Fuzzing · 🐰

It is tool to perform fuzz testing. Sometimes it's hard to understand if your function would crash if you pass null or undefined or any unusual value into it. To check that you're doing right you can use this package for stress testing your functions and APIs under different payloads.

Installation

First of all you need to install package like:

npm i fuzzing -SD

Usage

To start fuzzing follow few steps:

• pick function you want to fuzz
• select set of input values, which will be used as arguments passed into the function - input presets
• choose output type if you want to see errors or warnings produced after function execution - output types

import {fuzz} from 'fuzzing';

/**
 * Let's assume that you want to test your perfectly written sum function to find some bugs or unexpected behaviours 
 */
function sum(arr) {
   return arr.reduce((accumulator, item) => accumulator + item, 0);
}

const errors = fuzz(sum) // pick function you want to fuzz
   .numberArray()        // select preset of input values
   .errors();            // choose output type

// print result to see what's going on
console.log(errors);

// it would give you array of executed tests
// showing your input values, test results and stack trace in case of any error
/*
Array [
  Object {
    "description": "SUCCESS: Function returned result is OK and no errors happened",
    "input": Array [
      2.718281828459045,
      3.141592653589793,
      0.6931471805599453,
      2.302585092994046,
      1.4426950408889634,
      0.4342944819032518,
      0.7071067811865476,
      1.4142135623730951,
    ],
    "result": 12.853916621954687,
    "type": "success",
  },
  Object {
    "description": "WARNING: Function returned result might be nullable or dangerous in some way",
    "input": Array [
      undefined,
      undefined,
      undefined,
      undefined,
      undefined,
      undefined,
      undefined,
      undefined,
      undefined,
      undefined,
      undefined,
      undefined,
      undefined,
      13,
    ],
    "result": NaN,
    "type": "warning",
  },
  Object {
    "description": "WARNING: Function returned result might be nullable or dangerous in some way",
    "input": Array [
      0,
      1,
      -1,
      5e-324,
      1.7976931348623157e+308,
      NaN,
      -Infinity,
      Infinity,
    ],
    "result": NaN,
    "type": "warning",
  },
  Object {
    "description": "FAILED: Function execution failed, check error stack trace",
    "error": [TypeError: Cannot read property 'length' of undefined],
    "input": undefined,
    "type": "error",
  },
  Object {
    "description": "FAILED: Function execution failed, check error stack trace",
    "error": [TypeError: Cannot read property 'length' of null],
    "input": null,
    "type": "error",
  },
]
*/

Passing Multiple Parameters

To start fuzzing functions which expects multiple arguments you should import preset const which contains all the presets.

import {fuzz, preset} from 'fuzzing';

function multiply(a, b) {
   return a * b;
}

const warnings = fuzz(multiply)
   .under(preset.number(), preset.number())    // select presets for each function argument
   .errors();
   
console.log(warnings);

Testing API's

The same way you can use fuzzing to call API endpoint with different payloads to test the behavior of your web server.

import {fuzz} from 'fuzzing';

/**
 * For example you want to ping github
 */
function pingGithub(url) {
    return fetch('https://github.io/' + url, { mode: 'no-cors' });
}

const errors = await fuzz(pingGithub)
   .string()
   .errors();

console.log(errors);

// OR

fuzz(pingGithub)
   .string()
   .errors()
   .then(console.log);

Presets of input parameters

Available sets of parameters:

• boolean - Boolean
• number - Number
• string - String
• object - Object
• booleanArray - Array of a booleans
• numberArray - Array of a numbers
• stringArray - Array of strings
• objectArray - Array of objects
• all - All the data sets

Types of output

Available types of output are available as array of result items:

• successes - for passed tests
• warnings - for tests resulted with tricky or danger returned value
• errors - for failed tests
• all - for all tests

Contribution

Feel free to give a valuable feedback ❤️ Igor Golopolosov