JRuby Utils

A library for creating and interacting with a pool of JRuby instances in Clojure.

Usage

This is a brief overview of the functionality of this library; TODO add more docs.

This library provides three public namespaces: puppetlabs.services.jruby-pool-manager.jruby-schemas, puppetlabs.services.jruby-pool-manager.jruby-pool-manager-service, and puppetlabs.services.jruby-pool-manager.jruby-core.

The entry-point into most of the functionality this library provides is the jruby-pool-manager-service Trapperkeeper service's create-pool, which creates and returns a pool context and asynchronously starts filling the pool.

create-pool takes a config matching the JRubyConfig schema; this config can be created and provided defaults by the jruby-core/initialize-config function. The non-optional config settings that you must provide are ruby-load-path and gem-home.

You can provide custom initialization and callback logic by providing lifecycle functions for initializing the scripting container, initializing the pool instance, cleaning up a pool instance, and cleaning up the pool for shutdown.

After the pool has been created, use the borrow and return functions to work with instances. There is a with-jruby-instance macro to make this easier. There is also a with-lock macro that holds a lock on the pool (so that no borrows can take place) while you execute some logic.

In most TK apps where you want to work with JRuby instances, you will want to call create-pool in the init lifecycle of your service, and then call jruby-core/flush-pool-for-shutdown! in the stop lifecycle function.

Running tests

Clojure unit tests can be run with either a 1.7-based JRuby or 9k-based JRuby version. To run tests with a 1.7-based JRuby, just run:

lein test

To run tests with a 9k-based JRuby, run:

lein with-profile +jruby9k test

License

See LICENSE.

Support

We use the Trapperkeeper project on JIRA for tickets on this project, although Github issues are welcome too.