mirror of
https://github.com/silverstripe/silverstripe-framework
synced 2024-10-22 14:05:37 +02:00
| .. | ||
| bin | ||
| lib | ||
| server | ||
| spec | ||
| templates | ||
| .piston.yml | ||
| History.rdoc | ||
| jspec.gemspec | ||
| Manifest | ||
| Rakefile | ||
| README.rdoc | ||
= JSpec
JSpec is a minimalistic JavaScript behavior driven development framework,
providing simple installation, extremely low learning curve, absolutely no pollution
to core prototypes, async request support, and incredibly sexy syntax, tons of matchers
and much more.
== Features
* Highly readable
* Framework / DOM independent
* Modular via JSpec Module's and hooks
* Mock Ajax Requests
* Rhino support
* Node.js support
* Async support
* Fixture support
* Ruby JavaScript testing server
* Nested describes
* Does not pollute core object prototypes
* Cascading before/after/before_each/after_each hooks
* Extremely simple and intuitive matcher declaration
* Over 45 core matchers
* Allows parens to be optional when using matchers to increase readability
* Several helpful formatters (DOM, Console, Terminal, ...)
* Assertion graphs displaying how many, and which assertions pass or failed
* Default / customizable evaluation contexts
* DOM sandbox support
* Great looking default DOM theme
* `jspec` command-line utility for auto-running specs, and initializing project templates
* Proxy or 'Spy' assertions
* Method Stubbing
* Shared behaviors
* Profiling
* Interactive Shell
* Ruby on Rails Integration
* Tiny (15 kb compressed, 1300-ish LOC)
== Installation
Simply download JSpec and include JSpec.css and JSpec.js in your markup.
Head over to the downloads section on Github, clone this public repo, or
add JSpec as a git submodule with in your project. Alternatively JSpec is
also available as a Ruby Gem (though this is not required), which also
provides the `jspec` executable. First install [Gemcutter](http://gemcutter.org/) then execute:
$ sudo gem install jspec
At which point you may:
$ jspec init myproject
By default, the command above will use absolute path for all JSpec library files.
This behavior can be a problem when you're working across different computers or
operating systems. You can freeze the library or symlink it.
$ jspec init myproject --freeze
$ jspec init myproject --symlink
JSpec scripts should NOT be referenced via the <script> tag, they should be
loaded using the exec method (unless you are using the grammar-less alternative).
Below is an example:
...
<script>
function runSuites() {
JSpec
.exec('spec.core.js')
.exec('spec.jquery.js')
.run({ failuresOnly : true })
.report()
}
</script>
<body onLoad="runSuites()">
...
You may optionally want to use sources in the /pkg directory
for your project, since it includes compressed alternatives generated
each release.
== Example
describe 'ShoppingCart'
before_each
cart = new ShoppingCart
end
describe 'addProducts'
it 'should add several products'
cart.addProduct('cookie')
cart.addProduct('icecream')
cart.should.have 2, 'products'
end
end
describe 'checkout'
it 'should throw an error when checking out with no products'
-{ cart.clear().checkout() }.should.throw_error EmptyCart
end
end
end
== Grammar-less Example
JSpec's grammar is optional, you may also use the equivalent grammar-less
alternative below using pure JavaScript (when using the JSpec grammar you
may also use grammar-less assertions):
JSpec.describe('ShoppingCart', function(){
before_each(function{
cart = new ShoppingCart
})
describe('addProducts', function(){
it ('should add several products', function(){
cart.addProducts('cookie')
cart.addProducts('icecream')
expect(cart).to(have, 2, 'products')
})
})
describe('checkout', function(){
it ('should throw an error when checking out with no products', function(){
expect(function(){ cart.clear().checkout() }).to(throw_error, EmptyCart)
})
})
})
== Options
You may alter the way JSpec operates by assigning options via the
JSpec.options hash, by passing string-based option values via the
query string, or passing a hash to run(). For example
JSpec.options.failuresOnly = true, and ?failuresOnly=1 will both work.
