NPM Failsafe

The npm-failsafe lets you execute a sequence of NPM scripts and return the correct exit code should any of them fail.

Installation

npm install --save-dev npm-failsafe

Usage

Failsafe is a command line tool. To view the available options, run the following command in your terminal:

npx failsafe --help

Configuration

You can use failsafe to run scripts defined in your package.json file.

Configuring Failsafe in package.json

If your scripts require any arguments, those can be specified upfront in your package.json file.

For example, given the below package.json file:

{
    "scripts": {
        "test": "jest",
        "test:coverage": "npm run test -- --coverage",
        "lint": "eslint 'src/*'",
        "start": "pm2 start server.js --port=3000",
        "ci": "failsafe start lint test:coverage"
    }
}

Running: npm run ci will execute:

• npm run start
• npm run lint
• npm run test:coverage, in that order

It will also return the highest exit code of all the executed scripts.

Configuring Failsafe at runtime

If you need to pass arguments to your scripts at runtime, pass them to failsafe directly, and then Failsafe will pass them to your script as per the configuration in your package.json file.

For example, given the below package.json file:

{
    "scripts": {
        "test:clean": "rimraf reports",
        "test:execute": "cucumber-js",
        "test:report": "serenity-bdd run",
        "test": "failsafe test:clean test:execute [--name,--tags] test:report"
    }
}

Running npm test -- --name="Authentication" will execute:

• npm run test:clean
• npm run test:execute -- --name="Authentication", which in turn will execute cucumber-js --name="Authentication"
• npm run test:report

The same applies to npm test -- --tags="@smoke", which will execute:

• npm run test:clean
• npm run test:execute -- --tags="@smoke", which in turn will execute cucumber-js --name="@smoke"
• npm run test:report

Using wildcards

To help you avoid configuration errors, Failsafe will complain if you try to execute a script that doesn't exist, or use an argument that is not configured for a given script.

However, you can configure Failsafe with a wildcard of [...]. This instructs Failsafe to pass any arguments it receives to the script configured with the wildcard.

For example, given the below package.json file:

{
    "scripts": {
        "test:clean": "rimraf reports",
        "test:execute": "cucumber-js",
        "test:report": "serenity-bdd run",
        "test": "failsafe test:clean test:execute [...] test:report"
    }
}

Running npm test -- --name="Authentication" or npm test -- --tags="@smoke" will instruct Failsafe to pass those arguments to the test:execute script - the receiver of the wildcard.

Motivation

Assume a package.json with the following scripts defined:

"scripts": {
    "preintegration": "bin/start_the_server.sh",
    "integration": "bin/run_some_tests_that_require_the_server.sh",
    "cleanup": "bin/shutdown_the_server.sh"
}

In this example, we want to execute the integration script. The script runs some integration tests against some server, which means that we need to start the server up before the tests and shut it down afterwards.

The server itself is started in the preintegration phase (check out the node docs to learn more about the pre- and post- commands).

The question is: how do we shut it down?

We could add the following test script to our package.json: "test": "integration && cleanup".

The problem with this is that because of how the && operator works, the cleanup script will only get executed when the integration script succeeds. This is no good because we need to shut down the server even if the integration tests fail.

We could try to use the || operator instead, which executes the second script no matter the result of the first one: "test": "integration || cleanup". However, the problem with this approach is that the exit code of the "integration || cleanup" combo will always take the value of 0, incorrectly indicating that the test script has succeeded. This could for example cause a continuous integration server to publish your project even if the tests have failed...

Enter npm-failsafe!

With npm-failsafe you can execute a sequence of arbitrary NPM scripts and return the correct exit code should any of them fail:

"scripts": {
     "preintegration": "bin/start_the_server.sh",
     "integration": "bin/run_some_tests_that_require_the_server.sh",
     "cleanup": "bin/shutdown_the_server.sh",

     "test": "failsafe integration cleanup"
}

Your feedback matters!

Did you find this project useful? Give it a star on github! ★

Found a bug? Raise an issue or submit a pull request.

Have feedback? Let me know on twitter: @JanMolak