NAME

JavaScript::SpiderMonkey - Perl interface to the JavaScript Engine

SYNOPSIS

use JavaScript::SpiderMonkey;

my $js = JavaScript::SpiderMonkey->new();

$js->init();  # Initialize Runtime/Context

              # Define a perl callback for a new JavaScript function
$js->function_set("print_to_perl", sub { print "@_\n"; });

              # Create a new (nested) object and a property
$js->property_by_path("document.location.href");

              # Execute some code
my $rc = $js->eval(q!
    document.location.href = append("http://", "www.aol.com");

    print_to_perl("URL is ", document.location.href);

    function append(first, second) {
         return first + second;
    }
!);

    # Get the value of a property set in JS
my $url = $js->property_get("document.location.href");

$js->destroy();

INSTALL

JavaScript::SpiderMonkey requires Mozilla's readily compiled SpiderMonkey 1.5 distribution or better. Please check "SpiderMonkey Installation".

DESCRIPTION

JavaScript::SpiderMonkey is a Perl Interface to the SpiderMonkey JavaScript Engine. It is different from Claes Jacobsson's JavaScript.pm in that it offers two different levels of access:

[1]

A 1:1 mapping of the SpiderMonkey API to Perl

[2]

A more Perl-like API

This document describes [2], for [1], please check SpiderMonkey.xs.

new()

$js = JavaScript::SpiderMonkey->new() creates a new object to work with. To initialize the JS runtime, call $js->init() afterwards.

$js->destroy()

$js->destroy() destroys the current runtime and frees up all memory.

$js->init()

$js->init() initializes the SpiderMonkey engine by creating a context, default classes and objects and adding an error reporter.

$js->array_by_path($name)

Creates an object of type Array in the JS runtime:

$js->array_by_path("document.form");

will first create an object with the name document (unless it exists already) and then define a property named form to it, which is an object of type Array. Therefore, in the JS code, you're going to be able define things like

document.form[0] = "value";

$js->function_set($name, $funcref, [$obj])

Binds a Perl function provided as a coderef ($funcref) to a newly created JS function named $name in JS land. It's a real function (therefore bound to the global object) if $obj is omitted. However, if $obj is ref to a JS object (retrieved via $js->object_by_path($path) or the like), the function will be a method of the specified object.

$js->function_set("write", sub { print @_ });
    # write("hello"); // In JS land

$obj = $j->object_by_path("navigator");
$js->function_set("write", sub { print @_ }, $obj);
    # navigator.write("hello"); // In JS land

$js->array_set_element($obj, $idx, $val)

Sets the element of the array $obj at index position $idx to the value $val. $obj is a reference to an object of type array (retrieved via $js->object_by_path($path) or the like).

$js->array_set_element_as_object($obj, $idx, $elobj)

Sets the element of the array $obj at index position $idx to the object $elobj (both $obj and $elobj have been retrieved via $js->object_by_path($path) or the like).

$js->array_get_element($obj, $idx)

Gets the value of of the element at index $idx of the object of type Array $obj.

$js->property_by_path($path, $value, [$getter], [$setter])

Sets the specified property of an object in $path to the value $value. $path is the full name of the property, including the object(s) in JS land it belongs to:

$js-E<gt>property_by_path("document.location.href", "abc");

This first creates the object document (if it doesn't exist already), then the object document.location, then attaches the property href to it and sets it to "abc".

$getter and $setter are coderefs that will be called by the JavaScript engine when the respective property's value is requested or set:

sub getter {
    my($property_path, $value) = @_;
    print "$property_path has value $value\n";
}

sub setter {
    my($property_path, $value) = @_;
    print "$property_path set to value $value\n";
}

$js->property_by_path("document.location.href", "abc",
                          \&getter, \&setter);

If you leave out $getter and $setter, no callbacks are going to be triggered while the property is set or queried. If you just want to specify a $setter, but no $getter, set the $getter to undef.

$js->object_by_path($path, [$newobj])

Get a pointer to an object with the path specified. Create it if it's not there yet. If $newobj is provided, the ref is used to bind the existing object to the name in $path.

$js->property_get($path)

Fetch the property specified by the $path.

my $val = $js->property_get("document.location.href");

$js->eval($code)

Runs the specified piece of <$code> in the JS engine. Afterwards, property values of objects previously defined will be available via $j->property_get($path) and the like.

my $rc = $js->eval("write('hello');");

The method returns 1 on success or else if there was an error in JS land. In case of an error, the JS error text will be available in $@.

$js->set_max_branch_operations($max_branch_operations)

Set the maximum number of allowed branch operations. This protects against infinite loops and guarantees that the eval operation will terminate.

SpiderMonkey Installation

First, get the latest SpiderMonkey distribution from mozilla.org: http://www.mozilla.org/js/spidermonkey shows which releases are available. js-1.6.tar.gz has been proven to work.

Untar it at the same directory level as you just untarred the JavaScript::SpiderMonkey distribution you're currently reading. So, if you're currently in /my/path/JavaScript-SpiderMonkey-v.vv, do this:

cp js-1.6.tar.gz /my/path
cd /my/path
tar zxfv js-1.6.tar.gz

Then, compile the SpiderMonkey distribution, if you're on Linux, just use:

cd js/src
make -f Makefile.ref

It's important that the js and JavaScript-SpiderMonkey-v.vv directories are at the same level:

[/my/path]$ ls
JavaScript-SpiderMonkey-v.vv
js
js-1.6.tar.gz
[/my/path]$

(Note that you *can* untar the SpiderMonkey distribution elsewhere, but, if so, then you need to edit the setting of $JSLIBPATH in Makefile.PL).

Next, you need to copy the shared library file thus constructed (e.g., libjs.so or js32.dll) to an appropriate directory on your library path. On Windows, this can also be the directory where the perl executable lives. On Unix, this has been shown to work without copying, but this way you need to keep the compiled binary in the js build directory forever. Copying js/src/Your_OS_DBG.OBJ/libjs.so to /usr/local/lib and making sure that /usr/local/lib is in your LD_LIBRARY_PATH seems to be safest bet.

Now, build JavaScript::SpiderMonkey in the standard way:

cd JavaScript-SpiderMonkey-v.vv
perl Makefile.PL
make
make test
make install

E4X SUPPORT

To build JavaScript-SpiderMonkey with E4X (ECMAScript for XML) support:

perl Makefile.PL -E4X

Please note that E4X support is only supported since SpiderMonkey release 1.6.

THREAD SAFETY

To build JavaScript-SpiderMonkey when using a thread safe version of SpiderMonkey:

perl Makefile.PL -JS_THREADSAFE

AUTHORS

Mike Schilli, <m at perlmeister dot com>
Thomas Busch, <tbusch at cpan dot org> (current maintainer)

COPYRIGHT AND LICENSE

Copyright (c) 2002-2005 Mike Schilli
Copyright (c) 2006-2007 Thomas Busch

This library is free software; you can redistribute it and/or modify it under the same terms as Perl itself.