Mocha reporter to generate coverage report of istanbul instrumented code, for grunt
This doesn't force you to use PhantomJS, or instrument code for server or client-side.
- Install needed dependencies using:
npm install grunt mocha istanbul --save-dev
- Install this package using:
npm install grunt-mocha-istanbul --save-dev
- Call inside
Gruntfile.js
:grunt.loadNpmTasks('grunt-mocha-istanbul')
Although it's intended for use with Istanbul, you can freely
use this task with other command line compatible tool, by changing the scriptPath
option, such as:
grunt.initConfig({
mocha_istanbul: {
target: {
options: {
scriptPath: require.resolve('coverage-tool/the/path/to/bin'), // usually in nameofcoveragelibrary/lib/cli
}
}
}
});
Before anything, install babel required stuff.
npm install babel-cli babel-presets-2015 babel-register
Define your .babelrc file on the same level as Gruntfile.js
{
"presets": ["es2015"]
}
Then you'll need to use Isparta, until Istanbul 1.0 is released. (or you can use 1.0.0-alpha.2)
{
mocha_istanbul: {
src: 'test',
options: {
scriptPath: require.resolve('isparta/lib/cli'),
nodeExec: require.resolve('.bin/babel-node') // for Windows, you MUST use .bin/babel-node.cmd instead
mochaOptions: ['--compilers', 'js:babel-register'], // if you are writing your tests with ES2015+ as well
}
}
}
NOTE: for some unknown reason, using babel-node as nodeExec, with print
option makes it fail.
Most of the options that you pass to mocha is available in options
:
module.exports = function(grunt){
grunt.initConfig({
mocha_istanbul: {
coverage: {
src: 'test', // a folder works nicely
options: {
mask: '*.spec.js'
}
},
coverageSpecial: {
src: ['testSpecial/*/*.js', 'testUnique/*/*.js'], // specifying file patterns works as well
options: {
coverageFolder: 'coverageSpecial',
mask: '*.spec.js',
mochaOptions: ['--harmony','--async-only'], // any extra options
istanbulOptions: ['--harmony','--handle-sigint']
}
},
coveralls: {
src: ['test', 'testSpecial', 'testUnique'], // multiple folders also works
options: {
coverage:true, // this will make the grunt.event.on('coverage') event listener to be triggered
check: {
lines: 75,
statements: 75
},
root: './lib', // define where the cover task should consider the root of libraries that are covered by tests
reportFormats: ['cobertura','lcovonly']
}
}
},
istanbul_check_coverage: {
default: {
options: {
coverageFolder: 'coverage*', // will check both coverage folders and merge the coverage results
check: {
lines: 80,
statements: 80
}
}
}
}
});
grunt.event.on('coverage', function(lcovFileContents, done){
// Check below on the section "The coverage event"
done();
});
grunt.loadNpmTasks('grunt-mocha-istanbul');
grunt.registerTask('coveralls', ['mocha_istanbul:coveralls']);
grunt.registerTask('coverage', ['mocha_istanbul:coverage']);
};
If there's a mocha.opts
file inside the first src
folder or file defined, it will warn if you are overwriting any options.
Coverage is written to coverage
folder by default, in the same level as the Gruntfile.js
The check
will fail the build if the thresholds are not met. It's a great possibility for CI-builds.
Mochas parameters, check https://mochajs.org/#usage
Type: Array
Default Value: []
Type: Boolean
Default Value: false
Type: Array
Default Value: []
Type: String
Default Value: false
Type: Number
Default Value: false
Type: Boolean
Default Value: false
Type: String
Default Value: false
Type: Boolean
Default Value: false
Type: Boolean
Default Value: false
Type: Array
Default Value: false
An array of strings, any additional node executable parameters, manually set.
Eg.: nodeOptions: ['--throw-deprecation', '--require', 'some/module']
Type: String
Default Value: false
The mask for the tests to be ran. By default, mocha will execute the test
folder and all test files.
Will override any files specified in src
and instead use the mask on those files' folders.
Type: Array
Default Value: false
An array of strings, any additional mocha parameters, manually set.
Eg.: mochaOptions: ['--harmony', '-s', '100']
Type: Array
Default Value: false
Setting this exclude files from coverage report, check istanbul help cover
. You may use glob matching in here.
Type: Array
Default Value: false
Setting this includes only those files in the coverage report, check istanbul help cover
. You may use glob matching in here.
Type: Array
Default Value: false
An array of strings, any additional istanbul parameters, manually set.
Eg.: istanbulOptions: ['--harmony', '--handle-sigint', 'some=value', '-s', 'value']
Type: String
Default Value: 'coverage'
Name of the output of the coverage folder
Type: Array
Default Value: ['lcov']
Name of report formats. You can specify more than one. If you intend to use the coverage
option to
true
or do any checks, you must add: ['yourformat','lcovonly']
, since it's needed for the lcov.info
file to be created.
html - produces a bunch of HTML files with annotated source code
lcovonly - produces an lcov.info file
lcov - produces html + lcov files. This is the default format
cobertura - produces a cobertura-coverage.xml file for easy Hudson integration
text-summary - produces a compact text summary of coverage, typically to console
text - produces a detailed text table with coverage for all files
teamcity - produces service messages to report code coverage to TeamCity
Type: String
Default Value: false
The root path to look for files to instrument, defaults to .
. Can help to exclude directories that are not
part of the code whose coverage should be checked.
Type: String
Default Value: false
The type of report to print to console. Can be one of 'summary', 'detail', 'both', or 'none'. By
Default, Istanbul will print the 'summary' report.
Type: String
Default Value: 'istanbulPath'
Allows to override the default istanbul path to use another coverage library, such as ibrik.
Need to set the full path to the bin (script that accepts stdin arguments) and is compatible with cover
.
Type: Boolean
Default Value: false
Setting this to true
makes the task emit a grunt event coverage
, that will contain the lcov data from
the file, containing the following callback function(lcovcontent, done)
, and you must manually call
done()
when you are finished, else the grunt task will HANG, and won't allow any other tasks to finish.
See more information below
Type: Boolean
Default Value: false
Spits out the command line that would be called, just to make sure everything is alright
Type: String
Default Value: process.execPath
Sets the node executable that will invoke Istanbul and Mocha. Useful for setting something else than node, like babel-node
Type: String
Default Value: process.cwd()
Sets the current working directly. Note that changing this might have unexpected results, since the plugin and Grunt expects
to be working on the same level of Gruntfile.js
Type: Boolean
Default Value: false
Suppresses the output from Mocha and Istanbul
Type: Number
Default Value: false
Number of statements threshold to consider the coverage valid
Type: Number
Default Value: false
Number of lines threshold to consider the coverage valid
Type: Number
Default Value: false
Number of branches threshold to consider the coverage valid
Type: Number
Default Value: false
Number of functions threshold to consider the coverage valid
When you set the option coverage
to true
, you'll receive the coverage/lcov.info
file contents:
grunt.event.on('coverage', function(lcov, done){
console.log(lcov);
done(); // or done(false); in case of error
});
This is mainly useful so you can send it to, for example, coveralls (using coveralls):
grunt.event.on('coverage', function(lcov, done){
require('coveralls').handleInput(lcov, function(err){
if (err) {
return done(err);
}
done();
});
});
This way, Travis-CI can send the Istanbul generated LCOV directly to Coveralls.io website in this example, but you could create any transform for Jenkins, TeamCity, Hudson, etc.
MIT