Sep 12 2009

Capturing Values with Java Regular Expression Groups

Category: JavaPhil @ 6:48 pm

For whatever reason, I have been using a lot of Java Regular expressions at work lately; I started taking advantage of the grouping abilities provided by the standard Java Regular Expression implementation. I am certainly no RegEx expert, especial after reading this overview of backreferences and forward references! I use them every so often, probably not as much as I should; I know just enough to be dangerous! With my new exposure to grouping, I will probably use regular expressions much more frequently. Historically, if I needed to split a String into pieces, my goto class was the StringTokenizer. More often than not, I now use the String.split() method; you use create a simple regular expression and the method returns an array of the split values;  It is actually easier to use and handles the case where no matches are discovered fairly well.

If you are going to create and use regular expressions, you are probably going to need to test them. I have been using web-based expression testers for several years; here is one that I use often; Tonight, I found a FireFox plug-in I like ever better!

The point of this post was to show how easy it is to capture data using regular expressions. I created a little class to decode an id and status from a String. As illustrated in this simple example, you can use the Matcher class to do more than simply determine if the regular expression matched. With the addition of round brackets (parentheses) to the regular expression, we can actually capture the matched values in one simple step. This approach can also be used to capture optional values, the id and status, as in the included example. Unlike using a tokenizer or the split method, the Matcher class maintains the expected positions of non-specified, optional values; simply including an empty String for the optional attributes, when they are not provided.  As you can see the following Eclipse debugger screen shot, the groups are easily accessible, even allowing you to build nested groups, as my example demonstrates. You can play with the example or give the Matcher JavaDoc a quick read to learn more about how this class works.

public class StringDecoder {

    private static final Logger LOGGER = LoggerFactory.getLogger(StringDecoder.class);

    private Collection<String>  transactionIdentiferPatterns;

    public class Result {

        private long   transactionId;
        private String status;
        private String filename;

        public long getTransactionId() {
            return transactionId;

        public void setTransactionId(long transactionId) {
            this.transactionId = transactionId;

        public String getStatus() {
            return status;

        public void setStatus(String status) {
            this.status = status;

        public String getFilename() {
            return filename;

        public void setFilename(String filename) {
            this.filename = filename;

    private class Match {

        private String[] group;
        private int      start;
        private int      end;

        public String toString() {
            StringBuilder sb = new StringBuilder();
            sb.append("Match [end=" + end + ", start=" + start + ", value=");
            String comma = "";
            for (String s : group) {
                comma = ",";
            return sb.toString();

    public void setTransactionIdentiferPatterns(final Collection<String> patterns) {
        this.transactionIdentiferPatterns = patterns;

    public Result decode(final String filename) {
        Result rc = null;
        for (final String pattern : transactionIdentiferPatterns) {
            final Match match = find(StringUtils.trim(pattern), filename);
            if (match == null) {
                LOGGER.warn("No match found using pattern [{}]", pattern);
            else {
      "Mache found {}", match);
                rc = new Result();
                rc.setFilename(filename.substring(0, match.start) + filename.substring(match.end));
                return rc;

        return null;


    private Match find(final String patternStr, final String valueStr) {

        final Pattern pattern = Pattern.compile(patternStr);
        final Matcher matcher = pattern.matcher(valueStr);

        if (matcher.find()) {
            final int groupCount = matcher.groupCount();
            final Match match = new Match();

   = new String[groupCount + 1];
            for (int i = 0; i <= groupCount; i++) {
      [i] =;
                if (i == 0) {
                    match.start = matcher.start();
                    match.end = matcher.end();

            LOGGER.debug("Good {}", match);

            return match;

        return null;


    public void simple() {
        Collection<String> patterns = new ArrayList<String>();

        String name = "SomeStuff###OID122325F###.csv";
        Result result = this.decode(name);

        assertEquals(122325, result.getTransactionId());
        assertEquals("SomeStuff.csv", result.getFilename());
        assertEquals("F", result.getStatus());


Sep 09 2009

Hudson Task Scanner Plug-in

Category: Continuous Integration,Software DevelopmentPhil @ 8:21 pm

Ok, that is just about enough Hudson for this month, unless I get fired up about Selenium and try out the Hudson integration! I did find one more useful plug-in this week, the Task Scanner.  Because I tend to check in code early and often, my code is typically in an incomplete state; you will tend see a lot of TODO or FIXME comments in the code. Managing these task inside of Eclipse is easy enough, but I thought it might be interesting to also monitor them on Hudson’s dashboard.

