If you have used Hibernate and mapped a one-to-many relationship you've probably come across the "delete orphan" feature. This feature offers cascade delete of objects orphaned by code like the following:

Preference pref = getSomePreference();
user.getPreferences().remove(pref);

In the above code, a specific Preference is removed from a User. With the delete orphan feature, and assuming there is an active transaction associated with a session, the preference that was removed from the user is automatically deleted from the database when the transaction commits. This feature is pretty handy, but can be tricky if you try to write clever code in you setter methods, e.g. something like this:

// Do not do this!
public void setPreferences(Set newPreferences) {
    this.preferences = newPreferences == null ? new HashSet<Preference>() : newPreferences;
}

Code like the above results in a HibernateException with the following message if you pass null into setPreferences and try to save the user object:

A collection with cascade="all-delete-orphan" was no longer referenced by the owning entity instance

What is happening here is that Hibernate requires complete ownership of the preferences collection in the User object. If you simply set it to a new object as in the above code sample, Hibernate is unable to track changes to that collection and thus has no idea how to apply the cascading persistence to your objects! The same error will occur if you passed in a different collection, e.g.:

user.setPreferences(new HashSet<Preference>());

So the point is that Hibernate's delete orphan abstraction is leaking into your domain model object. This is pretty much unavoidable but is a leaky abstraction nonetheless that developers need to be aware of lest they run into the error mentioned above.

So how can you avoid this problem? The only sure way that I know of is to make the setter method private, since passing any new collection or null results in the "owning entity" error. This way only Hibernate will use the setter method to load up user objects (it invokes the method reflectively after setting it accessible via the Reflection API). Then you could add a method addPreference to your code which is the public API for adding preferences. Anyone could of course use reflection to do the same thing Hibernate is doing, but then all bets are off as they are subverting your public API. For example:

public void addPreference(Preference p) {
    getPreferences().add(p);
    p.setUser(this);
}

This has the nice side effect of establishing the bi-directional relationship between user and preference, assuming your model allows bi-directional navigation. You could also add a null check if you are paranoid. Removing a preference from a user is equally simple. You can write a helper method removePreference or you could call the getter and then call remove as shown here:

user.getPreferences().remove(aPref);

Essentially, you can operate on the collection returned by getPreferences as a normal Java collection and Hibernate will do the right thing. Since you are operating on the collection maintained and observed by Hibernate, it is able to track your changes, whereas replacing the collection wholesale makes Hibernate really mad, since it believes it is the sole proprietor of the collection, not you! For example, if you want to remove all the user's preferences you could write the following:

user.getPreferences().clear();

Note that all the above discussion refers to one-to-many relationships that specify the delete orphan cascade type; usually you will specify both "all" and "delete-orphan." In cases where you are only using the "all" cascade option, the semantics are quite different. Assuming the normal case where the "many" side of the relationship owns the relationship -- i.e. you used inverse="true" in your mapping file or @OneToMany(mappedBy = "user") if using annotations - then you must explicitly delete the child objects as Hibernate will only track that side of the relationship. For example, if cascade is "all" and you remove a preference from a user and then save the user, nothing happens! You would need to explicitly delete the preference object, as shown here:

// Assume only using cascade="all" and an inverse="true" mapping in User.hbm.xml
user.getPreferences().remove(aPref);  // does not cause database delete
session.delete(aPref);                // this causes the database deletion

One last thing to note in the above is that you must remove aPref from the user, or else Hibernate will throw an exception stating that the preference would be re-saved upon cascade! So if the User object is in the Session, remember you need to undo both sides of the relationship for things to work properly.

For anyone who uses PMD, the title of this blog appears in their list of PMD errors if they don't declare their loggers static and final. Specifically, the LoggerIsNotStaticFinal rule simply says that a log should be declared static and final. I also like to make sure they are private as well. For example:

// Jakarta Commons Logging
private static final Log log = LogFactory.getLog(MyClass.class);

The above code also shows another good practice, which is to pass the Class object to the getLog() method, instead of a string. Why the java.util.logging.Logger class doesn't even provide a method accepting a Class object is simply beyond me. Why did the people who developed the java.util.logging package base their API on Log4j yet omit some of the most useful parts of it? Oh well.

