How to tell when protractor/jasmine is done? (Reporting Tool logic)

Question

So I had to essentially rewrite a reporting tool I was using for HTML reports for Protractor/Jasmine E2E tests. Right now I'm able to write nicely to a file all test results for one test file being ran (e.g test1.spec.js).

My issue comes from when I run more than one file at a time (e.g. ['./test1.spec.js', './test2.spec.js']). Ideally what I would want to do is to store all the information from both tests into a, let's say, fileData object/variable and then write that variable to a file once ALL test files have completed.

I've tried writing in the jasmineDone function, but that occurs after each test file completes so it either duplicates information I have from the report or overwrites/messes up the styling of the report.

this.jasmineDone = function () {
    fileData += styles.closeReportTags;
    fileData += styles.closeBodyHtmlTag;
    fs.appendFileSync("C:/Reports/testing1.html", fileData);
}

I have also created an afterLaunch method defined in my protractor config file, and used in the reporting tool file as such:

this.afterLaunch = function (callback) {
    fs.writeFile("C:/Reports/testing1.html", fileData);
}

In the above I tried to append all the test results into the fileData variable throughout each run and then just write it all back after it's completed, but the fileData doesn't seem to get passed properly to the afterLaunch method (it doesnt write anything to the file, just creates a blank html file).

So my question is, In what ways can I keep a running count, or flag, to be able to determine if there are more test files queued up for protractor/jasmine in order to figure out when the actual test run is complete and in which scope can I use that to then write my file.

Thanks.

Answer 1

I had the same problem. Ultimately, all the hooks available in Jasmine/Protractor weren't good enough (especially if you're sharding tests). What ended up working for me was creating my own cli, parsing/storing the results I wanted, and then doing reporting.

Here's an example cli:

#!/usr/bin/env node

var spawn = require('child_process').spawn;
var output = [];

// get protractor args from cli, and skip first two (node and self)
var args = process.argv.slice(2);
// run protractor
protractor = spawn('node_modules/.bin/protractor', args);

/**
 * process listeners 
 */
protractor.stdout.on('data', function(buffer) {
    var text = buffer.toString();
    // parse results here...
    console.log(text);
});

protractor.stderr.on('data', function(buffer) {
    var text = buffer.toString();
    console.log(text);
});

protractor.on('exit', function(status) {
    // protractor is _done_ done, report your results here...
});

Make sure to chmod the file to be executable (eg. chmod 755). Then run it thusly:

./myCI conf.js --baseUrl http://example.com