* profile {bool} when enabled, uses console.time() in order to display performance information in your console log as specs are completed. (DOM, Console)
* failuresOnly {bool} displays only failing specs, making them quick to discover and fix (DOM, Terminal, Server)
* reportToId {string} an element id to report to when using the DOM formatter (DOM)
* verbose {bool} verbose server output, defaults to false (Server)
== Matchers
* Core
- equal, be ===
- be_a, be_an have constructor of x
- be_an_instance_of instanceof x
- be_at_least >=
- be_at_most <=
- be_null == null
- be_empty length < 0 or {}
- be_true == true
- be_false == false
- be_type be type of x
- be_greater_than >
- be_less_than <
- be_undefined check if variable passed is undefined
- throw_error should throw an error, optionally supply the error string or regexp for message comparison
- have object should have n of property (person.should.have(2, 'pets'))
- have_at_least object should have at least n of property
- have_at_most object should have a maximum n of property
- have_within object should have within n..n of property (person.should.have_within(1..3, 'pets')
- have_length length of n
- have_prop object should have property x, optionally supplying an expected value
- have_property strict version of have_prop
- be_within checks if n is within the range passed
- include include substring, array element, or hash key
- match string should match regexp x
- respond_to property x should be a function
- eql matches simple literals (strings, numbers) with ==
However composites like arrays or 'hashes' are recursively matched,
meaning that [1, 2, [3]].should_eql([1, 2, [3]]) will be true.
* jQuery
- have_tag, have_one have exactly one tag
- have_tags, have_many have more than one tag
- have_child have exactly one child
- have_children have more than one child
- have_text have plain text
- have_attr have an attribute, with optional value
- have_type
- have_id
- have_title
- have_alt
- have_href
- have_rel
- have_rev
- have_name
- have_target
- have_value
- have_class
- have_classes
- be_visible
- be_hidden
- be_enabled
- be_disabled
- be_selected
- be_checked
== Async Support With Mock Timers
The javascript mock timers library is available at http://github.com/visionmedia/js-mock-timers
although it is already bundled with JSpec at lib/jspec.timers.js
Timers return ids and may be passed to clearInterval(), however
they do not execute in threads, they must be manually scheduled and
controlled via the tick() function.
setTimeout(function(){
alert('Wahoo!')
}, 400)
tick(200) // Nothing happens
tick(400) // Wahoo!
setInterval() works as expected, although it persists, where as setTimeout()
is destroyed after a single call. As conveyed by the last tick() call below,
a large increment in milliseconds may cause the callbacks to be called several times
to 'catch up'.
progress = ''
var id = setInterval(function(){
progress += '.'
}, 100)
tick(50), print(progress) // ''
tick(50), print(progress) // '.'
tick(100), print(progress) // '..'
tick(100), print(progress) // '...'
tick(300), print(progress) // '......'
clearInterval(id)
tick(800) // Nothing happens
You may also reset at any time using resetTimers()
== Proxy Assertions
Proxy or 'Spy' assertions allow you to assert that a method is called n number
of times, with x arguments, returning x value. For example:
person = { getPets : function(species){ return ['izzy'] }}
person.should.receive('getPets', 'twice').with_args(an_instance_of(String))and_return(['izzy'])
person.getPets('dog') // This will pass
person.getPets() // This will fail because we asked an instance of String
This is a useful mechanism for testing the behavior of your object, as well as
how other methods may interact with it. Below is another example:
array = ['foo', 'bar']
array.should.receive('toString').and_return('foo,bar')
'array: ' + array // This line causes the spec to pass due to calling toString()
For more examples view spec/spec.matchers.js
== Method Stubbing
JSpec currently provides very simple stubbing support shown below:
person = { toString : function(){ return '<Person>' } }
stub(person, 'toString').and_return('Ive been stubbed!')
