Rules to detect constructs that are either broken, extremely confusing or prone to runtime errors.
Edit me

ApexCSRF

Since: PMD 5.5.3

Priority: Medium (3)

Having DML operations in Apex class constructor or initializers can have unexpected side effects: By just accessing a page, the DML statements would be executed and the database would be modified. Just querying the database is permitted.

In addition to constructors and initializers, any method called init is checked as well.

Salesforce Apex already protects against this scenario and raises a runtime exception.

Note: This rule has been moved from category "Security" to "Error Prone" with PMD 6.21.0, since using DML in constructors is not a security problem, but crashes the application.

This rule is defined by the following Java class: net.sourceforge.pmd.lang.apex.rule.errorprone.ApexCSRFRule

Example(s):

public class Foo {
    // initializer
    {
        insert data;
    }

    // static initializer
    static {
        insert data;
    }

    // constructor
    public Foo() {
        insert data;
    }
}

Use this rule by referencing it:

<rule ref="category/apex/errorprone.xml/ApexCSRF" />

AvoidDirectAccessTriggerMap

Since: PMD 6.0.0

Priority: Medium (3)

Avoid directly accessing Trigger.old and Trigger.new as it can lead to a bug. Triggers should be bulkified and iterate through the map to handle the actions for each item separately.

This rule is defined by the following XPath expression:

//ArrayLoadExpression[TriggerVariableExpression and LiteralExpression]

Example(s):

trigger AccountTrigger on Account (before insert, before update) {
   Account a = Trigger.new[0]; //Bad: Accessing the trigger array directly is not recommended.

   for ( Account a : Trigger.new ) {
        //Good: Iterate through the trigger.new array instead.
   }
}

Use this rule by referencing it:

<rule ref="category/apex/errorprone.xml/AvoidDirectAccessTriggerMap" />

AvoidHardcodingId

Since: PMD 6.0.0

Priority: Medium (3)

When deploying Apex code between sandbox and production environments, or installing Force.com AppExchange packages, it is essential to avoid hardcoding IDs in the Apex code. By doing so, if the record IDs change between environments, the logic can dynamically identify the proper data to operate against and not fail.

This rule is defined by the following Java class: net.sourceforge.pmd.lang.apex.rule.errorprone.AvoidHardcodingIdRule

Example(s):

public without sharing class Foo {
    void foo() {
        //Error - hardcoded the record type id
        if (a.RecordTypeId == '012500000009WAr') {
            //do some logic here.....
        } else if (a.RecordTypeId == '0123000000095Km') {
            //do some logic here for a different record type...
        }
    }
}

Use this rule by referencing it:

<rule ref="category/apex/errorprone.xml/AvoidHardcodingId" />

AvoidNonExistentAnnotations

Since: PMD 6.5.0

Priority: Medium (3)

Apex supported non existent annotations for legacy reasons. In the future, use of such non-existent annotations could result in broken apex code that will not compile. This will prevent users of garbage annotations from being able to use legitimate annotations added to Apex in the future. A full list of supported annotations can be found at https://developer.salesforce.com/docs/atlas.en-us.apexcode.meta/apexcode/apex_classes_annotation.htm

This rule is defined by the following Java class: net.sourceforge.pmd.lang.apex.rule.errorprone.AvoidNonExistentAnnotationsRule

Example(s):

@NonExistentAnnotation public class ClassWithNonexistentAnnotation {
    @NonExistentAnnotation public void methodWithNonExistentAnnotation() {
        // ...
    }
}

Use this rule by referencing it:

<rule ref="category/apex/errorprone.xml/AvoidNonExistentAnnotations" />

EmptyCatchBlock

Since: PMD 6.0.0

Priority: Medium (3)

Empty Catch Block finds instances where an exception is caught, but nothing is done. In most circumstances, this swallows an exception which should either be acted on or reported.

This rule is defined by the following XPath expression:

//CatchBlockStatement[./BlockStatement[count(*) = 0]]

Example(s):

public void doSomething() {
    ...
    try {
        insert accounts;
    } catch (DmlException dmle) {
        // not good
    }
}

Use this rule by referencing it:

<rule ref="category/apex/errorprone.xml/EmptyCatchBlock" />

EmptyIfStmt

Since: PMD 6.0.0

Priority: Medium (3)

Empty If Statement finds instances where a condition is checked but nothing is done about it.

This rule is defined by the following XPath expression:

//IfBlockStatement
 [BlockStatement[count(*) = 0]]

Example(s):

public class Foo {
    public void bar(Integer x) {
        if (x == 0) {
            // empty!
        }
    }
}

Use this rule by referencing it:

<rule ref="category/apex/errorprone.xml/EmptyIfStmt" />

EmptyStatementBlock

Since: PMD 6.0.0

Priority: Medium (3)

Empty block statements serve no purpose and should be removed.

