buster-server-cli

Command-line interface API for running ramp instances with a simple interface that allows capturing and viewing a list of connected browsers.

In Buster, this module is the implementation of the buster server command. It does not define the binary however, as it is intended to be generic enough to be reused outside of Buster.

Possible use cases

The capture server is the central piece in Buster's multiple browser automation capabilities. This module can be used as is to run tests for any framework, as it does not know anything about tests at all. However, if you're shipping a capture server for your own framework, you may want to brand your server a little.

The following example shows how to create a custom capture server for the fictional checkit test framework.

The binary

// checkit/bin/checkit-server
var path = require("path");
var serverCli = require("buster-server-cli");

serverCli.create(process.stdout, process.stderr, {
    missionStatement: "Checkit crazy multi-browser test runner server",
    description: "checkit-server [options]",
    templateRoot: path.join(__dirname, "..", "views"),
    documentRoot: path.join(__dirname, "..", "public")
}).run(process.argv.slice(2));

The index template

You need to define two templates for the server to work correctly. The first one is index.ejs, which is an ejs template for the index page of the server. Buster's index template renders a list of captured browsers and a link to /capture, which is the URL that causes the browser to become a captured slave.

The index.ejs template is rendered with one piece of data — slaves — which is an array of slave objects:

• slave.browser A string, i.e. "Firefox"
• slave.platform A string, i.e. "Linux"
• slave.version A string, i.e. "12.0"
• slave.os A string, contains a richer OS/platform description
• slave.userAgent The original user agent

The header template

The second template is the header.ejs template. It is used in the top frame in the frameset that is displayed in captured slaves. Currently this is just a static template, but future versions will expose an API to communicate with the server here to display progress etc.

See Buster's header template for a reference implementation.

Changelog

0.4.0 (2015-Nov-28)

• Upgraded all dependencies
• Replaced paperboy with serve-static

0.3.4 (2015-Mar-16)

• Additional param sessionId for method captureHeadlessBrowser.

0.3.3 (2015-Mar-12)

• Waits now until headless browser is started before calling callback of run. Needed for buster-ci.

0.3.2 (2015-Jan-23)

• Fix for issue #432 - buster-server says "running" even if address is already in use
• PR agent.name could somehow be null and will cause buster-server-cli to exit

0.3.1 (2014-Sep-17)

• Fix for issue #416 - buster-server crash with IE 11 on W7 only if there is two browsers captured