Compiling JAXB and JAX-WS classes with Maven in Java 11

Due to the deprecation of EE modules from SE, migrating projects to Java 11 can be somewhat trickier when you rely on Maven to compile your JAXB or JAX-WS classes. See https://jaxenter.com/jdk-11-java-ee-modules-140674.html for more information.

Here's how I got round the issues:

Dependencies

These were either the latest at time of writing, or the version that happened to work. Always check Maven Repository before throwing in dependencies - there may be a better version out there.

<javax.activation.version>1.1.1</javax.activation.version>
<jaxb-api.version>2.3.1</jaxb-api.version>
<jaxb-runtime.version>2.3.2</jaxb-runtime.version>
<jaxb2.maven.version>2.5.0</jaxb2.maven.version>
<jaxws.tools.version>2.3.2</jaxws.tools.version>

<dependency>
<groupId>javax.activation</groupId>
    <artifactId>activation</artifactId>
    <version>${javax.activation.version}</version>
</dependency>
<dependency>
    <groupId>javax.xml.bind</groupId>
    <artifactId>jaxb-api</artifactId>
    <version>${jaxb-api.version}</version>
</dependency>
<dependency>
    <groupId>org.glassfish.jaxb</groupId>
    <artifactId>jaxb-runtime</artifactId>
    <version>${jaxb-runtime.version}</version>
</dependency>
<dependency>
    <groupId>com.sun.xml.ws</groupId>
    <artifactId>jaxws-tools</artifactId>
    <version>${jaxws.tools.version}</version>
    <scope>compile</scope>
</dependency>

JAXB

Our legacy application was using the following dependencies from org.jvnet.jaxb2 to generate the classes based on the xsds/xjc:

<groupId>org.jvnet.jaxb2.maven2</groupId>
<artifactId>maven-jaxb2-plugin</artifactId>
<version>0.12.2</version>

<groupId>org.jvnet.jaxb2_commons</groupId>
<artifactId>jaxb2-basics</artifactId>
<version>0.6.4</version>

I won't go into detail as to how we configured those as it's not relevant here. Importantly, even by upgrading the versions, at time of writing, these simply do not work in Java 11! Here's the working solution which uses an xjc file to create the files. If you don't have one, they are easily created. There are many xjc guides out there.

<plugin>
    <groupId>org.codehaus.mojo</groupId>
    <artifactId>jaxb2-maven-plugin</artifactId>
    <version>${jaxb2.maven.version}</version>
    <executions>
        <execution>
            <id>mystuff</id>
            <goals>
                <goal>xjc</goal>
            </goals>
            <configuration>
                <sources>
                    <source>src/main/resources/xsd/</source>
                </sources>
                <packageName>my.package.name</packageName>
            </configuration>
        </execution>
    </executions>
</plugin>

JAX-WS

This is a bit different because we need the endpoints as well as the JAXB objects. Essentially we need to call wsimport differently than we would have using the fairly straightforward jaxws-maven-plugin.

<groupId>org.codehaus.mojo</groupId>
<artifactId>jaxws-maven-plugin</artifactId>
<version>2.4.1</version>

<plugin>
    <groupId>org.codehaus.mojo</groupId>
    <artifactId>exec-maven-plugin</artifactId>
    <version>${maven.exec.version}</version>
    <executions>
        <execution>
            <id>mkdir</id>
            <phase>generate-sources</phase>
            <goals>
                <goal>exec</goal>
            </goals>
            <configuration>
                <executable>mkdir</executable>
                <arguments>
                    <argument>-pv</argument>
                    <argument>target/generated-sources/wsimport</argument>
                </arguments>
            </configuration>
        </execution>
        <execution>
            <id>mystuff</id>
            <phase>generate-sources</phase>
            <goals>
                <goal>exec</goal>
            </goals>
            <configuration>
                <executable>java</executable>
                <arguments>
                    <argument>-classpath</argument>
                    <classpath/>
                    <argument>com.sun.tools.ws.WsImport</argument>
                    <argument>-extension</argument>
                    <argument>-Xnocompile</argument>
                    <argument>-wsdllocation</argument>
                    <argument>/xsd/file.wsdl</argument>
                    <argument>-s</argument>
                    <argument>target/generated-sources/wsimport</argument>
                    <argument>src/main/resources/xsd/file.wsdl</argument>
                </arguments>
            </configuration>
        </execution>
    </executions>
</plugin>

This plugin literally creates a directory and then calls wsimport from the com.sun.tools.ws.WsImport library.