The plug-in is very easy to configure and allows you to specify thresholds on the number of open tasks. Like most other plug-ins, it generates configurable trend graphs and provides a convenient task browsing screen.

Here was an interesting blog post that shows how you could use the plug-in to determine how messy your SVN merges will be… kind of cool..

Sep 08 2009

Continuous Deployment with Hudson…

Since I was looking through the Hudson plug-in list this week, I decided to try a few more of them. We currently have Hudson continuously deploying a simple web application to a Tomcat instance. Since both Hudson and Tomcat are running on the same server, it was pretty simple to have Tomcat point to the WAR file generated during the build process. Tomcat notices the new WAR file and automatically redeploys the application. Pretty simple and pretty handy; this is about a 100 times easier and faster than dealing with our standard Weblogic deployment process!

So, how do you deploy your application if Hudson is running on a different machine than your Tomcat instance? Fortunately, Hudson provides a plug-in to solve this problem, simply named the Deploy Plug-in. It supports a variety of containers, taking advantage of the Cargo Framework. Can’t get much easier than this, just point to your WAR file and fill in the container information. That’s it! I tried it out today and it worked perfectly. This was a much cleaner approach than adding the logic to the build process.

Having the current code set continuously deployed is a great benefit all by itself. This helps ensure that your application is built correctly and is deployable; How many times have you forgot to add that new JAR file to the WAR, only to discover the failure several days later? Continual deployment also makes the application constantly available to the analysis and test teams. This allows everyone to see the current state of development and bounce through the application to see how it will ultimately work. This early exposure also gives developers a change to resolve issues much earlier in the development cycle. We can achieve more value by adding the integration or system testing to the process as well… Just think how much we could reduce our development and testing cycles with this level of testing?

I have been starting to read about Selenium on the Hudson mailing list and found this interesting blog by Ben Rometsch on “Easy Automated Testing with Hudson and Selenium“. I really liked this picture from the article (hopefully it will spark your interest too!) and truly believe this is what we should be working towards… Maybe this can be one of my next educational adventures!!!

Sep 07 2009

Cleaning out my inbox!!!

Category: Blogging,JavaPhil @ 8:16 pm

I have been so busy with work and life, that my blogging had basically stopped for the summer.  I keep sending myself topics and notes about things that I would like to write about. Unfortunately, I have not had a chance to do anything with them and the topics keep piling up! I was talking with one of my coworkers last month, and he mentioned that he also had a blog. I mailed myself the link to his site,, and forgot all about it!

I thought It would be a good idea to share it, since many of my readers are coworkers as well! He has written some good stuff on the Spring Batch Framework,  REST, and Flex.  Please check it out…

Here is a  link to another one of my coworkers blogs… He  like to write about blogging and web security, you might find that interesting too…

Sep 07 2009

Why would anyone still be using IE6?

Category: Software DevelopmentPhil @ 7:54 pm

I have been working on a web application and was required to make the  work with IE6.  We were doing nothing fancy, but the application had numerous issues rendering in IE6. First, we implemented the presentation using  JSF and Facelets. All of our content was in XHTML documents, which IE6 apparently has no idea how to deal with! After doing some research,  we found a simple solution to the problem;  simply by removing the XML prologue, IE6 could now display the pages. We also had problems with some of the CommandLinks and CommandButtons not posting back to the server, by playing with different tag combinations, we finally got it working all of the browsers. What a pain in the behind!

