Q.E.D. is an abbreviation for the well known Latin phrase “Quod Erat Demonstrandum”, literally “which was to be demonstrated”, which is oft written in its abbreviated form at the end of a mathematical proof or philosophical argument to signify a successful conclusion. And so it is too for Ruby Q.E.D., though it might as easily be taken to stand for “Quality Ensured Documentation”.
QED is in fact both a test framework and a documentation system for Ruby developers. QED sits somewhere between lower-level testing tools like Test::Unit and grandiose requirement specifications systems like Cucumber. In practice it works exceptionally well for API-Driven Design, which is especially useful when designing reusable libraries, but it can be used to test code at any level of abstraction, from unit test to systems tests.
Write tests and documentation in the same breath!
Demos can be RDoc, Markdown or any other conforming text format.
Can use any BRASS compliant assertion framework, such as the the excellent AE (Assertive Expressive) library.
Data and Table macros allows large sets of data to be tested by the same code.
Documentation tool provides nice output with jQuery-based TOC.
QED can use any BRASS compliant assertions framework. Simply require the library in ones applique (see below). Traditionally this has been the AE (Assertive Expressive) library, which provides an elegant means to make assertions. To give a quick overview, assertion can be written as:
4.assert == 5
In this example, because 4 != 5, this expression will raise an Assertion exception. QED's Runner class is thus just a means of running and capturing code blocks containing such assertions.
QED documents are simply text files called demonstrandum (demos for short). Because they largely consist of free-form descriptive text, they are a practice pure Literate Programming. For example:
= Example Shows that the number 5 does not equal 4. 5.assert! == 4 But in fact equals 5. 5.assert == 5
In this example RDoc was chosen for the document format. However, almost any text format can be used. The only necessary distinction is that description text align to the left margin and all code be indented, although QED does recognize RDoc and Markdown single-line style headers, so any format that supports those (which covers many markup formats in use today) will have mildly improved console output. In any case, the essential take away here is that QED demonstrandum are simply descriptive documents with interspersed blocks of example code.
Give this design some thought. It should become clear that this approach is especially fruitful in that it allows documentation and specification to seamlessly merge into a unified demonstration.
If we were to run the above document through QED in verbatim mode the output would be identical (assuming we did not make a typo and the assertions passed). If there were errors or failures, we would see information detailing each.
To run a document through QED, simply use the qed command.
$ qed -v demo/01_example.rdoc
The -v option specifies verbatim mode, which outputs the entire document.
Notice we placed the QED document in a demo/ directory, there is no standard place that demonstrations have to go. They can be placed anywhere in a project that is preferred (lately I've been use `qed/`). However, the `qed` command will look for `qed/`, `demo/`, `demos/` and `spec/` in that order if no location is specified.
Also notice the use of `01_` prefix in front of the file name. While this is not strictly necessary, it helps order the documents nicely, in particular when generating QED documentation (“QEDocs”).
QED demonstrandum descriptive text is not strictly passive explanation. Using pattern matching techniques, document phrases can trigger underlying actions. These actions provide a support structure for running tests called the applique.
Creating an applique is easy. Along with your QED scripts, to which the applique will apply, create an applique/ directory. In this directory add Ruby scripts. When you run your demos every Ruby script in the directory will be automatically loaded.
Within these applique scripts advice can be defined. Advice can be either *event advice*, which is simply triggered by some fixed cycle of running, such as Before :each or After :all, and *pattern advice* which are used to match against descriptive phrases in the QED demos. An example would be:
When "a new round is started" do @round =  end
So that whenever the phrase “a new round is started” appears in a demo, the @round instance variable with be reset to an empty array.
It is rather amazing what can be accomplished with such a system, be sure to look at QED's own demonstrandum to get a better notion of how you can put the the system to use.
QED uses the RC library to handle configuration.
Here's a generally useful example of using SimpleCov to generate a test coverage report when running your QED demos. Add this to your `Config.rb` file:
config :qed, :coverage do require 'simplecov' SimpleCov.start do coverage_dir 'log/coverage' end end
You can use the profile via the `-p/–profile` option on the command line:
$ qed -p coverage
Or by setting the `profile` environment variable.
$ profile=coverage qed
To generate documentation from QED documents, use the qedoc command.
$ qedoc --output doc/qedoc --title "Example" demo/*.rdoc
When documenting, QED recognizes the format by the file extension and treats it accordingly. An extension of .qed is treated the same as .rdoc.
Use the --help options on each command to get more information on the use of these commands.
QED depends on the following external libraries:
BRASS - Assertions System
ANSI - ANSI Color Codes
RC - Runtime Configuration
Facets - Core Extensions
These will be automatically installed when installing QED via RubyGems, if they are not already installed.
Optional libraries that are generally useful with QED.
AE - Assertions Framework
Install these individually and require them in your applique to use.
QED uses itself for testing, which can be a bit tricky. But works fine for the most part. In the future we may add some addition tests via another test framework to ensure full coverage. But for now QED is proving sufficient.
To run the tests, use `qed` command line tool –ideally use `$ ruby -Ilib bin/qed` to ensure the current version of QED is being used.
For convenience, use `$ fire spec` to run the test specifications. To also generate a test coverage report use `$ fire spec:cov`.
(BSD 2 Clause license)
Copyright © 2009 Rubyworks, Thomas Sawyer
See COPYING.rdoc for details.