Inside vRO’s JavaScript Engine – Rhino 1.7R4 [CB10102]

vRealize Orchestrator a.k.a. vRO, which is a drag-and-drop automation tool, is quite an old tool developed and released as early as in 2007 by Dunes Technologies, Switzerland. After VMware acquired Dunes, there were 100’s of releases and updates came year after year. Lots and lots of improvements made over time in the UI, backend technologies, security, multi-language support, etc. However, one thing that remains the same is its JavaScript Engine. vRO uses Mozilla Rhino Engine 1.7R4 which was released in 2012.

In this post, My goal is to provide some insights on this Rhino Engine as it is almost extinct from the internet. However, I am still behind the JavaScript Engine which provides IntelliSense support to vRO 8.x. As you might have noticed and probably be wondering how CTRL+SPACE shows options only available to recent versions of JavaScript. I guess it’s for Node.js runtime.

What is Rhino Engine?

Rhino Engine converts JavaScript scripts into classes. It is intended to be used in desktop or server-side applications, hence there is no built-in support for the Web browser objects that are commonly associated with JavaScript which makes it very suitable for vRO. Rhino works in both compiled and interpreted mode. Rhino Engine got its name from the animal on the cover of the O’Reilly book about JavaScript published many years back.

The Rhino project was started at Netscape in the autumn of 1997. At the time, Netscape was planning to produce a version of Navigator written entirely in Java and so it needed an implementation of JavaScript written in Java. When Netscape stopped work on “Javagator,” as it was called, somehow Rhino escaped the axe (rumor had it that the executives “forgot” it existed). For a time, a couple of major companies (including Sun) licensed Rhino for use in their products and paid Netscape to do so, allowing work on Rhino to continue. Now Rhino is part of Mozilla’s open-source repository.

Released 10 years ago

Released in 2012-06-18, Rhino 1.7R4 is almost prehistoric for today’s standards and that’s been always a point of discussion in the vRO Community.

Compatibility with JavaScript features

While trying to look deep into its compatibility matrix with ES, I found this Kangax’s Compat-table which gives an excellent and detailed view of all the possible features that Rhino 1.7R4 supports. Click the link to know more.

Rhino Compatibility Matrix with JavaScript

ECMAScript 5.1 Specifications

This document will give you a very in-depth knowledge of the ECMAScript 5.1 that vRO leverages to understand the language better. Learn more at https://262.ecma-international.org/5.1.

You can download this document and read about all the fine details by yourself.

ecma-262_5.1_edition_june_2011 Download

Rhino on GitHub

Currently, on GitHub page of Mozilla, ~~version 1.7R4 is not available~~. However, you may find some very old scripts that were written at the time of 1.7R4 as I can validate using web-achieve. You can explore their GitHub repo here.

Source: https://github.com/mozilla/rhino

Rhino Engine Limitations in vRO

When writing scripts for workflows, you must consider the following limitations of the Mozilla Rhino implementation in Orchestrator.

When a workflow runs, the objects that pass from one workflow element to another are not JavaScript objects. What is passed from one element to the next is the serialization of a Java object that has a JavaScript image. As a consequence, you cannot use the whole JavaScript language, but only the classes that are present in the API Explorer. You cannot pass function objects from one workflow element to another.
Orchestrator runs the code in scriptable task elements in a context that is not the Rhino root context. Orchestrator transparently wraps scriptable task elements and actions into JavaScript functions, which it then runs. A scriptable task element that contains System.log(this); does not display the global object this in the same way as a standard Rhino implementation does.
You can only call actions that return nonserializable objects from scripting, and not from workflows. To call an action that returns a nonserializable object, you must write a scriptable task element that calls the action by using the System.getModuleModuleName.action() method.
Workflow validation does not check whether a workflow attribute type is different from an input type of an action or subworkflow. If you change the type of a workflow input parameter, for example from VIM3:VirtualMachine to VC:VirtualMachine, but you do not update any scriptable tasks or actions that use the original input type, the workflow validates but does not run.

Access to additional Java Classes

By default, vRealize Orchestrator restricts JavaScript access to a limited set of Java classes. If you require JavaScript access to a wider range of Java classes, you must set an vRealize Orchestrator system property.

Allowing the JavaScript engine full access to the Java virtual machine (JVM) presents potential security issues. Malformed or malicious scripts might have access to all the system components to which the user who runs the vRealize Orchestrator server has access. Therefore, by default the vRealize Orchestrator JavaScript engine can access only the classes in the java.util.* package.

If you require JavaScript access to classes outside of the java.util.* package, you can list in a configuration file the Java packages to which to allow JavaScript access. You then set the com.vmware.scripting.rhino-class-shutter-file system property to point to this file.

Procedure

Create a text configuration file to store the list of Java packages to which to allow JavaScript access.For example, to allow JavaScript access to all the classes in the java.net package and to the java.lang.Object class, you add the following content to the file.java.net.* java.lang.Object
Enter a name for the configuration file.
Save the configuration file in a subdirectory of /data/vco/usr/lib/vco. The configuration file cannot be saved under another directory.
Log in to Control Center as root.
Click System Properties.
Click New.
In the Key text box, enter com.vmware.scripting.rhino-class-shutter-file.
In the Value text box, enter vco/usr/lib/vco/your_configuration_file_subdirectory.
In the Description text box, enter a description for the system property.
Click Add.
Click Save changes from the pop-up menu.A message indicates that you have saved successfully.
Wait for the vRealize Orchestrator server to restart.

See an implementation example of accessing external Java classes by BlueCat here. Here, the code implements new java.lang.Long(0)

.
.
.
var testConfig = BCNProteusAPI.createAPIEntity(new java.lang.Long(0),configName,"","Configuration" );
var args = new Array( new java.lang.Long(0), testConfig );
configId = new java.lang.Long( BCNProteusAPI.call( profileName,"addEntity",args ));
System.log( "New configuration was created, id=" + configId );
var addTFTPGroupArgs = new Array( configId, "tftpGroupName1", "" );
var tftpGroupId = new java.lang.Long( BCNProteusAPI.call(profileName,"addTFTPGroup", addTFTPGroupArgs ) );
System.log( "New TFTP Group was created, id=" + tftpGroupId );
.
.
.