Now to the point. Why it is good practice to declare loggers private, static, and final? A logger is an internal implementation detail, so it should be private. You only need one logger for all instances of a class, hence static. And a logger should not be able to be replaced, thus final. So if this is good, what's not so good (at least in my opinion)? Simple - any logger that is not private, static, final, and which doesn't pass in a Class object to getLog()! For example, consider this common bit of code, declared in some base class:

// Not so good logger declaration
protected final Log log = LogFactory.getLog(getClass());

Why is this bad? Well, it isn't static for one thing. For another, it uses getClass() to obtain the log. At first this seems efficient since now all subclasses automatically inherit a ready-made log of the correct runtime type. So what's the issue here? The biggest problem with loggers declared in this manner is that you now get all the logging from the superclass mixed in with the logging from the subclass, and it is impossible in the log output to discern which messages came from which class unless you look at the source. This is really annoying if the superclass has a lot of logging that you don't want to see, since you cannot filter it out.

Another problem is that your ability to set log levels differently goes away, for example if a subclass resides in a different package than the superclass. In that case, if you try to filter out logging from the superclass, you can't because the actual runtime class was used to obtain the logger.

Last, having a protected logger just seems to violate basic object-oriented principles. Why in the world should subclasses know about an internal implementation detail from a superclass that is a cross-cutting concern, no less? Anyway, though this is a silly little rant it really is annoying when you extend a superclass that declares a protected logger like this.

Fun things happen in Java when you have the same class in two different places in your CLASSPATH. For example, maybe your application server has decided to bundle an open source package in their product but has not repackaged it, which can cause hours of fun debugging the problem. Another common example is, or at least used to be, XML parsers. Anyway, if you ever need to find out exactly where a class is being loaded from, you can plop a few simple lines of code in your application somewhere, run it, and Voila! you now at least know which JAR is causing you headaches which might help you figure out how to solve your problem.

The only reason I remembered how to do this was because Stu Halloway showed this trick in a No Fluff Just Stuff session on class loading years ago. The code is pretty simple, and here it is:

Class aClass = EmbeddedDriver.class;
ProtectionDomain protectionDomain = aClass.getProtectionDomain();
CodeSource codeSource = protectionDomain.getCodeSource();
if (codeSource != null) {
    URL location = codeSource.getLocation();
    System.out.println("location = " + location);
}

The output I received when trying to find where the Apache Derby EmbeddedDriver was being loaded from looked like this:

location = file:/Users/sleberkn/Workspace/learning_derby/lib/derby.jar

You could also print out information about the ProtectionDomain and CodeSource for additional information. For example, here's what I got:

protectionDomain = ProtectionDomain  (file:/Users/sleberkn/Workspace/learning_derby/lib/derby.jar )
 sun.misc.Launcher$AppClassLoader@a9c85c
 
 java.security.Permissions@5740bb (
 (java.io.FilePermission /Users/sleberkn/Workspace/learning_derby/lib/derby.jar read)
 (java.lang.RuntimePermission exitVM)
)


codeSource = (file:/Users/sleberkn/Workspace/learning_derby/lib/derby.jar )
location = file:/Users/sleberkn/Workspace/learning_derby/lib/derby.jar

The information contained in the ProtectionDomain can be very useful as it tells you information about class loaders, permissions, etc. Another trick, which maybe I'll write about some other time, is to find out which class loader is loading a particular class. The code is also simple as you basically just walk the tree of class loaders until there are no more. This can aid in debugging JEE applications that package things in EAR, EAR, and EJB-JAR files, as the various class loading schemes of application servers can be, well, annoying to say the least.

I now use mainly Eclipse when coding in Java and hate when I have to use the mouse to perform some task as it normally slows me down. Some of the things I used to use the mouse for included moving between editors and moving between the various task panes, like Package Explorer, Navigator, and Console. One day recently I decided to learn a few new shortcut keys for navigation and some miscellaneous things in Eclipse. Now, these shortcuts are on the Mac but I think the equivalents are a relatively straightforward mapping on Windows, e.g. Ctrl instead of Command/Apple/Clover.

Navigating Between Editors

To navigate between open editors, I've found two options. First, you can use Ctrl + Tab to get into select mode. You then use the right and left arrow keys to move between the open editor tabs. Once you find the one you want, hit Ctrl + Tab again to get out of select mode and back into edit mode.

The second way is using Apple + F6. Repeatedly hitting this keystroke switches between the two most recent editor tabs. If instead you hit Apple + F6 and keep the Apple key down, a small popup window appears with a list of open editors. You can then use the up and down arrow keys to select which editor you want to open. Using either of these key combinations, you never need to use the mouse again to switch editors.

