Jispy

Jispy: A JavaScript interpreter In Python.

Jispy is an interpreter for a strict subset of JavaScript, fondly called LittleJ (LJ). It employs recursive descent for parsing and is very easily extendable.

Built for embedding JavaScript

Jispy's original vision was to seamlessly allow embedding JavaScript programs in Python projects. By default, it doesn't expose the host's file system or any other sensitive element. Some checks on infinite looping and infinite recursion are provided to tackle possibly malicious code.

Maintainance and Licensing

The project is currently maintined by the engineering team at Polydojo. See LICENSE.txt and DEDICATION.md for project license and dedication respectively.

Installation

Simply include jispy.py in your project directory. For convenience, you may wish to also include stdlib.l.js, which provides multiple utilities and manipulators.

Jispy comes armed with a console and an API. While the API provides far greater control, the console is great for getting started.

The console()

You could start the console in two ways:

1. Run jispy.py in a terminal:

 $ python jispy.py

2. From Python's interactive interpreter:

 >>> import jispy
 >>> jispy.console()

In either case, you'll unleash a LittleJ REPL (Read-Evaluate-Print Loop).

LJ> "Hello World!";
"Hello World!"
LJ> var obj = {i: 'am', an: 'object'}, arr = ['i', 'am', 'an', 'array'];
LJ> obj.i === arr[1];
true
LJ> type(obj) === type(arr);
false
LJ> "Because type() of an array is " + type(arr);
"Because type() of an array is array"
LJ> // A function to compute factorial:
LJ> var f = function (n) { if (n <= 1) { return 1; } return n * f(n-1); };
LJ> f(6);
720
LJ> // Functions are first class!!
LJ> var foo = function () { return foo; };
LJ> foo === foo() && foo === foo()();
true
LJ> // Hit Crtl-D to exit
LJ> 

To enter multiple lines of code, end lines with tabs.
Here's an example with for:

LJ> var obj = {a: 'apple', b: 'ball', c: 'cat', d: 'dog'}, // tab..       
...     i = null, keyz = keys(obj);
LJ> for (i = 0; i < len(obj); i += 1) {                    // tab.. 
...     keyz[i] + ' for ' + obj[keyz[i]];                  // tab..       
... }
"a for apple"
"c for cat"
"b for ball"
"d for dog"
LJ> 

Note: Future versions of LittleJ (and hence Jispy) may no longer support for loops in favor of while loops and _.each(.) style iteration. If you'd like to maintain backward compatability, we recommend using array.each from stdlib.l.js.

The above example uses the inbuilt len() and keys() functions. These, along with other inbuilts have been described in LittleJ's specification.

The API

Consider the following LittleJ program for checking if a number is prime:

var isPrime = function (n) {
        var countF = 0, i = null; // countF counts the number of factors
        for (i = 1; i <= n; i += 1) { if (n % i === 0) { countF += 1; } }
        return countF === 2;
    };
print(isPrime(7)); // true
print(isPrime(8)); // false

If the above program has been stored as a Python-string in the variable lj_isPrime, the following python code would execute it.

 # ... set lj_isPrime ...
from jispy import Runtime
rt = Runtime()
rt.runX(lj_isPrime);

Runtimes

A Runtime provides a wrapper around a single global environment or scope. It allows you to run multiple programs in the same environment; at the same time allowing you to run a programs in foreign environments.

Runtime's constructor optionally accepts three arguments (in order):

• maxLoopTime: The maximum time in seconds for which a loop may run.
• maxDepth: The maximum allowable depth of nested environments (scopes).
• writer: write method of a file (opened for writing). It defaults to sys.stdout.write.

maxLoopTime and maxDepth both default to None. Unless set to a positive value, there shall be no checks on infinite looping and/or infinite recursion. If writer is set to None, the inbuilt print() shall not be made available.

Here's an example involving maxLoopTime and maxDepth:

 >>> from jispy import Runtime
 >>> prog1 = """var i = 1; while (true) { i += 0; }"""
 >>> prog2 = """var foo = function () { return foo(); }; foo();"""
 >>> rt = Runtime(maxLoopTime=13, maxDepth=100);
 >>> rt.runX(prog1);
 RuntimeError: looping for to long
 >>> rt.runX(prog2);
 RuntimeError: maximum call depth exceeded
 >>> 

Running programs in the right Runtime:

An instance of Runtime (say rt) provides 3 ways in which you may run a program. There's a method corresponding to each:

• rt.runG() runs programs in rt's global environment.
• rt.runC() runs programs in a first-level child of rt's global environment.
• rt.runX() runs programs in a completely different (foreign) environment.

In LittleJ, once a variable is defined, it may not be redefined in the same environment. Thus, the method used to run a program is significant.

Additionally, there's an rt.run() method, which accepts the environment in which a program should be run. Each of the above listed method internally calls rt.run(). However, calling rt.run() explicitly is not advised.

Input to rt.runG(), rt.runC() & rt.runX():

Each of these methods accept at least 1 and at most 2 inputs (in order):

• prog: A LittleJ program. It must be supplied.
• console: A boolean. (Optional, defaults to False.)

prog may be any one of:

• A Python string in which a LittleJ program is stored. (Like lj_isPrime above.)
• The name of an LittleJ program file. LittleJ program files have the .l.js extension. Any string ending in .l.js is treated as a file name.
• A parse tree generated by Jispy's (lex() and) yacc(). This option is explored in later sections.

