Skip to content

HTTPS clone URL

Subversion checkout URL

You can clone with
or
.
Download ZIP
Integration layer between TestSwarm and BrowserStack
JavaScript Shell
Branch: master

cli: Use return instead of callback for --ua2bs flag

This method is actually a plain getter that returns the value,
it doesn't take a callback. Because of this, 'twbs --ua2bs <name>'
resulted in no output.
latest commit 1b22aca35a
@Krinkle Krinkle authored
Failed to load latest commit information.
src cli: Use return instead of callback for --ua2bs flag
.gitignore
.jscsrc
.jshintignore
.jshintrc
.travis.yml build: Enable jscs and jshint
LICENSE Add myself to the credits
README.md readme: Update build status badge
package.json build: Run jscs from npm-test
sample-config.json Rename sample files for consistency.
sample-logrotate.conf
sample-run.sh Recommend run-loop 60 instead of 50

README.md

Build Status

testswarm-browserstack

This is a light weight integration layer between TestSwarm and BrowserStack. Use it to spawn BrowserStack workers needed by TestSwarm on demand.

This script is currently compatible with:

Install


git clone https://github.com/clarkbox/testswarm-browserstack.git testswarm-browserstack
cd testswarm-browserstack/
npm install

How to use CLI:


node src/cli.js --run

This above command will create and terminate BrowserStack workers as needed according to the information the TestSwarm swarmstate API provides. This command should be executed on a regular interval, either via a scheduler (such as crontab) or by letting node do a continuous loop (using the --run-loop option). Be sure to do start it from a scheduler still in a way that it will only start it if it isn't running anymore (in case of an exception).

If you plan to run it from a scheduler and keep log files, you're recommended to use the run-sample.sh file as a start. It contains the basic cli invocation as a template. Fill in the argument values and adjust the script and log paths. Also, as a reminder that log files can run out of hand quickly, we've provided a sample file to use in logrotate (e.g. on Ubuntu). To learn about logrotate, checkout Ubuntu manpages or the Slicehost tutorial on the subject. To install it, copy the file to logrotate.conf within this directory, adjust the path and (if you want to) set different settings. Then move it to /etc/logrotate.d/testswarm-browserstack.conf.

Main scripts:


  1. testswarm-browserstack.js - Abstraction of TestSwarm API, and Scott González's BrowserStack API. Use it to automatically spawn BrowserStack workers based on your swarm's needs.
  2. cli.js - nodejs cli wrapper around it all. Allows for scripted or generally easy manual invocation of the script.

testswarm-browserstack.js


Options documentation:

  • browserstack.user: BrowserStack username
  • browserstack.pass: BrowserStack password
  • browserstack.workerTimeout: Maximum lifetime of the worker (in seconds). Use 0 for indefinitely (BrowserStack will terminate the worker after the maximum run time, as of writing that maximum is 30 minutes).
  • browserstack.dryRun: Enable to only simulate spawning and termination of browserstack workers.
  • browserstack.totalLimit: Maximum number of simultaneous workers allowed under this BrowserStack account.
  • browserstack.eqLimit: How many simultaneous workers with the same browser are allowed.
  • testswarm.root: URL to the root of the TestSwarm installation. Relative to here should be ./index.php and ./api.php.
  • testswarm.runUrl: URL to the TestSwarm run page (including client name), for BrowserStack workers to open. If your swarm is protected with a token, this is the place to enter the token.
  • verbose: Output debug information (through console.log).

Example config.json:

{
  "browserstack": {
    "user": "example",
    "pass": "*******"
  },
  "testswarm": {
    "root": "http://ci.example.org/testswarm",
    "runUrl": "http://c.example.org/testswarm/run/JohnDoe"
  }
}

cli.js


This is a nodejs cli wrapper around testswarm-browserstack.js. Use --help to get all the information you need to know (see above for example usage):

  Usage: cli.js [options]

  Options:

    -h, --help            output usage information
    -V, --version         output the version number
    --config [path]       path to config file with options (defaults to ./config.json)
    --run                 Retrieve TestSwarm state and spawn/terminate BrowserStack workers as needed
    --run-loop <timeout>  Execute --run in a non-overlapping loop with set timeout (in seconds) between iterations
    --worker <id>         Get info abuot a specific BrowserStack worker
    --spawn <uaId>        Spwawn a BrowserStack worker by swarm useragent id (joining the swarm)
    --terminate <id>      Terminate a specific BrowserStack worker
    --terminateAll        Terminate all BrowserStack workers
    -v, --verbose         Output debug information (through console.log)
    --dry, --dry-run      Simulate spawning and termination of browserstack workers
Something went wrong with that request. Please try again.