otto.go

/*
Package otto is a JavaScript parser and interpreter written natively in Go.

http://godoc.org/github.com/robertkrimen/otto

	import (
	    "github.com/robertkrimen/otto"
	)

Run something in the VM

	vm := otto.New()
	vm.Run(`
	    abc = 2 + 2;
		console.log("The value of abc is " + abc); // 4
	`)

Get a value out of the VM

	value, err := vm.Get("abc")
		value, _ := value.ToInteger()
	}

Set a number

	vm.Set("def", 11)
	vm.Run(`
		console.log("The value of def is " + def);
		// The value of def is 11
	`)

Set a string

	vm.Set("xyzzy", "Nothing happens.")
	vm.Run(`
		console.log(xyzzy.length); // 16
	`)

Get the value of an expression

	value, _ = vm.Run("xyzzy.length")
	{
		// value is an int64 with a value of 16
		value, _ := value.ToInteger()
	}

An error happens

	value, err = vm.Run("abcdefghijlmnopqrstuvwxyz.length")
	if err != nil {
		// err = ReferenceError: abcdefghijlmnopqrstuvwxyz is not defined
		// If there is an error, then value.IsUndefined() is true
		...
	}

Set a Go function

	vm.Set("sayHello", func(call otto.FunctionCall) otto.Value {
	    fmt.Printf("Hello, %s.\n", call.Argument(0).String())
	    return otto.Value{}
	})

Set a Go function that returns something useful

	vm.Set("twoPlus", func(call otto.FunctionCall) otto.Value {
	    right, _ := call.Argument(0).ToInteger()
	    result, _ := vm.ToValue(2 + right)
	    return result
	})

Use the functions in JavaScript

	result, _ = vm.Run(`
	    sayHello("Xyzzy");      // Hello, Xyzzy.
	    sayHello();             // Hello, undefined

	    result = twoPlus(2.0); // 4
	`)

# Parser

A separate parser is available in the parser package if you're just interested in building an AST.

http://godoc.org/github.com/robertkrimen/otto/parser

Parse and return an AST

	filename := "" // A filename is optional
	src := `
	    // Sample xyzzy example
	    (function(){
	        if (3.14159 > 0) {
	            console.log("Hello, World.");
	            return;
	        }

	        var xyzzy = NaN;
	        console.log("Nothing happens.");
	        return xyzzy;
	    })();
	`

	// Parse some JavaScript, yielding a *ast.Program and/or an ErrorList
	program, err := parser.ParseFile(nil, filename, src, 0)

otto

You can run (Go) JavaScript from the commandline with: http://github.com/robertkrimen/otto/tree/master/otto

	$ go get -v github.com/robertkrimen/otto/otto

Run JavaScript by entering some source on stdin or by giving otto a filename:

	$ otto example.js

underscore

Optionally include the JavaScript utility-belt library, underscore, with this import:

	import (
		"github.com/robertkrimen/otto"
		_ "github.com/robertkrimen/otto/underscore"
	)

	// Now every otto runtime will come loaded with underscore

For more information: http://github.com/robertkrimen/otto/tree/master/underscore

# Caveat Emptor

The following are some limitations with otto:

  - "use strict" will parse, but does nothing.
  - The regular expression engine (re2/regexp) is not fully compatible with the ECMA5 specification.
  - Otto targets ES5. ES6 features (eg: Typed Arrays) are not supported.

# Regular Expression Incompatibility

Go translates JavaScript-style regular expressions into something that is "regexp" compatible via `parser.TransformRegExp`.
Unfortunately, RegExp requires backtracking for some patterns, and backtracking is not supported by the standard Go engine: https://code.google.com/p/re2/wiki/Syntax

Therefore, the following syntax is incompatible:

	(?=)  // Lookahead (positive), currently a parsing error
	(?!)  // Lookahead (backhead), currently a parsing error
	\1    // Backreference (\1, \2, \3, ...), currently a parsing error

A brief discussion of these limitations: "Regexp (?!re)" https://groups.google.com/forum/?fromgroups=#%21topic/golang-nuts/7qgSDWPIh_E

More information about re2: https://code.google.com/p/re2/

In addition to the above, re2 (Go) has a different definition for \s: [\t\n\f\r ].
The JavaScript definition, on the other hand, also includes \v, Unicode "Separator, Space", etc.

