minilog

Lightweight client & server-side logging with Stream-API backends

Features

• Works in the browser and on the server
• Themes for Node console output, and for the Chrome console (with fallbacks)
• Interface compatibility with Node and browser consoles, that is, you can replace console.* calls with Minilog since it supports (.log, .debug, .info, .warn and .error)
• Each log message can be associated with a namespace.
• Log output can be filtered by namespace and log level.
• Pipe to one or more backend streams at the same time
• Backends:
  • Node: Console, File (and all other WritableStreams), Redis
  • Browser: Console, LocalStorage, jQuery.ajax
• In Chrome, we support theming the dev console output.

See the docs at http://mixu.net/minilog/.

Upgrading from 2.x to 3.x

v3.0 changes the ajax logger's POST payload from newline-separated JSON lines to a hash { logs: [ ... ] }. Previously, this format was only used if logger.extras was set. This makes the POST payload parseable as JSON directly, but may require you to update your logging endpoint if you were using v2.x before.

Pipes everywhere

minilog is more convention than code. The logger is a EventEmitter, and backends are Writable streams. Filters and formatters are duplex (readable + writable) streams.

minilog works in Node, and in the browser:

// logs are scoped to a namespace for easy filtering (here, the namespace is "app")
var log = require('minilog')('app');
require('minilog').enable();

in the browser (via a single exported global window.Minilog):

<script src="dist/minilog.js"></script>
<script>
var log = Minilog('app');
Minilog.enable();
</script>

Usage:

// assuming you've done the two things above
log
  .debug('debug message')
  .info('info message')
  .log('info message')
  .warn('warning')
  .error('this is an error message');

Output:

To log to the console:

require('minilog').enable();
// or .pipe(process.stdout), if you don't want the default formatting and filtering

To log into a file:

require('minilog').pipe(fs.createWriteStream('./temp.log'));

You can also log to Redis and over HTTP to a RESTful API, see the backends at the end of this page.

You can pipe to more than one endpoint if you want.

Installation

For Node:

$ npm install minilog

You can find a ready-made file for the web in ./dist/minilog.js.

Upgrading from minilog v1

Everything is now a pipe, which means that the .format() and .filter() functions are deprecated. Check out the new filter mechanism docs. To apply a formatter, you should pipe the input into the formatter, and then pipe it to the desired backend:

var Minilog = require('minilog');

Minilog.pipe(Minilog.backends.console.formatWithStack)
       .pipe(Minilog.backends.console);

Enabling logging

Minilog output is suppressed by default. To enable logging, append minilog=1 to the page URL:

http://www.example.com/index.html?minilog=1

or call Minilog.enable() from the dev console or in code. On the browser, this also sets a value in LocalStorage so that logging is enabled on subsequent reloads. Call Minilog.disable() (new in v2) to stop logging.

Filtering

Minilog supports filtering via the log scope name and the log level, as well as a number of nifty features. See the filtering docs for more.

Formatting & themes

Minilog supports themes and custom formatters, and comes several with built-in themes:

To enable a specific theme, pipe to the formatter and then to the console:

var Minilog = require('minilog');

Minilog
    // formatter
    .pipe(Minilog.backends.console.formatClean)
    // backend
    .pipe(Minilog.backends.console);

Have a look at ./test/examples/themes_example.js.

To write your own formatter, have a look at the source code for the formatters - they inherit from Minilog.Transform.

Using Minilog as a console replacement

If you use an injected console object to log browser or Node.js activity, you can use Minilog instead: they have similar interfaces. Minilog provides a log() method, which proxies to debug().

So for instance, the following snippet:

function doThings(console) {
    if (problem) {
        console.error('problem');
        return;
    }
    console.log('no problem');
}

Works seamlessly with Minilog instead of console:

var Minilog = require('minilog');
doThings(Minilog);

Backends

Backends are Writable streams which handle stringification.

Node: Console, Redis

The console backend is literally this (plus code for pretty printing log lines in various ways):

{
  write: function(str) { process.stdout.write(str); }
}

The Redis backend is almost equally simple - it accepts client (an instance of node-redis) and key and uses rpush() to add to the list at the specified key.

Browser: Array, Console, jQuery, localStorage

The Array backend stores the log lines into an array. This is useful if you want to keep a list of all the log lines, e.g. for error reporting. Call .get() to get the array, and .clear() to empty it.

The Console backend makes sure that console.log is available. On IE8 and IE9, it tries to make the console a bit less aweful by using JSON.stringify to convert objects into strings (rather than "[Object object]").

The jQuery backend is useful for logging client-side log lines on the server side:

• it sends new log messages as a POST request to a given URL every 30 seconds
• if localStorage is available, logs are written to localStorage as well. This is helpful because it reduces the risk that you lose log lines just because the client navigates to a different page.
• Unsent logs from localStorage are sent the next time the backend is activated (on your domain, localStorage is isolated).
• No errors, even if localStorage is not available or jQuery is not defined (though no POST requests if no jQuery).

The localStorage backend just writes logs to the given key in localstorage.

Have a look at the example server setup in ./test/examples/jquery_server.js.