Table of Contents

Using System Properties in Native Images

Assume you have the following Java Program:

public class App {
    public static void main(String[] args) {

If you compile that with, e.g., native-image -Dfoo=bar App the system property foo will be available at image build time. For example, whenever you are in the code that is part of your application but executed at image build time (usually static field initializations and static initializers). Thus if you execute the image above it will not contain foo in the list of properties.

If, on the other hand, you execute the image with app -Dfoo=bar, it will show foo in the list of properties because you specified it for image run time.

In other words:

  • Passing -D<key>=<value> to native-image affects properties seen at image build time.
  • Passing -D<key>=<value> to an image execution affects properties seen at image run time.

Access Environment Variables at Run Time

Native image can also access environment variables at runtime. Consider the following example.

  1. Save this Java code into the file:
  import java.util.Map;
  public class EnvMap {
      public static void main (String[] args) {
          var filter = args.length > 0 ? args[0] : "";
          Map<String, String> env = System.getenv();
          for (String envName : env.keySet()) {
              if(envName.contains(filter)) continue;

This code iterates over the environment variables and prints out the ones passing through the filter, passed as the command line argument.

  1. Compile and build a native image:
  native-image EnvMap
  1. Run the resulting native image and pass some argument. It will correctly print out the values of the environment variables. For example:
  ./envmap HELLO
  HELLOWORLD=hello world
  export HELLOWORLD="world"
  ./envmap HELLO