After each spec all stubs are restored to their original methods so
there is no reason to explicitly call destub(). To persist stubs,
use a before_each hook:
before_each
stub(someObject, 'method').and_return({ some : thing })
end
To destub a method simply call destub() at any time:
destub(person, 'toString')
If you would like to whipe an object clear of stubs simply pass it
to destub() without an additional method argument:
destub(person)
Alternatively both these utility functions may be called as methods
on any object when using the JSpec grammar:
someObject.stub('method').and_return('whatever')
// Converted to stub(someObject, 'method').and_return('whatever')
== Helpers
* Core
- an_instance_of used in conjunction with the 'receive' matcher
- mockRequest, mock_request mock a request (requires jspec.xhr.js)
- unmockRequest, unmock_request unmock requests (requests jspec.xhr.js)
* jQuery
- sandbox used to generate new DOM sandbox, using jQuery object
- element same as invoking jQuery, just reads better and no need to worry about $ collisions
- elements alias of element
== Shared Behaviors
JSpec's support for shared behaviors allows multiple suites or describe blocks to share
common functionality. For example an Admin, would inherit all specs of User:
describe 'User'
before
User = function(name) { this.name = name }
user = new User('joe')
end
it 'should have a name'
user.should.have_property 'name'
end
describe 'Administrator'
should_behave_like('User')
before
Admin = function(name) { this.name = name }
Admin.prototype.may = function(perm){ return true }
user = new Admin('tj')
end
it 'should have access to all permissions'
user.may('edit pages').should.be_true
end
end
end
NOTE: both User and Administrator's before hooks implement the 'user' variable
== Mock Ajax Requests
JSpec supports generic Ajax mocking which is usable with any JavaScript framework via 'jspec.xhr.js'. The
API is comprised of two functions, mockRequest() and unmockRequest(). unmockRequest() is
automatically called after each specification to restore the default functionality of XMLHttpRequest,
so it is uncommon to call unmockRequest() directly. Below is a jQuery example:
it 'should mock requests'
mockRequest().and_return('{ foo : "bar" }', 'application/json')
$.getJSON('foo', function(response, statusText){
response.foo.should.eql 'bar'
})
end
The mockRequest().and_return signature is as follows:
mockRequest().and_return(<data>, [content-type], [response-status-code], [headers-hash])
At the moment mockRequest() itself does not accept any arguments, however in the future
this will be used to target specific uris for mocking.
NOTE: works with Rhino as well
== Hooks
Currently the following hooks are supported, and may be utilized any number of times as they
are simply pushed to a stack. So for instance you may have two before_each blocks within the same
scope, they will both run, but this can help keep your specs readable.
* before run once before the suite is executed
* after run once after the suite is executed
* before_each run before each specification
* after_each run after each specification
== Custom Contexts
Custom contexts can be applied to supply helper
methods or properties to all subsequent bodies (other hooks, or specs).
Keep in mind that when replacing the default context you will loose
functionality provided by it, unless you manually merge it with your
custom context.
To reset the context simply assign null to obtain the original context.
...
before
JSpec.context = { foo : 'bar' }
end
after
JSpec.context = null
end
it 'will work ;)'
foo.should_equal 'bar'
end
...
== Async Support
Currently only jspec.jquery.js supports async requests. JSpec uses jQuery.ajaxSetup and sets all
requests to sync, which preserves execution order, and reports correctly.
it 'should load mah cookies (textfile)'
$.post('async', function(text){
text.should_eql 'cookies!'
})
end
== Pre-processor
The pre-processing capability of JSpec is extremely powerful. Your JavaScript
code is not necessarily what it seems. For example when you seemingly invoke a
object's prototype like below:
'foobar'.should.include 'bar'
First parens are added:
'foobar'.should.include('bar')
Secondly the matcher invocation is converted to a non-polluting match() call:
JSpec.match('foobar', 'should', 'include', 'bar')
This also means instead of:
var object = { foo : 'bar' }
object.should.include 'foo'
We can do:
{ foo : 'bar' }.should.include 'foo'
=== Closure Literal
These are equivalent:
-{ throw 'test' }.should.throw_error
function() { throw 'test' }.should.throw_error
=== Inclusive Range Literal
The following expands to the array of [1,2,3,4,5]
n.should.be_within 1..5
== Formatters
To change a formatter simply alter the options hash like below, assigning
a new constructor, or pass it within the hash to run():
JSpec.options.formatter = JSpec.formatters.Console
OR
JSpec
.exec('...')
