different context was posted here: http://javahowto.blogspot.com/2008/10/solve-javalangarraystoreexception.html

Today I learned that Hibernate has an extra-lazy mode that detects such cases. It issues the proper “select count(id) from A” SQL statement instead of “select … from A” in the background: @LazyCollection(LazyCollectionOption.EXTRA)

The enum’s JavaDoc is very terse but there’s a nice article here: http://sites.google.com/a/pintailconsultingllc.com/java/hibernate-extra-lazy-collection-fetching (also stresses the difference of list vs. bag semantics). If you happen to have a copy of “Java Persistence with Hibernate” you’d find this in chapter 13.1.3 on page 567.

I recommend sticking with the sample application at http://repo1.maven.org/maven2/org/springframework/security/spring-security-samples-openid/<your_version>/spring-security-samples-openid-<your_version>.war as a reference. Also, I found http://www.packtpub.com/article/opening-up-to-openid-with-spring-security quite helpful.

One caveat though is worth mentioning here: the Spring sample application does not work with Google and Yahoo OpenID (and probably many more). The reason is explained in the above article, just scroll down to “The OpenID user registration problem”.

For my sample application I didn’t implement registration just yet. So, when I used Google’s generic OpenID URL https://www.google.com/accounts/o8/id for the input field in openidlogin.jsp I was able to authenticate against my Google account but Spring complained as none of the user names listed in applicationContext-security.xml matched what Google returned. Quick-fix: check your application’s log file for “Verification succeeded for: https://www.google.com/acc”….That URL is the effective user name you should paste to applicationContext-security.xml.

I asked my myself: where the heck is this place? What’s the story behind that? My first guess was northern Scandinavia. Turns out I wasn’t quite right but not very wrong either.

I compared it with various coastlines on Google Maps and found a matching spot in northern Iceland. I zoomed closer and bingo! in the center of the logo is a town called Dalvik. Now it all makes sense. There’s even a few words about the link between the town Dalvik and Android on Wikipedia.

This is yet another idiotic WebSphere behavior that you need to work around. Set

com.ibm.ws.webcontainer.invokefilterscompatibility=true

in Servers -> Server -> Web Container Settings -> Web Container -> Custom Properties.

IBM has more information in their custom properties documentation (search for “Invoking the filter capability”) and on the respective support document (for fix pack 6.1.07). Since the URLs to those documents might become invalid in the future I persisted the latter in a PDF.