Grunt plugin for Karma
This current version uses karma@^3.0.0
. For using older versions see the
old releases of grunt-karma.
From the same directory as your project's Gruntfile and package.json, install karma and grunt-karma with the following commands:
$ npm install karma --save-dev
$ npm install grunt-karma --save-dev
Once that's done, add this line to your project's Gruntfile:
grunt.loadNpmTasks('grunt-karma');
Inside your Gruntfile.js
file, add a section named karma
, containing
any number of configurations for running karma. You can either put your
config in a [karma config file] or leave it all in your Gruntfile (recommended).
karma: {
unit: {
configFile: 'karma.conf.js'
}
}
karma: {
unit: {
options: {
files: ['test/**/*.js']
}
}
}
You can override any of the config file's settings by putting them directly in the Gruntfile:
karma: {
unit: {
configFile: 'karma.conf.js',
port: 9999,
singleRun: true,
browsers: ['PhantomJS'],
logLevel: 'ERROR'
}
}
To change the logLevel
in the grunt config file instead of the karma config, use one of the following strings:
OFF
, ERROR
, WARN
, INFO
, DEBUG
The files
option can be extended "per-target" in the typical way
Grunt handles files:
karma: {
options: {
files: ['lib/**/*.js']
},
unit: {
files: [
{ src: ['test/**/*.js'] }
]
}
}
When using the "Grunt way" of specifying files, you can also extend the file objects with the options supported by karma:
karma: {
unit: {
files: [
{ src: ['test/**/*.js'], served: true },
{ src: ['lib/**/*.js'], served: true, included: false }
]
}
}
When using template strings in the files
option, the results will flattened. Therefore, if you include a variable that includes an array, the array will be flattened before being passed to Karma.
Example:
meta: {
jsFiles: ['jquery.js','angular.js']
},
karma: {
options: {
files: ['<%= meta.jsFiles %>','angular-mocks.js','**/*-spec.js']
}
}
If you have multiple targets, it may be helpful to share common
configuration settings between them. Grunt-karma supports this by
using the options
property:
karma: {
options: {
configFile: 'karma.conf.js',
port: 9999,
browsers: ['Chrome', 'Firefox']
},
continuous: {
singleRun: true,
browsers: ['PhantomJS']
},
dev: {
reporters: 'dots'
}
}
In this example the continuous
and dev
targets will both use
the configFile
and port
specified in the options
. But
the continuous
target will override the browser setting to use
PhantomJS, and also run as a singleRun. The dev
target will simply
change the reporter to dots.
There are three ways to run your tests with karma:
Setting the autoWatch
option to true will instruct karma to start
a server and watch for changes to files, running tests automatically:
karma: {
unit: {
configFile: 'karma.conf.js',
autoWatch: true
}
}
Now run $ grunt karma
Many Grunt projects watch several types of files using grunt-contrib-watch.
Config karma like usual (without the autoWatch option), and add
background:true
:
karma: {
unit: {
configFile: 'karma.conf.js',
background: true,
singleRun: false
}
}
The background
option will tell grunt to run karma in a child process
so it doesn't block subsequent grunt tasks.
The singleRun: false
option will tell grunt to keep the karma server up
after a test run.
Config your watch
task to run the karma task with the :run
flag. For example:
watch: {
//run unit tests with karma (server needs to be already running)
karma: {
files: ['app/js/**/*.js', 'test/browser/**/*.js'],
tasks: ['karma:unit:run'] //NOTE the :run flag
}
},
In your terminal window run $ grunt karma:unit:start watch
, which starts the
karma server and the watch task. Now when grunt watch detects a change to
one of your watched files, it will run the tests specified in the unit
target using the already running karma server. This is the preferred method
for development.
Keeping a browser window & karma server running during development is
productive, but not a good solution for build processes. For that reason karma
provides a "continuous integration" mode, which will launch the specified
browser(s), run the tests, and close the browser(s). It also supports running
tests in PhantomJS, a headless webkit browser which is great for running tests as part of a build. To run tests in continous integration mode just add the singleRun
option:
karma: {
unit: {
configFile: 'config/karma.conf.js',
},
//continuous integration mode: run tests once in PhantomJS browser.
continuous: {
configFile: 'config/karma.conf.js',
singleRun: true,
browsers: ['PhantomJS']
},
}
The build would then run grunt karma:continuous
to start PhantomJS,
run tests, and close PhantomJS.
You can pass arbitrary client.args
through the commandline like this:
$ grunt karma:dev watch --grep=mypattern
MIT License