Introduction of Google Error-Prone

Introdution to Google Error-Prone in Maven, which augments the compiler's type analysis and catches more mistakes at compile time!


In this article, I want to share with you a powerful Java Compiler plugin: Error Prone. Error Prone is an annotation processor that can be plugged into the Java compiler to augment the compiler’s type analysis. After reading this article, you will understand:

  • Why should we use Error Prone?
  • How to use it in Maven?
  • Some bug patterns in real
  • How to suppress warnings?
  • Some limits of Error Prone
  • How to go further in this topic?

Now, let’s get started!


Why should I use Error Prone?

Error Prone makes the Java compiler more powerful by analyzing the code during compile time. It has several advantages:

  • Shift-left. It makes errors discovered early in the software development lifecycle. Instead of being identified at build time, during code review, or in production, now they are identified at compile time. They are caught before they cost your time.
  • Neutral to build system. Google Error Prone is a Java annotation processor that is plugged in the compiler. So it can be used in any build system, e.g. Bazel, Maven, Gradle, Ant.
  • Suggestions for fixes. The fix of the problem is usually suggested at the same time when the bug is identified.

If you know others, please let me know by leaving a comment :)


How to use Error Prone in Maven?

To use Error Prone in Maven, you need to register the Google Error Prone Core artifact in the list of annotation processor paths of the Maven Compiler Plugin. It should look something like this in your POM file (pom.xml):


If you are running on JDK 8, you need to use error-prone-javac instead of the standard javac. More detail can be reached in the official installation guide here: In that page, you can also find the solutions for other build systems, such as Bazel, Gradle, Ant.

Bug Patterns

How does the error look like when a bug is discovered?

In this section, let’s take a look at some demos extracted from my pull-request to see how the source code looks like and how the error message and suggestion looks like.



LocalDateTime d = LocalDateTime.parse("2009-06-15T13:45:30");
d.format(DateTimeFormatter.ISO_LOCAL_DATE_TIME); // ERROR
assertEquals("2009-06-15T13:45:30", d.format(DateTimeFormatter.ISO_LOCAL_DATE_TIME));


lang: en

Error: Failed to execute goal org.apache.maven.plugins:maven-compiler-plugin:3.8.1:testCompile (default-testCompile) on project java-examples-date: Compilation failure Error: /home/runner/work/java-examples/java-examples/date/src/test/java/io/mincongh/date/so42364818/[23,13] [ReturnValueIgnored] Return value of this method must be used Error: (see Error: Did you mean to remove this line?

From the error message provided by Error Prone, you can see the explanation of the failure and a link for the official website for more detail. Now, focusing on this error, it is triggered because the return value of the method LocalDateTime.format(String) is not used. Local date-time in Java Time is an immutable class, so it means that I wrote one line of dead code and simply need to remove it:

  LocalDateTime d = LocalDateTime.parse("2009-06-15T13:45:30");
- d.format(DateTimeFormatter.ISO_LOCAL_DATE_TIME);
  assertEquals("2009-06-15T13:45:30", d.format(DateTimeFormatter.ISO_LOCAL_DATE_TIME));



public class PhoneNumberNoHash {


  public boolean equals(Object o) {
    if (o == this) return true;
    if (!(o instanceof PhoneNumberNoHash)) return false;
    PhoneNumberNoHash pn = (PhoneNumberNoHash) o;
    return pn.lineNumber == lineNumber && pn.prefix == prefix && pn.areaCode == areaCode;

  // Broken - no hashCode method!


Error: Failed to execute goal org.apache.maven.plugins:maven-compiler-plugin:3.8.1:compile (default-compile) on project java-examples-hashcode: Compilation failure: Compilation failure: Error: /home/runner/work/java-examples/java-examples/hashcode/src/main/java/io/mincongh/hashcode/bad/[26,18] [EqualsHashCode] Classes that override equals should also override hashCode. Error: (see Error: /home/runner/work/java-examples/java-examples/hashcode/src/main/java/io/mincongh/hashcode/bad/[24,18] [EqualsHashCode] Classes that override equals should also override hashCode. Error: (see

As you can see, this is triggered because the class only overrides the equals and does not override the hashCode. By doing so, it causes broken behavior when trying to store the object in a collection.