# Halting Problem

If you want to stop long running executions (like third-party code), you can use the interrupt channel to do this:

	package main

	import (
	    "errors"
	    "fmt"
	    "os"
	    "time"

	    "github.com/robertkrimen/otto"
	)

	var halt = errors.New("Stahp")

	func main() {
	    runUnsafe(`var abc = [];`)
	    runUnsafe(`
	    while (true) {
	        // Loop forever
	    }`)
	}

	func runUnsafe(unsafe string) {
	    start := time.Now()
	    defer func() {
	        duration := time.Since(start)
	        if caught := recover(); caught != nil {
	            if caught == halt {
	                fmt.Fprintf(os.Stderr, "Some code took to long! Stopping after: %v\n", duration)
	                return
	            }
	            panic(caught) // Something else happened, repanic!
	        }
	        fmt.Fprintf(os.Stderr, "Ran code successfully: %v\n", duration)
	    }()

	    vm := otto.New()
	    vm.Interrupt = make(chan func(), 1) // The buffer prevents blocking

	    go func() {
	        time.Sleep(2 * time.Second) // Stop after two seconds
	        vm.Interrupt <- func() {
	            panic(halt)
	        }
	    }()

	    vm.Run(unsafe) // Here be dragons (risky code)
	}

Where is setTimeout/setInterval?

These timing functions are not actually part of the ECMA-262 specification. Typically, they belong to the `windows` object (in the browser).
It would not be difficult to provide something like these via Go, but you probably want to wrap otto in an event loop in that case.

For an example of how this could be done in Go with otto, see natto:

http://github.com/robertkrimen/natto

Here is some more discussion of the issue:

* http://book.mixu.net/node/ch2.html

* http://en.wikipedia.org/wiki/Reentrancy_%28computing%29

* http://aaroncrane.co.uk/2009/02/perl_safe_signals/
*/
package otto

import (
	"encoding/json"
	"fmt"
	"strings"

	"github.com/robertkrimen/otto/file"
	"github.com/robertkrimen/otto/registry"
)

// Otto is the representation of the JavaScript runtime. Each instance of Otto has a self-contained namespace.
type Otto struct {
	// Interrupt is a channel for interrupting the runtime. You can use this to halt a long running execution, for example.
	// See "Halting Problem" for more information.
	Interrupt chan func()
	runtime   *_runtime
}

// New will allocate a new JavaScript runtime
func New() *Otto {
	self := &Otto{
		runtime: newContext(),
	}
	self.runtime.otto = self
	self.runtime.traceLimit = 10
	self.Set("console", self.runtime.newConsole())

	registry.Apply(func(entry registry.Entry) {
		self.Run(entry.Source())
	})

	return self
}

func (otto *Otto) clone() *Otto {
	self := &Otto{
		runtime: otto.runtime.clone(),
	}
	self.runtime.otto = self
	return self
}

// Run will allocate a new JavaScript runtime, run the given source
// on the allocated runtime, and return the runtime, resulting value, and
// error (if any).
//
// src may be a string, a byte slice, a bytes.Buffer, or an io.Reader, but it MUST always be in UTF-8.
//
// src may also be a Script.
//
// src may also be a Program, but if the AST has been modified, then runtime behavior is undefined.
func Run(src interface{}) (*Otto, Value, error) {
	otto := New()
	value, err := otto.Run(src) // This already does safety checking
	return otto, value, err
}

// Run will run the given source (parsing it first if necessary), returning the resulting value and error (if any)
//
// src may be a string, a byte slice, a bytes.Buffer, or an io.Reader, but it MUST always be in UTF-8.
//
// If the runtime is unable to parse source, then this function will return undefined and the parse error (nothing
// will be evaluated in this case).
//
// src may also be a Script.
//
// src may also be a Program, but if the AST has been modified, then runtime behavior is undefined.
func (self Otto) Run(src interface{}) (Value, error) {
	value, err := self.runtime.cmpl_run(src, nil)
	if !value.safe() {
		value = Value{}
	}
	return value, err
}

// Eval will do the same thing as Run, except without leaving the current scope.
//
// By staying in the same scope, the code evaluated has access to everything
// already defined in the current stack frame. This is most useful in, for
// example, a debugger call.
func (self Otto) Eval(src interface{}) (Value, error) {
	if self.runtime.scope == nil {
		self.runtime.enterGlobalScope()
		defer self.runtime.leaveScope()
	}

	value, err := self.runtime.cmpl_eval(src, nil)
	if !value.safe() {
		value = Value{}
	}
	return value, err
}