I just cannot image how anyone could actually still be using IE6… Unfortunately, It is still the corporate standard where I work too. Somehow, I have no idea how, my Internet Explorer magically upgraded itself to IE7 a couple of years ago, and then again to IE8. It is amazing how that happened, and even more amazingly, everything still works fine! How do people surf without tabs? That is the question I really want answered! I really don’t use IE anyway, as the rich supply of plug-ins makes Firefox the best option out there, in my humble opinion.

Anyway, one of my friends forwarded me a couple of posts that I just had to share, the best one being “IE6 must die“! You just really have to love what gets people fired up… People are even spending their time building Javascript libraries and  WordPress plug-ins, to encourage people to upgrade. I found some of these URLs rather funny…


The most interesting post was about “IE6 Offenders“. I checked out the site,, which highlighted organizations that were still using IE6. It reported the number of users and the reason why they had not upgraded. In searching for the site tonight, the guy actually took his site down! I guess the guy must have caught some flack? I actually thought it was quite interesting!

I also like some of the banners, example below, that people are adding to their sites… I found this post tonight, The browser that will not die. It says that Microsoft will support IE6 until at least, 2014! That is amazingly sad!

Sep 07 2009

Hudson Plug-in for Custom Documentation, integrating an Ivy Report

Category: Continuous Integration,Software DevelopmentPhil @ 7:14 am

One of the best things about the Hudson Continuous Integration platform, is the amount of activity. People are constantly fixing problems and adding new functionality. I’m sure that some would say this is a bad thing, but even with the frequent releases, it is amazingly stable.

I have been moving projects to the Ivy Dependency Manager for the last couple of years. We have established a multi-repository solution, that allows us to control access to the repositories without impeding the progress of the development teams.  Introducing multiple repositories and transitive dependencies takes a little getting used to, so I have been adding an task to each of the project’s build scripts.  Here is an example of Ivy dependency report. I think it is quite handy on multiple levels, but the primary benefit is displaying the dependencies of each component. Side benefits include showing the different types licenses used by each component, kind of nice for corporate environments. The report also shows where the components are resolved from, such as the legacy or sandbox repositories (indicating that you might need to do some promotion or upgrade work before releasing the project).

My problem was, how to tie the Ivy dependency report into my continuous integration process. Building the report was easy, but how could I make it available on the main Hudson project page? Unfortunately, there was no plug-in to make this happen; I was actually thinking about writing my own plug-in this summer, but never quite got around to it (or blogging for that matter!) Fortunately, this appears to have been a common need and someone was nice enough to release a really nice plug-in this past July, called DocLinks. It allows you to add multiple links to items generated by the build process. The process could not be simpler.  The links will automatically show up on your main project page, but will not be active until the next build completes.

This also solved a problem that I have with the Testability Explorer plug-in. I have about a 50% success rate with this plug-in. I am able to generate both the XML and HTML files for all projects without issue, however, the Hudson Plug-in reports nothing for half of the projects; I have yet to figure out the pattern. Anyway, the DocLinks plug-in now lets me add a link to the Testability Explorer HTML report, so that everyone can always see the results, independent of the Hudson integration.

Suggestion!There is one area that the Hudson community could be a little more helpful. The Hudson tool is very good about showing you when updates are available for itself or installed plug-ins. However, there appears to be now good process for discovering new plug-ins. I even follow the Hudson mailing lists, and there are no notices about new or updated components; the only way I know about upgrades is to click the “configure” link every so often, when I actually remember! To find out about new plug-ins is even worse, I have to look at the Plug-in page, and notice if there is something new. There are so many plug-ins now, it is almost impossible to know if something has been added. Hopefully, they will add some kind of search for recently added or updated feature to their page; I’m sure that I’m not the only one that would appreciated it!

