1. Overview
Simply put, PMD is a source code analyzer to find common programming flaws like unused variables, empty catch blocks, unnecessary object creation, and so forth.
It supports Java, JavaScript, Salesforce.com Apex, PLSQL, Apache Velocity, XML, XSL.
In this article, we’ll focus on how to use PMD to perform static analysis in a Java project.
2. Prerequisites
Let’s start with setting up PMD into a Maven project – using and configuring the maven-pmd-plugin:
<project>
...
<reporting>
<plugins>
<plugin>
<groupId>org.apache.maven.plugins</groupId>
<artifactId>maven-pmd-plugin</artifactId>
<version>3.7</version>
<configuration>
<rulesets>
<ruleset>/rulesets/java/braces.xml</ruleset>
<ruleset>/rulesets/java/naming.xml</ruleset>
</rulesets>
</configuration>
</plugin>
</plugins>
</reporting>
</project>
You can find the latest version of maven-pmd-plugin here.
Notice how we’re adding rulesets in the configuration here – these are a relative path to already define rules from the PMD core library.
Finally, before running everything, let’s create a simple Java class with some glaring issues – something that PMD can start reporting problems on:
public class Ct {
public int d(int a, int b) {
if (b == 0)
return Integer.MAX_VALUE;
else
return a / b;
}
}
3. Run PMD
With the simple PMD config and the sample code – let’s generate a report in the build target folder:
mvn site
The generated report is called pmd.html and is located in the target/site folder:
Files com/baeldung/pmd/Cnt.java Violation Line Avoid short class names like Cnt 1–10 Avoid using short method names 3 Avoid variables with short names like b 3 Avoid variables with short names like a 3 Avoid using if...else statements without curly braces 5 Avoid using if...else statements without curly braces 7
As you can see – we’re not getting results. The report shows violations and line numbers in your Java code, according to PMD.
4. Rulesets
The PMD plugin uses five default rulesets:
- basic.xml
- empty.xml
- imports.xml
- unnecessary.xml
- unusedcode.xml
You may use other rulesets or create your own rulesets, and configure these in the plugin:
<project>
...
<reporting>
<plugins>
<plugin>
<groupId>org.apache.maven.plugins</groupId>
<artifactId>maven-pmd-plugin</artifactId>
<version>3.7</version>
<configuration>
<rulesets>
<ruleset>/rulesets/java/braces.xml</ruleset>
<ruleset>/rulesets/java/naming.xml</ruleset>
<ruleset>/usr/pmd/rulesets/strings.xml</ruleset>
<ruleset>http://localhost/design.xml</ruleset>
</rulesets>
</configuration>
</plugin>
</plugins>
</reporting>
</project>
Notice that we’re using either a relative address, an absolute address or even a URL – as the value of the ‘ruleset’ value in configuration.
A clean strategy for customizing which rules to use for a project is to write a custom ruleset file. In this file, we can define which rules to use, add custom rules, and customize which rules to include/exclude from the official rulesets.
5. Custom Ruleset
Let’s now choose the specific rules we want to use from existing sets of rules in PMD – and let’s also customize them.
First, we’ll create a new ruleset.xml file. We can of course use one of the existing rulesets files as an example and copy and paste that into our new file, delete all the old rules from it, and change the name and description:
<?xml version="1.0"?>
<ruleset name="Custom ruleset"
xmlns="http://pmd.sourceforge.net/ruleset/2.0.0"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xsi:schemaLocation="http://pmd.sourceforge.net/ruleset/2.0.0
http://pmd.sourceforge.net/ruleset_2_0_0.xsd">
<description>
This ruleset checks my code for bad stuff
</description>
</ruleset>
Secondly, let’s add some rule references:
<!-- We'll use the entire 'strings' ruleset --> <rule ref="rulesets/java/strings.xml"/>
Or add some specific rules:
<rule ref="rulesets/java/unusedcode.xml/UnusedLocalVariable"/> <rule ref="rulesets/java/unusedcode.xml/UnusedPrivateField"/> <rule ref="rulesets/java/imports.xml/DuplicateImports"/> <rule ref="rulesets/java/basic.xml/UnnecessaryConversionTemporary"/>
We can customize the message and priority of the rule:
<rule ref="rulesets/java/basic.xml/EmptyCatchBlock"
message="Must handle exceptions">
<priority>2</priority>
</rule>
And you also can customize a rule’s property value like this:
<rule ref="rulesets/java/codesize.xml/CyclomaticComplexity">
<properties>
<property name="reportLevel" value="5"/>
</properties>
</rule>
Notice that you can customize individual referenced rules. Everything but the class of the rule can be overridden in your custom ruleset.
Next – you can also excluding rules from a ruleset:
<rule ref="rulesets/java/braces.xml">
<exclude name="WhileLoopsMustUseBraces"/>
<exclude name="IfElseStmtsMustUseBraces"/>
</rule>
Next – you can also exclude files from a ruleset using exclude patterns, with an optional overriding include pattern.
A file will be excluded from processing when there is a matching exclude pattern, but no matching include pattern.
Path separators in the source file path are normalized to be the ‘/’ character, so the same ruleset can be used on multiple platforms transparently.
Additionally, this exclude/include technique works regardless of how PMD is used (e.g. command line, IDE, Ant), making it easier to keep the application of your PMD rules consistent throughout your environment.
Here’s a quick example:
<?xml version="1.0"?>
<ruleset ...>
<description>My ruleset</description>
<exclude-pattern>.*/some/package/.*</exclude-pattern>
<exclude-pattern>
.*/some/other/package/FunkyClassNamePrefix.*
</exclude-pattern>
<include-pattern>.*/some/package/ButNotThisClass.*</include-pattern>
<rule>...
</ruleset>
6. Conclusion
In this quick article we introduced PMD – a flexible and highly configurable tool focused on static analysis of Java code
As always, the full code presented in this tutorial is available over on Github.