Home - News ( United States | United Kingdom | Italy | Germany ) - Football scores

Showing content from https://docs.pmd-code.org/pmd-doc-7.3.0/pmd_userdocs_extending_defining_properties.html below:

Defining rule properties | PMD Source Code Analyzer

Learn how to define your own properties both for Java and XPath rules.

Rule properties are a way to make your rules configurable directly from the ruleset XML. Their usage is described on the Configuring Rules page.

If you’re a rule developer, you may want to think about what would be useful for a user of your rule to parameterize. It could be a numeric report level, a boolean flag changing the behaviour of your rule… Chances are there is some detail that can be abstracted away from your implementation, and in that case, this page can help you squeeze that sweet flexibility out of your rule.

The basic thing you need to do as a developer is to define a property descriptor and declare that your rule uses it. A property descriptor defines a number of attributes for your property:

• Its name, with which the user will refer to your property;
• Its description, for documentation purposes;
• Its default value

All of these attributes can be specified in a single Java statement (or XML element for XPath rules).

The procedure to define a property is quite straightforward:

• Create a property descriptor of the type you want, by using a builder from PropertyFactory
• Call definePropertyDescriptor(PropertyDescriptor) in the rule’s noarg constructor.

You can then retrieve the value of the property at any time using getProperty(PropertyDescriptor).

Note: The base class for all rule implementations is AbstractRule, which is a PropertySource. So you can directly call definePropertyDescriptor(...) or getProperty(...) within your rule.

Properties can be built using type-specific builders, which can be obtained from the factory methods of PropertyFactory. For example, to build a string property, you would call

PropertyFactory.stringProperty("myProperty")
               .desc("This is my property")
               .defaultValue("foo")
               .build();

This is fairly more readable than a constructor call, but keep in mind the description and the default value are not optional.

For numeric properties, you can add constraints on the range of acceptable values, e.g.

PropertyFactory.intProperty("myIntProperty")
               .desc("This is my property")
               .defaultValue(3)
               .require(positive()) // must be > 0
               .require(below(100)) // must be <= 100
               .build();

Predefined constraints such as positive and below are available in the class NumericConstraints. A custom constraint can be implemented by implementing the interface PropertyConstraint.

Enumerated properties do not have a specific value type, instead, you can choose any type of value, provided the values are from a closed set. To make that actionable, you give string labels to each of the acceptable values, and the user will provide one of those labels as a value in the XML. The property will give you back the associated value, not the label. Here’s an example:

enum Mode {
    Easy, Hard
}

// Using this method, the labels are the `toString` of each enum constant.
// To customize this look at the overloads of `enumProperty`.
static PropertyDescriptor<Mode> modeProperty
 = PropertyFactory.enumProperty("modeProperty", Mode.class)
                  .desc("This is my property")
                  .defaultValue(Mode.Easy)
                  .build();

You can see an example of properties used in a PMD rule such as AvoidReassigningLoopVariables. There are several things to notice here:

• The property descriptors are declared static final, which should generally be the case, as descriptors are immutable and can be shared between instances of the same rule;
• The property is declared using definePropertyDescriptor` in the constructor, which ensures the property gets recognised by PMD at the time the properties are overridden (which happens before rule execution);
• The value of the property is not retrieved in the constructor, but in one of the visit methods (typically on the highest node in the tree, since the property doesn’t change).

XPath rules can also define their own properties. To do so, you must add a property element in the properties element of your rule, which declares the type attribute. This attribute conditions what type the underlying property has, and can have the following values:

Note that enumerated properties are not available in XPath rules.

Properties defined in XPath also must declare the description attribute. Numeric properties also expect the min and max attributes for now. Here are a few examples to sum it up:

<property name="stringProp" type="Boolean" value="true" description="A BooleanProperty."/>
<property name="intProp" type="Integer" value="3" min="1" max="20" description="An IntegerProperty."/>

You can then use the property in XPath with the syntax $propertyName, for example:

<rule name="MyXpathRule" ...>
  <properties>
    <property name="maxStatements" type="Integer" value="10" min="1" max="40"
              description="Max number of statements per method"/>
    <property name="xpath">
      <value>
        <![CDATA[
            //MethodDeclaration/Block[count(./*) > $maxStatements]
        ]]></value>
    </property>
  </properties>
</rule>

Multivalued properties are also allowed and their type attribute has the form List[Boolean] or List[Character], with every above type allowed. These properties make use of the sequence datatype provided by XPath 2.0 and above. When providing multiple values, the delimiter is a simple comma (“,”). The comma can be escaped with a backslash when needed.

<rule name="MyXpathRule" ...>
  <properties>
    <property name="intProp" type="List[Integer]" value="1,2,5" description="An IntegerMultiProperty." />
    <property name="reportedIdentifiers" type="List[String]" value="foo,bar"
              description="A StringMultiProperty." />
    <property name="xpath">
        <value><![CDATA[
            //VariableId[@Image = $reportedIdentifiers]
        ]]></value>
    </property>
  </properties>
</rule>

Notice that in the example above, @Name = $reportedIdentifiers doesn’t test @Name for equality with the whole sequence ('foo', 'bar'), it tests whether the sequence contains @Name. That is, the above rule will report all variables named foo or bar. All other XPath 2.0 functions operating on sequences are supported.

RetroSearch is an open source project built by @garambo | Open a GitHub Issue

Search and Browse the WWW like it's 1997 | Search results from DuckDuckGo

HTML: 3.2 | Encoding: UTF-8 | Version: 0.7.4