BCOMM: Browser Command

Dispatch commands, tests, logs, etc to and from remote browsers.

Usage

The following are example usages. See the help for each command for more details.

(shell commands are preceded by "$" and bcomm console commands are preceded by ">")

Startup the server:

$ bcomm server

After the server is started, point multiple web browsers at the server (e.x http://localhost:8080/ if on local machine with default port) to start them as runners.

Set the client tool's (the "bcomm" command) server url (if different than localhost:8080):

$ bcomm config server.url http://12.34.56.78:1234/

List all connected runners (browser) along with their browser user agent and current location:

$ bcomm runners
6ba1da: Safari 5 (Mac) => http://12.34.56.78:1234/static/dispatcher.html
bbf29f: Chrome 6 (Mac) => http://12.34.56.78:1234/static/dispatcher.html
72fcfd: Opera 9.8 (Mac) => http://12.34.56.78:1234/static/dispatcher.html
238d84: Firefox 3.6 (Mac) => http://12.34.56.78:1234/static/dispatcher.html

Each runner (or uploaded test or task) is identified by a 6 digit hex string.

Open a simple JavaScript console to the connected runners and issue commands:

$ bcomm console
> 1+1
238d84: 2
bbf29f: 2
72fcfd: 2
6ba1da: 2
> console.log("hello!")
6ba1da: console.log("hello!")
6ba1da: undefined
bbf29f: console.log("hello!")
bbf29f: undefined
72fcfd: console.log("hello!")
72fcfd: undefined
238d84: console.log("hello!")
238d84: undefined
>

("undefined" is printed here because console.log() has no return value)

Upload a test (can be a HTML or JavaScript file, or a directory containing an "index.html" file):

$ bcomm upload example/standalone.js
cf1091

Run the last uploaded test:

$ bcomm run
Running tests: cf1091
238d84: console.log("hello!")
238d84: console.log("world!")
238d84: {"type": "complete", "result": "test result goes here", "complete": true, "taskID": "8c0f28", "runnerID": "238d84"}
6ba1da: console.log("hello!")
6ba1da: console.log("world!")
6ba1da: {"type": "complete", "result": "test result goes here", "complete": true, "taskID": "4fe957", "runnerID": "6ba1da"}
bbf29f: console.log("hello!")
bbf29f: console.log("world!")
bbf29f: {"type": "complete", "result": "test result goes here", "complete": true, "taskID": "65702f", "runnerID": "bbf29f"}
72fcfd: console.log("hello!")
72fcfd: console.log("world!")
72fcfd: {"type": "complete", "result": "test result goes here", "complete": true, "taskID": "a34875", "runnerID": "72fcfd"}

BCOMM must be notified when the test is done by calling BCOMM.complete(result) otherwise the runner will be blocked.

Other commands include:

• reset: refreshes runner(s)
• abort: aborts the runner(s) current task (useful if it finished without calling BCOMM.complete())
• tail: lists all console output, connects, disconnects, etc

TODO

• Better output (print browser/version/os instead of runnerID)
• More robust (timeout tests, etc)
• Cleanup old tests
• Auto-run uploads
• Port server to non-blocking platform