// Get the value of the top-level binding of the given name.
//
// If there is an error (like the binding does not exist), then the value
// will be undefined.
func (self Otto) Get(name string) (Value, error) {
	value := Value{}
	err := catchPanic(func() {
		value = self.getValue(name)
	})
	if !value.safe() {
		value = Value{}
	}
	return value, err
}

func (self Otto) getValue(name string) Value {
	return self.runtime.globalStash.getBinding(name, false)
}

// Set the top-level binding of the given name to the given value.
//
// Set will automatically apply ToValue to the given value in order
// to convert it to a JavaScript value (type Value).
//
// If there is an error (like the binding is read-only, or the ToValue conversion
// fails), then an error is returned.
//
// If the top-level binding does not exist, it will be created.
func (self Otto) Set(name string, value interface{}) error {
	{
		value, err := self.ToValue(value)
		if err != nil {
			return err
		}
		err = catchPanic(func() {
			self.setValue(name, value)
		})
		return err
	}
}

func (self Otto) setValue(name string, value Value) {
	self.runtime.globalStash.setValue(name, value, false)
}

func (self Otto) SetDebuggerHandler(fn func(vm *Otto)) {
	self.runtime.debugger = fn
}

func (self Otto) SetRandomSource(fn func() float64) {
	self.runtime.random = fn
}

// SetStackDepthLimit sets an upper limit to the depth of the JavaScript
// stack. In simpler terms, this limits the number of "nested" function calls
// you can make in a particular interpreter instance.
//
// Note that this doesn't take into account the Go stack depth. If your
// JavaScript makes a call to a Go function, otto won't keep track of what
// happens outside the interpreter. So if your Go function is infinitely
// recursive, you're still in trouble.
func (self Otto) SetStackDepthLimit(limit int) {
	self.runtime.stackLimit = limit
}

// SetStackTraceLimit sets an upper limit to the number of stack frames that
// otto will use when formatting an error's stack trace. By default, the limit
// is 10. This is consistent with V8 and SpiderMonkey.
//
// TODO: expose via `Error.stackTraceLimit`
func (self Otto) SetStackTraceLimit(limit int) {
	self.runtime.traceLimit = limit
}

// MakeCustomError creates a new Error object with the given name and message,
// returning it as a Value.
func (self Otto) MakeCustomError(name, message string) Value {
	return self.runtime.toValue(self.runtime.newError(name, self.runtime.toValue(message), 0))
}

// MakeRangeError creates a new RangeError object with the given message,
// returning it as a Value.
func (self Otto) MakeRangeError(message string) Value {
	return self.runtime.toValue(self.runtime.newRangeError(self.runtime.toValue(message)))
}

// MakeSyntaxError creates a new SyntaxError object with the given message,
// returning it as a Value.
func (self Otto) MakeSyntaxError(message string) Value {
	return self.runtime.toValue(self.runtime.newSyntaxError(self.runtime.toValue(message)))
}

// MakeTypeError creates a new TypeError object with the given message,
// returning it as a Value.
func (self Otto) MakeTypeError(message string) Value {
	return self.runtime.toValue(self.runtime.newTypeError(self.runtime.toValue(message)))
}

// Context is a structure that contains information about the current execution
// context.
type Context struct {
	Filename   string
	Line       int
	Column     int
	Callee     string
	Symbols    map[string]Value
	This       Value
	Stacktrace []string
}

// Context returns the current execution context of the vm, traversing up to
// ten stack frames, and skipping any innermost native function stack frames.
func (self Otto) Context() Context {
	return self.ContextSkip(10, true)
}

// ContextLimit returns the current execution context of the vm, with a
// specific limit on the number of stack frames to traverse, skipping any
// innermost native function stack frames.
func (self Otto) ContextLimit(limit int) Context {
	return self.ContextSkip(limit, true)
}

