errors

Errors is a comprehensive and robust, yet lightweight, set of error utilities for node.js enabling you to do errors more effectively.

Features

• Parameterized error factory allowing you do define how errors should behave based on your project needs.
• Support for enterprise level error attributes including more detailed error cause and operator response messages.
• Predefined error constructors for all HTTP 4xx-5xx based errors allowing you to leverage HTTP errors out of the box.
• express.js integration permitting your code to send() any type of Error directly using Express's response object.
• connect.js support allowing you to use custom errors with connect's errorHandler middleware or this libraries custom error handler middleware.
• Error mapping via registered mapping function permitting you to map between errors when needed.

Installation

Install using npm:

$ npm install errors

Running the tests

From the errors directory first install the dev dependencies:

npm install

Then run the tests:

npm test

API Docs

The API docs are provided in html and md format and are located under errors/docs/. If you want to rebuild them for any reason, you can run the following from the errors directory:

make doc

Defining error messages

The examples assume you've required the errors module like so:

require('errors');

Create a very barebones error -- you must specify at least the error name:

// barebones
errors.create({name: 'RuntimeError'});
console.log(new errors.RuntimeError().toString());

produces:

RuntimeError: An unexpected RuntimeError occurred.
Code: 601

You can define a default message for the error:

// default message
errors.create({
    name: 'RuntimeError',
    defaultMessage: 'A runtime error occurred during processing'
});
console.log(new errors.RuntimeError().toString());

which outputs:

RuntimeError: A runtime error occurred during processing
Code: 602

Define a default message, explanation and response:

// default message, explanation and response
errors.create({
    name: 'FileNotFoundError',
    defaultMessage: 'The requested file could not be found',
    defaultExplanation: 'The file /home/boden/foo could not be found',
    defaultResponse: 'Verify the file exists and retry the operation'
});
console.log(new errors.FileNotFoundError().toString());

gives us:

FileNotFoundError: The requested file could not be found
Code: 603
Explanation: The file /home/boden/foo could not be found
Response: Verify the file exists and retry the operation

Override messages on instantiation:

// override messages
console.log(new errors.FileNotFoundError(
        'Cannot read file'
        , 'You do not have access to read /root/foo'
        , 'Request a file you have permissions to access').toString());

outputs:

FileNotFoundError: Cannot read file
Code: 603
Explanation: You do not have access to read /root/foo
Response: Request a file you have permissions to access

Use the options style constructor to assign standard properties:

console.log(new errors.Http401Error({
    message: "Expired Token",
    explanation: "Your token has expired"}).toString());

outputs:

Http401Error: Expired Token
Code: 401
Explanation: Your token has expired
Error: Expired Token

Using the options style constructor you can also assign arbitrary non-standard properties:

console.log(new errors.Http401Error({
    message: "Expired Token",
    explanation: "Your token has expired",
    expired: new Date()}).toString());

outputs:

Http401Error: Expired Token
Code: 401
Explanation: Your token has expired
expired: Fri Jun 20 2014 04:19:41 GMT-0400 (EDT)

Note however that you cannot assign values to the stack, name or code standard property:

console.log(new errors.Http401Error({
    name: "ExpiredToken"}).toString());

outputs:

/home/boden/workspace/errors/lib/errors.js:261
                throw Error("Properties 'stack', 'name' or 'code' " +
                      ^
Error: Properties 'stack', 'name' or 'code' cannot be overridden
    at Error (<anonymous>)
    at new scope.(anonymous function) (/home/boden/workspace/errors/lib/errors.js:261:14)
    at Object.<anonymous> (/home/boden/workspace/errors/examples/basic/usage.js:126:13)
    at Module._compile (module.js:456:26)
    at Object.Module._extensions..js (module.js:474:10)
    at Module.load (module.js:356:32)
    at Function.Module._load (module.js:312:12)
    at Function.Module.runMain (module.js:497:10)
    at startup (node.js:119:16)
    at node.js:906:3

Error codes

If you don't provide a code when defining the error, a unique code will be assigned for you. Unique codes start at 600 and increase by 1 for each error defined.

If you prefer to manage your own error codes, for example to group related errors into blocks of codes, just specify a code:

// define code
errors.create({
    name: 'SecurityError',
    code: 1100
});
console.log(new errors.SecurityError().toString());

which logs:

SecurityError: An unexpected SecurityError occurred.
Code: 1100

Inheritance

You can build a hierarchy of errors by specifying the parent your error should inherit from. If no parent is specified, the error will inherit from Error.

For example:

// inheritance
errors.create({
    name: 'FatalError',
    defaultMessage: 'A fatal error occurred',
});
errors.create({
    name: 'FatalSecurityError',
    defaultMessage: 'A security error occurred, the app must exit',
    parent: errors.FatalError
});
try {
    throw new errors.FatalSecurityError();
} catch (e) {
    if (e instanceof errors.FatalError) {
        // exit
        console.log("Application is shutting down...");
    }
}

will produce:

Application is shutting down...

Namespacing

By default, newly defined errors are created on the exports of the errors module, but you can specify where the error should be defined.

For example to define an error on your module's exports:

// namespace
errors.create({
    name: 'MalformedExpressionError',
    scope: exports
});
console.log(new exports.MalformedExpressionError().toString());

Looking up errors

For convenience, errors keeps track of all the errors you've defined via the errors module and allows you to look them up via name or code.

