Unit tests, written with QUnit, are used to
expose bugs for squashing, prevent bugs from respawning, and suppress new
bugs when adding new features and making changes.
The simplest way to run the tests is to open tests/tests.html in your browser.
The test suites will automatically run themselves and present their results.
To run the tests from the command line, download and install
PhantomJS, and run run-qunit.js with it:
$ cd tests/
$ phantomjs run-qunit.js tests.html
Failed tests and their failed assertions will be printed to the console. A
results summary will be printed at the end.
To generate coverage statistics, use JSCoverage
to instrument the js files:
$ cd tests/
$ jscoverage ../js/ ../instrumented/
$ phantomjs run-qunit.js tests.html
Coverage percentage will be included in the output summary, and a highlighted
line-by-line html file will be generated.
Thanks to Rod @ While One Fork for the
CIS guide
on putting the above together.
Tests go in js files in the tests/suites/ directory tree. QUnit organizes
tests into suites called "modules"; there is one module per js file. If the
tests you are adding do not fit into an existing module, create a new one attests/suites/<new module>.js, where <new module> is a broad yet
descriptive name for the suite. If tests have many year-specific cases (ie,
behave differently in leap years vs normal years, or have specific buggy
behavior in a certain year), create the module in a new directory,tests/suites/<new module>/<year>.js, where <new module> is the decriptive
name and <year> is the four-digit year the tests pertain to.
In order for new tests to be run, they must be imported into tests/tests.html.
Find the script includes headed by the html comment <!-- Test suites -->, and
add a new one to the list which includes the new js files.
By all means, please do! Just note that I stopped working on this structure
once it fit my needs, there's no real support for it, and it may change in the
future. Otherwise, have at it.