// ContextSkip returns the current execution context of the vm, with a
// specific limit on the number of stack frames to traverse, optionally
// skipping any innermost native function stack frames.
func (self Otto) ContextSkip(limit int, skipNative bool) (ctx Context) {
	// Ensure we are operating in a scope
	if self.runtime.scope == nil {
		self.runtime.enterGlobalScope()
		defer self.runtime.leaveScope()
	}

	scope := self.runtime.scope
	frame := scope.frame

	for skipNative && frame.native && scope.outer != nil {
		scope = scope.outer
		frame = scope.frame
	}

	// Get location information
	ctx.Filename = "<unknown>"
	ctx.Callee = frame.callee

	switch {
	case frame.native:
		ctx.Filename = frame.nativeFile
		ctx.Line = frame.nativeLine
		ctx.Column = 0
	case frame.file != nil:
		ctx.Filename = "<anonymous>"

		if p := frame.file.Position(file.Idx(frame.offset)); p != nil {
			ctx.Line = p.Line
			ctx.Column = p.Column

			if p.Filename != "" {
				ctx.Filename = p.Filename
			}
		}
	}

	// Get the current scope this Value
	ctx.This = toValue_object(scope.this)

	// Build stacktrace (up to 10 levels deep)
	ctx.Symbols = make(map[string]Value)
	ctx.Stacktrace = append(ctx.Stacktrace, frame.location())
	for limit != 0 {
		// Get variables
		stash := scope.lexical
		for {
			for _, name := range getStashProperties(stash) {
				if _, ok := ctx.Symbols[name]; !ok {
					ctx.Symbols[name] = stash.getBinding(name, true)
				}
			}
			stash = stash.outer()
			if stash == nil || stash.outer() == nil {
				break
			}
		}

		scope = scope.outer
		if scope == nil {
			break
		}
		if scope.frame.offset >= 0 {
			ctx.Stacktrace = append(ctx.Stacktrace, scope.frame.location())
		}
		limit--
	}

	return
}

// Call the given JavaScript with a given this and arguments.
//
// If this is nil, then some special handling takes place to determine the proper
// this value, falling back to a "standard" invocation if necessary (where this is
// undefined).
//
// If source begins with "new " (A lowercase new followed by a space), then
// Call will invoke the function constructor rather than performing a function call.
// In this case, the this argument has no effect.
//
//	// value is a String object
//	value, _ := vm.Call("Object", nil, "Hello, World.")
//
//	// Likewise...
//	value, _ := vm.Call("new Object", nil, "Hello, World.")
//
//	// This will perform a concat on the given array and return the result
//	// value is [ 1, 2, 3, undefined, 4, 5, 6, 7, "abc" ]
//	value, _ := vm.Call(`[ 1, 2, 3, undefined, 4 ].concat`, nil, 5, 6, 7, "abc")
func (self Otto) Call(source string, this interface{}, argumentList ...interface{}) (Value, error) {
	thisValue := Value{}

	construct := false
	if strings.HasPrefix(source, "new ") {
		source = source[4:]
		construct = true
	}

	// FIXME enterGlobalScope
	self.runtime.enterGlobalScope()
	defer func() {
		self.runtime.leaveScope()
	}()

	if !construct && this == nil {
		program, err := self.runtime.cmpl_parse("", source+"()", nil)
		if err == nil {
			if node, ok := program.body[0].(*_nodeExpressionStatement); ok {
				if node, ok := node.expression.(*_nodeCallExpression); ok {
					var value Value
					err := catchPanic(func() {
						value = self.runtime.cmpl_evaluate_nodeCallExpression(node, argumentList)
					})
					if err != nil {
						return Value{}, err
					}
					return value, nil
				}
			}
		}
	} else {
		value, err := self.ToValue(this)
		if err != nil {
			return Value{}, err
		}
		thisValue = value
	}

	{
		this := thisValue

		fn, err := self.Run(source)
		if err != nil {
			return Value{}, err
		}

		if construct {
			result, err := fn.constructSafe(self.runtime, this, argumentList...)
			if err != nil {
				return Value{}, err
			}
			return result, nil
		}

		result, err := fn.Call(this, argumentList...)
		if err != nil {
			return Value{}, err
		}
		return result, nil
	}
}

