This framework was originally part of angular/protractor and is now a separate module to decouple cucumber.js.
npm install --save-dev protractor-cucumber-framework
To implement this framework, utilize the protractor
custom framework config option:
exports.config = {
// set to "custom" instead of cucumber.
framework: 'custom',
// path relative to the current config file
frameworkPath: require.resolve('protractor-cucumber-framework'),
// require feature files
specs: [
'path/to/feature/files/**/*.feature' // accepts a glob
],
cucumberOpts: {
// require step definitions
require: [
'path/to/step/definitions/**/*.steps.js' // accepts a glob
]
}
};
All of the cucumberOpts
will be passed to cucumberjs
as arguments.
For example, to call cucumberjs with the --strict
, --no-colors
, and to specify custom formatters:
cucumberOpts: {
strict: true,
'no-colors': true,
format: ['progress', 'pretty:output.txt'],
// ...
}
The following parameters have special behavior:
require
- globs will be expanded to multiple--require
argumentsrerun
- value is passed as an argument; for use with the rerun feature
If you have a formatter that outputs to a path and your tests are sharded or you have multi capabilities then this library will add the PID to the path to make them unique. The reason for this is multiple processes can write to the same path which ends up clobbering each other. You'll end up with 1 file per process that protractor spawns.
exports.config = {
capabilities: {
shardTestFiles: true,
// ...
},
cucumberOpts: {
format: 'json:results.json',
// ...
}
};
If there were 2 feature files then you can expect the following output files...
results.11111.json
results.22222.json
...where the numbers will be the actual PIDs.
If your process abruptly stops with an exit code 199
then your steps most likely threw an uncaught
exception. Protractor is capturing these and exiting the process in this situation. The solution is
to upgrade to at least protractor version 4.0.10 and add the following to your protractor conf...
ignoreUncaughtExceptions: true
This allows cucumber to handle the exception and record it appropriately.
Pull requests are welcome. Commits should have an appropriate message and be squashed.
Ensure that the following dependencies are installed:
- Java SDK and JRE
- node.js
- Google Chrome
Clone the github repository:
git clone https://github.com/protractor-cucumber-framework/protractor-cucumber-framework
cd protractor-cucumber-framework
npm install
Start a selenium server:
npm run webdriver
Start the test app that tests will be run against in a separate shell:
npm start
Run the tests in a separate shell:
npm test
- bump version
npm publish
- tag release (
git tag vx.x.x && git push origin master --tags
) - build github release (
npx release
)