Navigating Between Windows

Many times I want to switch from an editor to Package Explorer, Console or any other view and then back again. I used to use the mouse but now I can switch easily using Apple + F7. This key combination behaves in the same way that Apple + F6 does for switching editors. The difference is now you can select any view in Eclipse. So, if you are editing a Java file and want to go to JUnit view, you can hit Apple + F7, keep holding the Apple key down, and use the up and down arrows to select the view you want. Nice.

Moving and Copying Lines of Code

Now of course no one copies code and pastes it somewhere else anymore, but if you do you might find these key combinations useful. Let's say you need to move one or more lines of code up or down in a source file, any source file. For example maybe you want to move one method below another method, or move a variable declaration. In Eclipse you can highlight one or more lines in an editor and then use Alt + up/down to move those lines of code up or down, respectively. You must continue to hold Alt down while you are moving the code.

If you need to duplicate one or more lines of code (you don't ever do that do you?) you can use the Alt + Apple + up/down key combination to duplicate several lines of code above or below the current location. Once you've duplicated the code, you can release the Apple key and then use the up and down arrow keys to move the code where you want it.

"Quick" Windows

These keyboard shortcuts are really useful while editing Java files. The first is Apple + O (the letter "oh" not a zero). This brings up the Quick Outline view and allows you to quickly find any member in the class. You simply start typing the first few letters of the member name, e.g. "get", and Eclipse will filter out only members whose name begins with "get." You can then select the member you want to navigate to and hit Enter. You can also use the up and down arrows in the Quick Outline view to navigate between members.

The second tip is Apple + T. This keystroke shows the Quick Type Hierarchy view. You select the name of a class, for example Set and hit Apple + T, you will get a popup showing the entire hierarchy of the Set interface. This is really useful if you code to interfaces instead of concrete classes, for example the way you would when building a Spring-based web application. Since everything is defined in terms of interfaces, you can use Apple + T to quickly find an implementation class and go to it. Or, when in a subclass you might want to see the class hierarchy or navigate to a superclass. Apple + T works well in this case as well.

Run in Background

Many times in Eclipse you want something to run in the background, for example when synchronizing with your source code repository. By default Eclipse pops up a dialog that shows the task status. You can always click the "Run in Background" button on the dialog and continue working, but that gets repetitive. So, edit your Eclipse preferences and in the General section, make sure the "Always run in background" check box is checked. That way tasks that require background processing execute in the background without requiring you to dismiss the dialog.

Show Heap Status

IntelliJ has long shown a display of the memory it is currently consuming, and allows you to suggest to the JVM to run the garbage collector. You can do this in Eclipse too, by ensuring that the "Show Heap Status" check box is checked in the General section of the Eclipse preferences. You'll now have a display in the lower right corner of Eclipse that contains the current memory consumption, allocated memory, and a little garbage can icon which allows you to ask Java to collect the trash if you click it.

Today I needed to install Subversion on my Macbook running OS X and was following Dan Benjamin's excellent instructions for accomplishing this task. The main difference was that I wanted to install the latest and greatest (as of today) version of Subversion, which is 1.4.2. Dan's instructions covered 1.3.1 and a co-worker was able to install SVN just fine using those instructions and the older version. I started following the instructions, substituting 1.4.2 for 1.3.1. Everything was going nicely until the configure, at which point I received an error stating "no suitable apr found". The following is the full output.

configure: Configuring Subversion 1.4.2
configure: creating config.nice
checking for gcc... gcc
checking for C compiler default output file name... a.out
checking whether the C compiler works... yes
checking whether we are cross compiling... no
checking for suffix of executables... 
checking for suffix of object files... o
checking whether we are using the GNU C compiler... yes
checking whether gcc accepts -g... yes
checking for gcc option to accept ANSI C... none needed
checking how to run the C preprocessor... gcc -E
checking build system type... i386-apple-darwin8.8.1
checking host system type... i386-apple-darwin8.8.1
checking target system type... i386-apple-darwin8.8.1
checking for egrep... grep -E
checking whether ln -s works... yes
checking for a BSD-compatible install... /usr/bin/install -c
checking for static Apache module support... no
checking for Apache module support via DSO through APXS... no - Unable to locate /usr/include/httpd/mod_dav.h
==================================================================
WARNING: skipping the build of mod_dav_svn
         --with-apxs or --with-apache must be used
==================================================================
configure: Apache Portable Runtime (APR) library configuration
checking for APR... no
configure: WARNING: APR not found
The Apache Portable Runtime (APR) library cannot be found.
Please install APR on this system and supply the appropriate
--with-apr option to 'configure'

or

get it with SVN and put it in a subdirectory of this source:

   svn co \
    http://svn.apache.org/repos/asf/apr/apr/branches/0.9.x \
    apr

Run that right here in the top level of the Subversion tree,
then run autogen.sh again.