// Object will run the given source and return the result as an object.
//
// For example, accessing an existing object:
//
//	object, _ := vm.Object(`Number`)
//
// Or, creating a new object:
//
//	object, _ := vm.Object(`({ xyzzy: "Nothing happens." })`)
//
// Or, creating and assigning an object:
//
//	object, _ := vm.Object(`xyzzy = {}`)
//	object.Set("volume", 11)
//
// If there is an error (like the source does not result in an object), then
// nil and an error is returned.
func (self Otto) Object(source string) (*Object, error) {
	value, err := self.runtime.cmpl_run(source, nil)
	if err != nil {
		return nil, err
	}
	if value.IsObject() {
		return value.Object(), nil
	}
	return nil, fmt.Errorf("value is not an object")
}

// ToValue will convert an interface{} value to a value digestible by otto/JavaScript.
func (self Otto) ToValue(value interface{}) (Value, error) {
	return self.runtime.safeToValue(value)
}

// Copy will create a copy/clone of the runtime.
//
// Copy is useful for saving some time when creating many similar runtimes.
//
// This method works by walking the original runtime and cloning each object, scope, stash,
// etc. into a new runtime.
//
// Be on the lookout for memory leaks or inadvertent sharing of resources.
func (in *Otto) Copy() *Otto {
	out := &Otto{
		runtime: in.runtime.clone(),
	}
	out.runtime.otto = out
	return out
}

// Object{}

// Object is the representation of a JavaScript object.
type Object struct {
	object *_object
	value  Value
}

func _newObject(object *_object, value Value) *Object {
	// value MUST contain object!
	return &Object{
		object: object,
		value:  value,
	}
}

// Call a method on the object.
//
// It is essentially equivalent to:
//
//	var method, _ := object.Get(name)
//	method.Call(object, argumentList...)
//
// An undefined value and an error will result if:
//
//  1. There is an error during conversion of the argument list
//  2. The property is not actually a function
//  3. An (uncaught) exception is thrown
func (self Object) Call(name string, argumentList ...interface{}) (Value, error) {
	// TODO: Insert an example using JavaScript below...
	// e.g., Object("JSON").Call("stringify", ...)

	function, err := self.Get(name)
	if err != nil {
		return Value{}, err
	}
	return function.Call(self.Value(), argumentList...)
}

// Value will return self as a value.
func (self Object) Value() Value {
	return self.value
}

// Get the value of the property with the given name.
func (self Object) Get(name string) (Value, error) {
	value := Value{}
	err := catchPanic(func() {
		value = self.object.get(name)
	})
	if !value.safe() {
		value = Value{}
	}
	return value, err
}

// Set the property of the given name to the given value.
//
// An error will result if the setting the property triggers an exception (i.e. read-only),
// or there is an error during conversion of the given value.
func (self Object) Set(name string, value interface{}) error {
	{
		value, err := self.object.runtime.safeToValue(value)
		if err != nil {
			return err
		}
		err = catchPanic(func() {
			self.object.put(name, value, true)
		})
		return err
	}
}

// Keys gets the keys for the given object.
//
// Equivalent to calling Object.keys on the object.
func (self Object) Keys() []string {
	var keys []string
	self.object.enumerate(false, func(name string) bool {
		keys = append(keys, name)
		return true
	})
	return keys
}

// KeysByParent gets the keys (and those of the parents) for the given object,
// in order of "closest" to "furthest".
func (self Object) KeysByParent() [][]string {
	var a [][]string

	for o := self.object; o != nil; o = o.prototype {
		var l []string

		o.enumerate(false, func(name string) bool {
			l = append(l, name)
			return true
		})

		a = append(a, l)
	}

	return a
}

// Class will return the class string of the object.
//
// The return value will (generally) be one of:
//
//	Object
//	Function
//	Array
//	String
//	Number
//	Boolean
//	Date
//	RegExp
func (self Object) Class() string {
	return self.object.class
}

func (self Object) MarshalJSON() ([]byte, error) {
	var goValue interface{}
	switch value := self.object.value.(type) {
	case *_goStructObject:
		goValue = value.value.Interface()
	case *_goMapObject:
		goValue = value.value.Interface()
	case *_goArrayObject:
		goValue = value.value.Interface()
	case *_goSliceObject:
		goValue = value.value.Interface()
	default:
		// It's a JS object; pass it to JSON.stringify:
		var result []byte
		err := catchPanic(func() {
			resultVal := builtinJSON_stringify(FunctionCall{
				runtime:      self.object.runtime,
				ArgumentList: []Value{self.value},
			})
			result = []byte(resultVal.String())
		})
		return result, err
	}
	return json.Marshal(goValue)
}