This rule is defined by the following XPath expression:

//Method/ModifierNode[@Abstract!= true() and ../BlockStatement[count(*) = 0]]
| //Method/BlockStatement//BlockStatement[count(*) = 0 and @Location != parent::*/@Location]

Example(s):

public class Foo {

   private Integer _bar;

   public void setBar(Integer bar) {
        // empty
   }

}

Use this rule by referencing it:

<rule ref="category/apex/errorprone.xml/EmptyStatementBlock" />

EmptyTryOrFinallyBlock

Since: PMD 6.0.0

Priority: Medium (3)

Avoid empty try or finally blocks - what’s the point?

This rule is defined by the following XPath expression:

//TryCatchFinallyBlockStatement[./BlockStatement[count(*) = 0]]

Example(s):

public class Foo {
    public void bar() {
        try {
          // empty !
        } catch (Exception e) {
            e.printStackTrace();
        }
    }
}

public class Foo {
    public void bar() {
        try {
            Integer x=2;
        } finally {
            // empty!
        }
    }
}

Use this rule by referencing it:

<rule ref="category/apex/errorprone.xml/EmptyTryOrFinallyBlock" />

EmptyWhileStmt

Since: PMD 6.0.0

Priority: Medium (3)

Empty While Statement finds all instances where a while statement does nothing. If it is a timing loop, then you should use Thread.sleep() for it; if it is a while loop that does a lot in the exit expression, rewrite it to make it clearer.

This rule is defined by the following XPath expression:

//WhileLoopStatement[./BlockStatement[count(*) = 0]]

Example(s):

public void bar(Integer a, Integer b) {
  while (a == b) {
    // empty!
  }
}

Use this rule by referencing it:

<rule ref="category/apex/errorprone.xml/EmptyWhileStmt" />

MethodWithSameNameAsEnclosingClass

Since: PMD 5.5.0

Priority: Medium (3)

Non-constructor methods should not have the same name as the enclosing class.

This rule is defined by the following Java class: net.sourceforge.pmd.lang.apex.rule.errorprone.MethodWithSameNameAsEnclosingClassRule

Example(s):

public class MyClass {
    // this is OK because it is a constructor
    public MyClass() {}
    // this is bad because it is a method
    public void MyClass() {}
}

Use this rule by referencing it:

<rule ref="category/apex/errorprone.xml/MethodWithSameNameAsEnclosingClass" />

OverrideBothEqualsAndHashcode

Since: PMD 6.31.0

Priority: Medium (3)

Override both public Boolean equals(Object obj), and public Integer hashCode(), or override neither. Even if you are inheriting a hashCode() from a parent class, consider implementing hashCode and explicitly delegating to your superclass.

This is especially important when Using Custom Types in Map Keys and Sets.

This rule is defined by the following Java class: net.sourceforge.pmd.lang.apex.rule.errorprone.OverrideBothEqualsAndHashcodeRule

Example(s):

public class Bar {        // poor, missing a hashCode() method
    public Boolean equals(Object o) {
      // do some comparison
    }
}
public class Baz {        // poor, missing an equals() method
    public Integer hashCode() {
      // return some hash value
    }
}
public class Foo {        // perfect, both methods provided
    public Boolean equals(Object other) {
      // do some comparison
    }
    public Integer hashCode() {
      // return some hash value
    }
}

Use this rule by referencing it:

<rule ref="category/apex/errorprone.xml/OverrideBothEqualsAndHashcode" />

TestMethodsMustBeInTestClasses

Since: PMD 6.22.0

Priority: Medium (3)

Test methods marked as a testMethod or annotated with @IsTest, but not residing in a test class should be moved to a proper class or have the @IsTest annotation added to the class.

Support for tests inside functional classes was removed in Spring-13 (API Version 27.0), making classes that violate this rule fail compile-time. This rule is mostly usable when dealing with legacy code.

This rule is defined by the following XPath expression:

//UserClass[
      not(./ModifierNode/Annotation[lower-case(@Image) = 'istest']) and
      (
        (./Method/ModifierNode/Annotation[lower-case(@Image) = 'istest']) or
        (./Method/ModifierNode[@Test = true()])
      )
    ]

Example(s):

// Violating
private class TestClass {
  @IsTest static void myTest() {
    // Code here
  }
}

private class TestClass {
  static testMethod void myTest() {
    // Code here
  }
}

// Compliant
@IsTest
private class TestClass {
  @IsTest static void myTest() {
    // Code here
  }
}

@IsTest
private class TestClass {
  static testMethod void myTest() {
    // Code here
  }
}

Use this rule by referencing it:

<rule ref="category/apex/errorprone.xml/TestMethodsMustBeInTestClasses" />