Sep 06 2009

Freemarker: Java Template Engine

Category: JavaPhil @ 8:35 pm

Freemarker has been around for several years and is an unbelievably simple and powerful tool. Freemarker is used by numerous tools and frameworks, just click the “powered by” link to the right. Although it has not had a major release for some time, it is still an active project. I believe the tool is so mature, that it essentially does exactly what it needs to, minimizing the need for new features/releases. I first became aware of Freemarker using the Hibernate reverse engineering tool. We had some specific code fragments and styles we wanted to generate; it was very simple to augment and override the base code generation functionality, once you knew what was going on! I also worked on a Struts2-based project, which also utilized Freemarker.

I liked the Freemarker overview picture, taken from the project’s web site; it shows just how simple the tool is to understand and use. I thought I would highlight a non-traditional use of Freemarker to demonstrate how easily it can be integrated into any project. And, hopefully get you thinking about some interesting ways it can be utilized. The API is extremely configurable, capable of utilizing multiple types of input, such as databases or files, and can even be used as an alternative to XSLT. Please read the feature page for a full list of capabilities.

To start with, I created a simple little wrapper class to hide the Freemarker specifics and make it easier to deal with different type input objects. Freemarker can populate templates using a Map or simple Java bean, my preferred method of interaction. The helper class has a simple API, one overloaded method named process. You simply provide the template as a String, along with the input object; the generated result String is returned from the method. I have not really needed to do anything super fancy and the helper class meets all of my needs, from conditional logic, iteration, to custom formating, which is all implemented in the template.

public class FreemarkerHelper {

    private static final Logger LOGGER = LoggerFactory.getLogger(FreemarkerHelper.class);

    public String process(final String template, final Object input) {
       final Configuration cfg = new Configuration();
       cfg.setObjectWrapper(new DefaultObjectWrapper());
       return process(cfg, template, input);

    public String process(final String template, final Map<String, Object> input) {
	return process(new Configuration(), template, input);

    private String process(final Configuration cfg, final String template, final Object input) {
        String rc = null;

       try {
           final Template temp =new Template("TemporaryTemplate", new StringReader(template), cfg);

           final Writer out = new StringWriter();
	   temp.process(input, out);
	   rc = out.toString();
       catch (final Exception e) {

       return rc;

In a recent project, I used Freemarker for multiple purposes, generating dynamic filenames and simple XML files. The system needed to dynamically generate filenames based on a specific set of meta data, such as the object id, user id, date, and other application specific data. We allowed the files names to be specified as a Freemarker template, such as….

String template = “MyFile_${userid}_${date?string(\”yyyyMMddHHmmss\”)}_${oid?string(\”#\”)}.xml“;
String template2 = “${oid?string(\”0000000000\”)}.${filename}“;

Using the basic formatting features, Freemarker allowed us to generate any filename (format) that was required, without changing the code. We could format the numeric object id as a simple number or padded with leading zeros. Likewise, dates could also be formatted using any combination of the date’s components, using just part of the date or the time. The more interesting part is that you just pass in a Java bean or map with those attributes defined and Freemarker does the real work. This is a pretty trivial usage of Freemarker, and might not be the most common usage pattern. However, I literally did not have to write any code, other than my little helper class, which I have used on multiple projects, and I can support unlimited filename combinations; Huge bang for the buck! Once I had the filename generated, I needed to generate a simple XML file to control some external processing. The file always had the same structure, just different values. This was perfect use of Freemarker, no need for for any heavy frameworks such as XML Beans or Castor. I just created a simple Java bean that captured the descriptive data and an XML template; within minutes, my XML control file was generated.

Another quick example, I recently built a little Servlet to act as a web service for a function that we wanted to make available to multiple applications. The primary output from was either a JSON or XML stream, but I also added a basic HTML response for testing and demonstrations purposes. I hacked together some basic HTML, but was disappointed by how ugly the application was! I Google for a free CSS template, and integrated it with my Servlet. I created a single page as a template and added variables for the title, menu, and body. Using my helper class, I was able to conditionally control the menus, as well as dynamically add the main body contents. Once again, this only took minutes to implement, and was quite a bit simpler than messing with Tiles or Sitemesh!

<#if current == "home" >
      <li id="current"><a href="/Context">Home</a></li>
      <li><a href="/Context">Home</a></li>