<plugin>
    <groupId>org.codehaus.mojo</groupId>
    <artifactId>build-helper-maven-plugin</artifactId>
    <version>${maven.build.helper.version}</version>
    <executions>
        <execution>
            <id>add-source</id>
            <phase>generate-sources</phase>
            <goals>
                <goal>add-source</goal>
            </goals>
            <configuration>
                <sources>
                    <source>${project.build.directory}/generated-sources/wsimport/</source>
                </sources>
            </configuration>
        </execution>
    </executions>
</plugin>

For the code to be picked up use build-helper-maven-plugin as you normally would.

Factorials and Trailing Zeros

Here’s a seemingly impossible and odd query that I’ve been asked in a technical interview:

Question: “How many trailing zeros are there in 126! (factorial)”

At the time I was left stumped and made sure to find out the solution later - so here it is! It turns out there’s a simple solution that works for all numbers.

Answer: 

Firstly, as a quick reminder, a factorial is the product of an integer and the integers below it. For 126! it is: 126 * 125 * 124 * … 3 * 2 * 1

Which can be shown as: n * (n – 1) * (n – 1) * …

Now the solution. The simplest way has three steps, i.e. for number n:

1. Divide the number n by 5 and write down the quotient. Ignore remainders (if any).
2. Then divide the quotient by 5 again and again till you get a quotient less than 5.
3. Add up all the resultant quotients to get the number of zeros in n!

Therefore, for 126! We get to the answer via:

126 / 5 = 25 (Actually 25.2 but we ignore the fraction. Note that 125 / 5 gives us the whole number of 5 – this will be useful in a moment)

25 / 5 = 5

5 / 5 = 1

25 + 5 + 1 = 31 zeros in 126!

This pattern can be shown as:

[ ] indicates only whole numbers; neglect any fractional part in the division process. This pattern may make more sense when we note that 52 is 25 and 53 is 125. The formula technically goes on till infinity but fractions are considered as 0 so do not affect the final result.

For 126! This resulted in [25] + [5] + [1] = 31.

Why does this work? Well, here’s an explanation: We know that a number gets a zero at the end of it if the number has 10 as a factor, e.g. 10 is a factor of 20, 150, and 99990. Also, 5 * 2 = 10, so we need to account for all the products of 5 and 2. However, every other factor is even, so there are far more factors of 2 than 5 - As such, we have to count the number of factors divisible by 5. To phrase it differently, there are many more numbers that are multiples of 2 (2, 4, 6, 8, 10, 12, 14, ...) than are multiples of 5 (5, 10, 15, ...). If we were to take all the numbers with 5 as a factor, we'll have way more than enough even numbers to pair with them to get factors of 10 (and another trailing zero on the factorial). Finally, we need to consider numbers which are factors of 5 and count them again, e.g. 25 (which is 5 * 5 or 52), 125 (5 * 5 * 5 or 53), 625 (5 * 5 * 5 * 5 or 54) etc.

To conclude, here is 126! in full (feel free to count the trailing zeros):

23721732428800468856771473051394170805702085973808045661837377170052497697783313457227249544076486314839447086187187275319400401837013955325179315652376928996065123321190898603130880000000000000000000000000000000

The approximate value is: 2.37217324288E+211

Cleaner code - Comments

I truly believe that if the code is written well enough, comments should not be required at all. Not all comments are bad. They can be useful, i.e. they add additional details, information, caveats or provide example inputs (e.g. good/bad inputs for complex regex). Sadly, most of the time comments are put in with little thought, added automatically and are not kept up to date. It has become a bug-bear of mine and this nonsense has to stop! The following are the types of comments that bug me the most!

Bad comments

Comments should be informative, explain the intent or clarify what the code is doing. Sadly this is not always the case. Please be on guard to remove these variances of bad commenting:

Journal comments

We don't need to document when code was created, last updated or by whom. That's what versioning systems are for.