Consider:

 >>> import jispy
 >>> rt = jispy.Runtime();
 >>> mappy = """
 ... var map = function (arr, func) {
 ...     var ans = [], i = null;
 ...     for (i = 0; i < len(arr); i += 1) {
 ...             append(ans, func(arr[i]));
 ...     }
 ...     return ans;
 ... };"""
 >>> rt.runG(mappy);
 >>> rt.runG('print(map([1, 2, 3], function (x) { return x*x; }));');
 [1, 4, 9]
 >>> rt.runC('print(map([1, 2, 3], function (x) { return x/2; }));');
 [0.5, 1, 1.5]
 >>> rt.runX('print(map([1, 2, 3], function (x) { return x+2; }));');
 ReferenceError: map is not defined
 >>> 

In the above example, since mappy is run in rt's global environment, the variable map it is available to not only the global environment but also all child environments. However, it is not available in any foreign environment.

Including stdlib.l.js:

stdlib.l.js contains many useful string, array and object related utilities. To use them in your programs, run stdlib.l.js using runG()

More about console()

Each call to Jispy's console() uses a single Runtime. Each input line (or lines) is executed as an independent program in the global environment (using the runG() method). Thus, each line has access to variables defined in earlier an earlier line.

Input to console()

The console accepts three optional arguments (in order):

• rt: The Runtime in which each line (or lines) should be run. Defaults to None, in which case a fresh Runtime is created.
• semify: A boolean, defaults to False. If True (or truthy), the console tries to add semicolons to the end of lines as required.
• prompt: The input prompt. Defaults to "LJ> "

Originally, the console() was not a REPL. It now is. (But the name has stuck.)

console() and stdlib.l.js:

 >>> import jispy
 >>> rt = jispy.Runtime(13, 100);
 >>> rt.runG('stdlib.l.js');
 >>> jispy.console(rt);

stdlib.l.js includes array.map() and array.filter(). As the library was run in the global environment, we may freely use those functions in our LittleJ programs:

 LJ> var a = [5, 8, 3, 4, 6, 7, 1];
 LJ> array.map(array.filter(a, function (x) { return x % 2 === 0; }),    
 ...             function (x) { return x * x * x; });
 [512, 64, 216]
 LJ> // [8^3, 4^3, 6^3]
 LJ> 

Adding Natives

Functions such as print() may be added via Runtime's addNatives method. It accepts a single dictionary as input. The key-value pairs are added to the global environment.

For example, let's add a native version of the factorial function and set native inbuilt_num to 7.0:

 >>> from jispy import Runtime
 >>> factorial = lambda n: 1.0 if n <= 1.0 else n * factorial(n-1)
 >>> rt = Runtime(maxLoopTime = 13, maxDepth = 100)
 >>> rt.addNatives({'factorial' : factorial, 'inbuilt_num' : 7.0})
 >>> demoProgo = 'print(factorial(inbuilt_num));'
 >>> rt.runC(demoProgo)
 5040
 >>> 

Please note that the LittleJ program may reassign all natives; and a native variable may be initialized (as a fresh variable) in a new environment.

Adding a native input() function:

 >>> import jispy as j
 >>> rt = j.Runtime();
 >>> rt.addNatives({'input': lambda prompt: raw_input(prompt)});
 >>> j.console(rt); # use input() on Python 3+

 LJ> var name = input('Please enter your name: ');
 Please enter your name: Awesome Person
 LJ> print('Dear ' + name + ', THANK YOU FOR READING THIS FAR!');
 Dear Awesome Person, THANK YOU FOR READING THIS FAR!
 LJ> 

Known Issues:

1. Phony && and || operators:

The && and ||, operators behave like the following functions:

var andand = function (a, b) { if (a) { return b; } return a; },
    oror   = function (a, b) { if (a) { return a; } return b; };

That is, both operands to these operators are ALWAYS evaluated.

Let's say we want to write a function pSquare which squares only positive numbers. If a non-positive number is passed as input, 'bad input' should be returned.

Here's an instinctive version of pSquare:

LJ> var pSquare = function (n) {        
...     if (type(n) === 'number' && n > 0) { return n * n; }    
...     return 'bad input';    
... };
LJ> pSquare('what?');
TypeError: bad operands for binary > ... "string" === "number" && "what?" > 0
LJ> 

With n as 'what?', since type('what?') === 'number' is false, the condition 'what?' > 0 should not be evaluated. But Jispy (incorrectly) tries to evaluate it and raises a TypeError in the process.

The workaround is to use conditionals:

LJ> var pSquare = function (n) {        
...     if (type(n) === 'number') { if (n > 0) { return n * n; } }      
...     return 'bad input';     
... };
LJ> pSquare('what?');
"bad input"
LJ> 

Fixing && and || will requires partially re-writing the parser and the interpreter. There are no immediate plans for such a fix.

2. Trailing dots aren't handled:

LJ> var obj = {alphas: {a: 'apple', b: 'ball'}};
LJ> obj.alphas.a;           // works as expected
"apple"
LJ> obj    .alphas    .a;   // also works as expected (leading dots)
"apple"
LJ> obj.    alphas.    a;   // doesn't yet work (trailing dots)
SyntaxError: unexpected trailing . (dot) obj.
LJ> // This issue usually pops up in refinements spread over multiple lines:
LJ> obj.alphas. 
...             a;
SyntaxError: unexpected trailing . (dot) obj.alphas.
LJ> 

The workaround is to use leading dots only.

2.1 The plague of semicolon insertion:

Due to JavaScript's semicolon insertion, using leading dots may change the meaning of your program (in JavaScript). Thus, spreading refinements over multiple lines is not advisable.

LittleJ does not (and will not) include semicolon insertion. This is one area where the semantic meaning of a LittleJ program may not be retained in JavaScript. Ways of mitigating this issue are currently being explored.

---

That's all folks!

Thank you for showing interest in Jispy.

You are welcome to experiment with Jispy. PLEASE FEEL FREE TO CONTRIBUTE.

If you use Jispy in any of your projects, PLEASE LET US KNOW.

---