Whichever of the above you do, you probably need to do
something similar for apr-util, either providing both
--with-apr and --with-apr-util to 'configure', or
getting both from SVN with:

   svn co \
    http://svn.apache.org/repos/asf/apr/apr-util/branches/0.9.x \
    apr-util

configure: error: no suitable apr found

Oops.

I searched around Google a bit to see if anyone else had encountered (and solved) this problem. Didn't find much so I decided to go ahead and follow the suggestion to install Apache Portable Runtime and then try the configuration again. In case you run into the same problem, here are the modified instructions I used to build SVN 1.4.2 for OS X. Thanks to Dan for the original instructions, and I hope the modified ones will help you out too. Admittedly I did not think too much about exactly why I needed apr and was interested in just getting it to work. So if I've done something stupid by installing apr please feel free to post a comment and let me know!

One last note is that the following commands contain my username, sleberkn, so of course you'll need to substitute your username, or follow Jason's suggestion to use $HOME instead. Before just trying these commands, first read Dan's entire post so you have all the prerequisites, such as setting your PATH correctly!

cd ~/Desktop
mkdir src
cd src
curl -O http://subversion.tigris.org/downloads/subversion-1.4.2.tar.gz
tar xzvf subversion-1.4.2.tar.gz
curl -O http://mirrors.24-7-solutions.net/pub/apache/apr/apr-1.2.7.tar.gz
tar xvzf apr-1.2.7.tar.gz
curl -O http://mirrors.24-7-solutions.net/pub/apache/apr/apr-util-1.2.7.tar.gz
tar xvzf apr-util-1.2.7.tar.gz
cd apr-1.2.7
./configure --prefix=/Users/sleberkn/Desktop/src/subversion-1.4.2
make
make install
cd ..
cd apr-util-1.2.7
./configure --prefix=/Users/sleberkn/Desktop/src/subversion-1.4.2 --with-apr=/Users/sleberkn/Desktop/src/subversion-1.4.2
make
make install
cd ..
cd subversion-1.4.2
./configure --prefix=/usr/local --with-openssl --with-ssl --with-zlib --with-apr=/Users/sleberkn/Desktop/src/subversion-1.4.2 --with-apr-util=/Users/sleberkn/Desktop/src/subversion-1.4.2
make
sudo make install

Assuming all that worked, verify your installation.

which svn
svn --version

The output from the above commands should show the Subversion is installed in /usr/local/bin. My output looked like this.

nic-sleberkn:~/Desktop/src/subversion-1.4.2 sleberkn$ which svn
/usr/local/bin/svn
nic-sleberkn:~/Desktop/src/subversion-1.4.2 sleberkn$ svn --version
svn, version 1.4.2 (r22196)
   compiled Nov 30 2006, 17:47:45

Copyright (C) 2000-2006 CollabNet.
Subversion is open source software, see http://subversion.tigris.org/
This product includes software developed by CollabNet (http://www.Collab.Net/).

The following repository access (RA) modules are available:

* ra_svn : Module for accessing a repository using the svn network protocol.
  - handles 'svn' scheme
* ra_local : Module for accessing a repository on local disk.
  - handles 'file' scheme

Recently I came across a couple of pretty interesting things in Java 5. The StringBuilder class was introduced in Java 5 as an unsynchronized version of the rather ubiquitous StringBuffer class which has been around for ages, and which the beginner Java programmer soon learns about the first time a more experienced Java developer sees his code littered with string concatentation using the '+' operator. The reason StringBuilder was introduced (I am guessing) is that you almost never need synchronization when building up a string, or at least I've never had multiple threads building a string at the same time! So the StringBuilder class is pretty much identical to StringBuffer except its methods are not synchronized.