* Changes (from 11-Oct-2001) * -------------------------- * 11-Oct-2001 : Re-organised the class and moved it to new package *               com.jrefinery.date (DG); * 05-Nov-2001 : Added a getDescription() method, and eliminated NotableDate *               class (DG); * 12-Nov-2001 : IBD requires setDescription() method, now that NotableDate *               class is gone (DG);  Changed getPreviousDayOfWeek(), *               getFollowingDayOfWeek() and getNearestDayOfWeek() to correct *               bugs (DG);

Attributions and Bylines

/* Added by Rick */

Thanks Rick. Remove anything like this.

Noise / Redundant comments

Sometimes you see comments that are nothing but noise. They restate the obvious and provide no new information.

/**   * Returns the day of the month.   *   * @return the day of the month.   */  public int getDayOfMonth() {    return dayOfMonth; }

Consider what purpose does the comment serve. If it’s not more informative than the code, remove.

Misleading comments

These are possibly the worst kind as they may state a fact about the code that may have been true once but is now false. Too often the comments get separated from the code they describe and become orphaned blurbs of ever-decreasing accuracy.

JavaDocs in Nonpublic code

Javadocs are useful for public APIs. Generating Javadoc pages for the classes and functions inside a system is not generally useful, and the extra formality of the Javadoc comments amounts to little more than a distraction. Comments like this just clutter up the code, propagate lies, and lend to general confusion and disorganization.

Print the content of a JavaScript object

I was working on an Eclipse plugin based on a 3rd party API. One feature of their API I used allowed me to display a HTML page in a window. Sadly, within the applications frame, the page is impossible to debug with modern browser tools (e.g. FireBug in FireFox or DevTools in Chrome). I needed a different way to debug the JavaScript Objects I had on the page. Here is the useful solution which ending up helping me:

function printObject(o) {
  var out = '';
  for (var p in o) {
    out += p + ': ' + o[p] + '\n';
  }
  alert(out);
}

// test:
var myObject = {'something': 1, 'other thing': 2};
printObject(myObject);

Here we can access all the elements of an object using a foreach loop. The printObject function alert()s the object showing all properties and respective values.

Oracle Serive Bus: Get file name from proxy service

The following is how to get the file name of a file that is being processed by a proxy service in OSB.

1. We can get the URI via the $inbound context variable with the following XPath:
$inbound/ctx:transport/ctx:request/tp:headers/file:fileName

This results in:
<file:fileName
xmlns:file="http://www.bea.com/wli/sb/transports/file">
c:\temp\input\2635978486281_
test.xml</file:fileName>

2. To get the filename only, use the tokenize function (which returns the characters after the last "\\":
tokenize($inbound/ctx:transport/ctx:request/tp:headers/file:fileName, "\\")[last()]

Auto increment primary key with Oracle Databases

Suppose you have a database and you want each entry to be identified by a unique number. This can be achieved easily in MySQL by specifying "auto_increment" for the column, but Oracle requires a little bit of extra work to get this done.

One way to do it is by creating two database objects, a sequence and a trigger:

1. As a simple example, imagine we have a table called "tbl_test" with two columns, "test_id" and "test_data".
create table tbl_test (test_id number, test_data varchar2(145)); 

 2. Now, we create a sequence for the test_id column. Note that the numbering starts with 1 and is incremented by 1.
create sequence seq_test 
start with 1 
increment by 1 
nomaxvalue; 

3. Next we create the trigger that will automatically insert the next number from the sequence into the test_id column.
create trigger tgr_test 
before insert on tbl_test 
for each row 
begin 
select seq_test.nextval into :new.test_id from dual; 
end;
/

XMLType in SQL Developer

I thought I'd add this as for someone new to SQL Developer, this is not obvious at first glance. When creating a table with column(s) of type XMLType, the type is not available via the standard Create Table wizard. Below are the full steps to use that datatype:

• Right-click on the 'Tables' node in the 'Connections' view
• Select 'New Table'
• Check the 'Advanced' check box located in the top right corner - This will show an alternate 'Create Table' dialogue which hosts considerably more functionality
• In the 'Column Properties' area, select the 'Complex' radio button for datatype
• Choose the SYS schema; the XMLType will be available in the drop down