More Bug Patterns

Now we understand how the bug patterns look like, I’m not going to go further in this section. If you want to know the complete list of patterns, you can visit the “Bug patterns” page of the official website: to find the complete list.

Suppress Warnings

Sometimes, you may want to suppress warnings because you find those bugs identified by Error Prone unsuitable in your situation. Here are some ways to suppress warnings.

Use @SuppressWarnings. You can suppress false positives by adding the suppression annotation @SuppressWarnings("MyBugPattern") to the enclosing element. In other words, you can add this annotation to the variable declaration, method block, class block, etc. The value to use inside the annotation is the exact name of the bug-pattern defined by Error Prone. You can also suppress multiple warnings by providing a string array instead of a single value.

public class FormattingTest { ... }
@SuppressWarnings({"FormatString", "ArrayToString"})
public class FormattingTest { ... }

Disable one check. If you want to disable one check completely, you can turn off the check as a compiler option. Each check has a severity, it is one of the values of “OFF”, “WARN”, and “ERROR”. For example, to turn off the check ReferenceEquality, you can do:

# A valid Error Prone command-line option looks like:
#     -Xep:<checkName>[:severity]
# To turn off ReferenceEquality check:

Disable all checks. You may also want to disable all checks. This can be useful when you have a big codebase with a large number of existing bug patterns identified. These bugs are low-priority and cannot be fixed immediately. Therefore, you need to adapt Error Prone progressively by enabling patterns explicitly one after another. To disable all the checks, you need to pass the following option to the compiler:

# Disable all checks
# Then enable some of them

Exclude some paths. You can also define paths to exclude as a regular expression. It allows matching against a source file’s path to determine whether it should be excluded.

# Exclude generated files

You can see more information from the command-line flags page of the official website:


What are the limits of Error Prone?

Error Prone cannot be the only code analysis tool. It’s great to have a code analysis tool running at compile-time, but I think other code analysis tools still have their place. Some tools, such as Sonar, can be executed at build time and generate report which can be stored in a server. It aggregates historical data, attaches code coverage report, assigns to people, evaluates the quality score, etc. These cannot be handled by Error Prone.

Error Prone is limited to Java. We cannot use it for other languages.

Error Prone command-line options cannot be defined at a module level in Maven. At least I don’t know how to do… Centralizing everything inside the parent POM is not a very flexible solution. Sometimes I would like to custom the checks by modules, but I cannot do that.

Error Prone is sometimes annoying. Whenever you have an error, it fails at compile time. Some other errors may not be displayed until the current one is fixed. This fail-fast strategy makes it a bit annoying when you are actively writing code. Maybe Google developers are so good that they only write bug-free software :p

Going Further

How to go further from here?

  • To know more about the Error Prone, visit their official website Some additional topics that are not covered by this blog, such as writing a custom check, automatic refactoring using Refaster, but they can be found from the official website.
  • To see the source code of Error Prone, visit GitHub project:
  • Annotation processor is a powerful tool to bring additional capacity to Java compiler, such as code generation, compile-time type checks. To learn more about annotation processing, visit Gunnar Morling’s GitHub page “Awesome Java Annotation Processing”
  • Want to eliminate NullPointerExceptions (NPEs)? Uber has developed “NullAway”, an Error Prone plugin to check the null-ability in your code. Visit this project on GitHub:


In this article, we discovered Google Error-Prone, a Java annotation processing tool to detect code errors at compile-time. We saw the advantages of using Error Prone, the usage in Maven, some bug patterns, the ways to suppress warnings, the limitation, and finally, how to go further from here. Interested to know more? You can subscribe to the feed of my blog, follow me on Twitter or GitHub. Hope you enjoy this article, see you the next time!