Another new feature Java 5 introduced is covariant return types, which permits an overriding method to return a more specialized type than the overriden method. Some good examples I've seen include this and this. In a nutshell, suppose you want to override the clone() method for a Person class. The clone() method in Object can only return, well, Object. Prior to Java 5 your overridden clone() method in Person would also have to return Object and a cast would be required. But now you can do this:

@Override
protected Person clone() throws CloneNotSupportedException {
    ...
}

So now there is no cast required since the overriden clone() can return the type you want - Person. This is a nice feature in general, but the thing I came across that was more interesting was the inheritance relationship between StringBuffer, StringBuilder, and AbstractStringBuilder. StringBuffer and StringBuilder both extend AbstractStringBuilder and use covariant return types in overridden methods. For example, here are the definitions of the append(String) methods for these three classes:

public AbstractStringBuilder append(String str) {
    ....
}

public synchronized StringBuffer append(String str) {
    ...
}

public StringBuilder append(String str) {
    ...
}

So what's cool about this is that the parent class defines the append() method to return an object of the parent type AbstractStringBuilder but the subclasses overriding this method can restrict the return type so that no casts are required. StringBuilder's append() method can return StringBuilder and StringBuffer's append() method can return StringBuffer. I found it interesting that the methods were not marked with the @Override annotation - maybe that would have been a gargantuan task to retrofit all existing code using the annotation?

But even though the covariance is cool, what was even more interesting is that AbstractStringBuilder is not a public class, as described by David Flanagan. It is defined as:

package java.lang;

// imports

abstract class AbstractStringBuilder implements Appendable, CharSequence {
   ...
}

So it is actually a package-private, or default - or whatever you call it - accessible class that is not visible to code outside the package. If you read the JavaDocs, they say both StringBuilder and StringBuffer extend Object, not AbstractStringBuilder! I guess JavaDoc figures what you don't know can't hurt you, right? This is interesting because I don't think I've ever seen a case where a superclass wasn't as visible as its subclasses, and David points out this is the first time he is aware it has happened (I assume in the JDK he means). I don't know if I necessarily think this is a good design, but I suppose it is an interesting way to limit who can create subclasses of an abstract class. Normally I've always wanted to openly permit subclasses of abstract classes as that is generally the point, i.e. the abstract class provides some base functionality that subclasses then must customize and extend. But if you for whatever reason need to restrict subclassing an abstract class, this is as good a way as any I suppose.

This past week I started adding a new set of features to a legacy EJB 2.x application that I wrote circa-2002 and which follows Core J2EE Patterns to the letter. In a nutshell, that means there are Struts Actions which call Business Delegates that delegate to Stateless Session Beans which in turn execute straight "Fast-Lane Reader" JDBC or talk to CMP Entity Beans with local interfaces. Whew! This is the first time in about three years that any new features have been requested for this application. Up until now it has been humming along nicely, though trying to migrate it to WebLogic 8.1 from WebLogic 7.0 was somewhat of a challenge with changed WebLogic deployment descriptor settings I needed to track down.

The new features are relatively minor, or at least they would be in a modern architecture using Spring, Hibernate, a more modern web framework than Struts, AJAX, etc. To add the new features as rapidly as possible - a week or two at most is what I was aiming for - I decided to stick to the existing architecture. The main reason I chose this path was internal consistency within the application, so there was not a mishmash of legacy and modern technologies which would undoubtedly make maintenance more difficult. The secondary reason is that I thought it would actually be faster to stick to the way it had originally been implemented, rather than try to introduce a bunch of new technologies that facilitate faster development. Whether it would have been faster to integrate newer technologies I'll never know as I am committed at this point and have implemented a fair chunk of the new features in the past few days.

But, what I've learned is just how slow and tedious using the legacy EJB 2.x architecture really is, especially now that I've done several projects using Spring, Hibernate, et. al. After all, I just want to implement a simple JDBC query and display it in a browser, right? Not with EJB 2.x! Dealing with JNDI lookups manually, obtaining and handling JDBC connections, rollbacks in Stateless Session Beans on exceptions, writing Business Delegates that essentially do nothing at all but "hide" RemoteExceptions and delegate to Session Beans and more, has really made me appreciate just how much easier things have become with the more lightweight frameworks supporting the Buzzwords-du-Jour such as IoC, Dependency Injection, Transparent Persistence, etc. This includes Spring, Hibernate, the new Java Persistence API (JPA), Ruby on Rails, and a lot more.

I now know exactly why it took me nine months to write the original version of this application back in 2002 using EJB 2.x technology. For a project of similar size and scope I'd estimate no more than three months and maybe less using modern frameworks in 2006. Probably even less if you are among those lucky enough to use Rails. So what's the point of all this? I think mainly that there has been a consistent progression towards lighter weight frameworks that just let you get your job done in a more efficient and less invasive manner. And also why DRY always results in faster development, more maintainable code, and more fun developing.