Using Eclipse with JRuby Truffle

This is an experiment, mainly to have a much better feedback from the Truffle annotation processor and ideally much faster compile time!

Using integrated Maven support

Create a new workspace in Eclipse.

Start by ensuring there is no leftover configuration:

$ rm -rf core/.classpath core/.settings core/.project

We can now import the project:

• From the main menu bar, select File > Import...
• Select Maven > Existing Maven Projects
• Select jruby/core directory as root directory (NOT just jruby)
• Click Finish

And now enable the Truffle annotation processor:

• From the main menu bar, select Project > Properties
• Select Java Compiler > Annotation Processing
• Click on Enable project specific settings, then OK
• Click Yes to rebuild the project.

You shall be set!

It is possible to run JRuby with a run configuration for org.jruby.Main.main(). Further integration is under investigation.

Old technique not using Maven support

Be sure to have the project built before with mvn.

Maven can be used to generate Eclipse project files:

mvn eclipse:eclipse

Create a new workspace in Eclipse, outside the jruby directory (otherwise Eclipse complains about overlapping dirs).

We need to set the path to your Maven repository (usually ~/.m2/repository):

• From the menu, select Window > Preferences
• In the search field type Classpath Variables
• Click on New and enter the variable name: M2_REPO
• Click on Folder... and select your home, then .m2 then repository

We can now import the project:

• From the main menu bar, select File > Import...
• Select General > Existing Projects into Workspace
• Enter the jruby directory in root directory
• Click Finish

Close the jruby-tests project since the setup is wrong for now.

There are some for now unsolved errors in JRuby generated files, so in the Package Explorer, open target/generated-sources, then right click on org.jruby and select Build Path > Exclude.

Now it should start to work to some extent!