cucumber.md

Cucumberjs parser

There are 2 parsers that can be used, the cucumber- and the cucumberMulti-parser. Just select one of them based on the below description and use them like this

// how to use the `cucumber`-parser
protractor-flake --protractor-path=/path/to/protractor --parser cucumber --node-bin node --max-attempts=3 -- path/to/protractor.conf.js

// how to use the `cucumberMulti`-parser
protractor-flake --protractor-path=/path/to/protractor --parser cucumberMulti --node-bin node --max-attempts=3 -- path/to/protractor.conf.js

cucumber

This parser only works with cucumberjs version < v0.9.0

The cucumber parser will search the testoutput for this piece of text /path/to/your/featurefile/flakey.feature:4 # Scenario: Flakey scenario. This is the line that says that a scenario in a featurefile failed. The /path/to/your/featurefile/flakey.feature piece will be stripped out and passed to protractor-flake to rerun the file(s) for the amount of retries that have been given.

cucumberMulti

This parser is based on cucumberjs > v0.9.0. These versions have a different logging and also a different way in returning the path of the runned specs. (See example logging).

This parser can parse testoutput from capabilities and multiCapabilities

The cucumberMulti parser will search the testoutput for the text combination Failures:and Specs: /path/to/your/featurefile/flakey.feature. The parser needs to search for both pieces of text because succeeded specs are also being logged. The /path/to/your/featurefile/flakey.feature piece will be stripped out and passed to protractor-flake to rerun the file(s) for the amount of retries that have been given.

Based on the way the capability in Protractor has been configured the Specs: are printed. See the table below for the results.

// Single
capabilities: {
    browserName: 'chrome',
    shardTestFiles: true
}
// Multi
multiCapabilities = [
    {
        browserName: 'chrome',
        shardTestFiles: true
    },
    {
        browserName: 'firefox',
        shardTestFiles: true
    }
]

Capability	shardTestFiles	Feature	Failures	Specs printed
single	true	1	true	false
single	false	1	true	false
single	true	1	false	false
single	false	1	false	false
single	true	>1	true	TRUE
single	false	>1	true	false
single	true	>1	false	TRUE
single	false	>1	false	false
multi	true	1	true	TRUE
multi	false	1	true	TRUE
multi	true	1	false	TRUE
multi	false	1	false	TRUE
multi	true	>1	true	TRUE
multi	false	>1	true	FALSE
multi	true	>1	false	TRUE
multi	false	>1	false	false

When no Specs: are printed in the logging Protractor will run the Protractor command again which means that all the specs are run on ALL the instances (also the succeeded specs).

Keep in mind

Keep in mind that if the testsuite holds multiple featurefiles and only 1 featurefile is flakey the following will occur. For example:

• Run 1:
  • 2 featurefiles are run
  • 1 is flakey => protractor-flake will parse the output, finds 1 failed specs output (see table) and reruns protractor
• Run 2:
  • 1 featurefile is run
  • featurefile is flakey => protractor-flake will parse the output and finds NO specs in the output (see table) and reruns protractor
• Run 3:
  • protractor is rerun by protractor-flake but gets no --specs, this means that ALL the featurefiles are run again

To avoid this behaviour a simple "workaround" can be implemented, see "Always print specs"

Always print specs

There is a way to always print Specs:. This can be done with the following hook, see cucumberjs for more hook info:

var afterHook = function () {
  this.After(function (scenario, callback) {
        if (scenario.isFailed()) {
            console.log('Specs:', scenario.getUri());
        }
        callback();
  });
};

module.exports = afterHook;

Make sure that the file is required in the protractor.conf.js file like this

cucumberOpts: {
    require: [
        'path/to/your/afterHooks.js'
    ]
}

This will always print the Specs: in the testoutput. The parser will filter double specs.