So from our previous example:

errors.find(1100);
errors.find('SecurityError')

Will both return the SecutiryError we defined.

Stack traces

By default stack traces are disabled which means that error methods like toString() and toJSON() return representation without stack traces. You can enable stack traces by leveraging the errors.stacks() method.

For example:

errors.stacks(true);
new errors.Http413Error().toString();
// => includes stack trace
new errors.Http413Error().toJSON();
// => includes a 'stack' property

You can also use the errors.stacks() method without arguments to retrieve the current value of stacks.

This allows you to write code like:

if (errors.stacks()) {
    // => stack traces enabled
}

Mappers

You can register and leverage mapper functions which allow you to map from one (or more) error types into another.

For example if you wanted to mask invalid user and password errors into a generic credentials error:

// mappers
errors.create({name: 'InvalidUsernameError'});
errors.create({name: 'InvalidPasswordError'});
errors.mapper(['InvalidUsernameError', 'InvalidPasswordError'], function(err) {
    return new errors.SecurityError('Invalid credentials supplied');
});
console.log(errors.mapError(new errors.InvalidUsernameError()).toString());
console.log(errors.mapError(new errors.InvalidPasswordError()).toString());

outputs:

SecurityError: Invalid credentials supplied
Code: 1100
SecurityError: Invalid credentials supplied
Code: 1100

Predefined HTTP 4xx-5xx errors

The errors module predefines a set of errors which represent HTTP 4xx-5xx responses. These errors are exported by the errors module and use the naming convention Http[code]Error. For example Http401Error and Http500Error which have a code of 401 and 500 respectively.

For example to leverage the HTTP errors:

throw new errors.Http401Error();
// ...
throw new errors.Http500Error('Something bad happened');

Connect/Express middleware integration

Compatibility Errors version 0.1.0 only works with Express < 4.0.0.

You can use your custom errors with connect's or express's errorHandler() middleware as you might expect:

// ...
app.use(function(req, res, next) {
    // bubble up to errorHandler
    throw new errors.Http401Error();
});
app.use(express.errorHandler());
// ...

However due to the additional information captured in custom errors such as the response and explanation, the default HTML formatting of connect/express errorHandler() is not as pretty as you might like.

Therefore errors provides its own flavor of middleware.

In its simplest form just use errors.errorHandler() as you would do with connect or express. This simple form of the middleware will include the additional datums stored in the custom error such as the explanation and response. But the errors.errorHandler() middleware also accepts some optional arguments to customize its behavior.

Specifically you can set the title to use for HTML based responses, override if the stack should be included and also specify if the middleware should use connectCompat mode. In connectCompat mode the HTML based responses look exactly as they would with connect/express errorHandler() and do not include the additional datums from your error.

For example

// ...
app.use(errors.errorHandler({title: 'Errors Middleware', includeStack: true}));
// ...

binds the errors errorHandler using a custom title and which will include stack traces. Note that using the includeStack property overrides the current value of errors.stacks().

Express send integration

When the errors module is first imported, it determines if express is installed. If express is installed, errors automatically patches express's response.send() method to support send()ing Error based objects.

So the following is valid:

app.get('/users/:user', function(req, res) {
    users.get(req.params.user, function(err, user) {
        return res.send(err || user || new errors.Http404Error('User does not exist'));
    });
});

By default both vanilla errors (those provided by the JS runtime) and errors which have a code which is not a valid HTTP status code are mapped to a 500 response.

So:

res.send(new Error('Vanilla JS error'));

and

res.send(new errors.find('MyErrorName'));

both will result in a 500 response.

Mappers can also be used with express's send() method.

For example:

errors.mapper('RangeError', function(rangeError) {
    return new errors.Http412Error('Invalid range requested');
})
.mapper('ReferenceError', function(refError) {
    return new errors.Http424Error('Bad reference given');
})
.mapper('SyntaxError', function(syntaxError) {
    return new errors.Http400Error('Invalid syntax');
});

// ...

res.send(new RangeError());
// => 412 response as per mapper

res.send(new ReferenceError());
// => 424 response as per mapper

res.send(new SyntaxError());
// => 400 response as per mapper

The implementation provides direct support for application/json, text/html and text/plain content types. If the request specifies a different Accept type, the response defaults to text/plain. Moreover application/json responses provide a complete JSONifed representation of the error.

For example the following setup:

errors.create({
    name: 'DatabaseConnectionError',
    defaultExplanation: 'Unable to connect to configured database.',
    defaultResponse: 'Verify the database is running and reachable.'
});

// ...
res.send(new errors.DatabaseConnectionError());

Will produce the JSON response below when application/json is used as the accept type:

{
    "explanation": "Unable to connect to configured database.",
    "response": "Verify the database is running and reachable.",
    "code": 601,
    "status": 500,
    "name": "DatabaseConnectionError",
    "message": "An unexpected DatabaseConnectionError occurred."
}

For HTML based responses, send()ing an error will produce a HTML response that looks like express's or connect's errorHandler() middleware. That is, it's an HTML page with minimal styling. Moreover you can control the HTML response page title using the errors.title('My Title') method. You can also control if stack traces should be included in the send() by using the errors.stacks() method.

License

(The MIT License)

Copyright (c) 2012 Boden Russell <bodensemail@gmail.com>

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the 'Software'), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED 'AS IS', WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.