.run({ formatter : JSpec.formatters.Terminal })
.report()
== Fixtures
The fixture() utility function may be used in order to load arbitrary file contents
for use with your specifications. JSpec will resolve fixture('data') in the following
manor:
- 'data'
- 'spec/data'
- 'spec/fixtures/data'
- 'spec/fixtures/data.html'
So if the file 'spec/fixtures/data.html' exists, we can simply use fixture('data'),
where as 'spec/fixtures/xml/data.xml' must be specified with fixture('xml/data.xml').
If you prefer not to store fixtures in the 'fixtures' directory you must be more specific
with the path supplied.
== Testing DOM Elements
When using jQuery testing DOM elements is very easy. Many may think they require specific
sandbox divs in their html, however you do not. Using the fixture support mentioned above
you may simply load some HTML, and use the 'elements()' utility which is an alias of jQuery:
describe 'JSpec DOM testing'
describe 'is so easy'
before_each
list = elements(fixture('users-list'))
// or list = jQuery(fixture('users-list'))
// or list = $(fixture('users-list'))
end
it 'should have users'
list.should.have_tag 'ul'
end
end
end
You may also use simple strings, since jQuery's constructor will convert them to DOM elements:
describe 'Something'
before_each
html = elements('<p>Foo</p>')
// or html = $('<p>Foo</p>') ...
end
it 'should do something'
html.should.have_text 'Foo'
end
end
== Custom Matchers
First lets create a simple equality matcher. In the case below JSpec is smart enough to realize
this is simply a binary operator, and simply transforms this into 'actual === expected'
JSpec.addMatchers({
equal : '==='
})
To alias a method to keep your specs readable you may alias them like below:
JSpec.addMatchers({
be : 'alias equal'
})
'foo'.should.equal 'foo'
true.should.be true
Matchers with string bodies implicitly return the expression value.
The expanded version of the equal matcher would then be:
JSpec.addMatchers({
equal : 'actual === expected'
})
Large matchers or those which require several parameters may wish
to utilize the hash method:
JSpec.addMatchers({
equal : { match : function(actual, expected){
return actual === expected
}}
})
To keep JSpec tiny, JSpec will default to generating failure messages
for you, how ever this can be explicitly defined:
JSpec.addMatchers({
equal : {
match : function(actual, expected){
return actual === expected
},
message : function(actual, expected, negate) {
return 'a message here'
}
}
})
When defining matchers that are extremely similar in functionality, however
require different names, you may use a prefixed list of words like below which
defines be_disabled, be_selected, be_checked, and have_type, have_id, etc. Each
function must return the matcher body which will be used.
JSpec.addMatchers({
'be disabled selected checked' : function(attr) {
return 'jQuery(actual).attr("' + attr + '")'
},
'have type id title alt href src sel rev name target' : function(attr) {
return function(actual, value) {
return value ? jQuery(actual).attr(attr) == value:
jQuery(actual).attr(attr)
}
}
})
== Extending Or Hooking Into JSpec
JSpec provides a hook architecture for extending or analyzing various
points in its execution, through the use of 'Modules'. For a Module
example view lib/jspec.jquery.js.
The following methods or properties are utilized by JSpec:
- name : module name string
- init : called to initialize a module
- formatters : hash of formatters merged with JSpec.formatters
- utilities : hash of utility functions merged with JSpec.defaultContext
- matchers : hash of matchers merged with JSpec's core matchers via JSpec.addMatchers()
- DSLs : hash of DSL methods; for example DSLs.snake contains before_each, after_each, etc.
Where as DSLs.camel may contain beforeEach, afterEach, etc.
Below is a list of hooks, descriptions, and valid return values which
may simply be implemented as module methods. beforeSuite, afterSuite, beforeSpec, and afterSpec have lower
precedence than before_each, after_each etc within the specs themselves, allowing them to override or undo
anything that has been done by a Module.
- running(options) : started running JSpec with the options passed : returning 'stop' will halt running
- loading(file) : loading a file : returning 'stop' will prevent loading
- executing(file) : executing a file : returning 'stop' will prevent execution
- posting(data, url) : posting data to a url : returning 'stop' will prevent request
- preprocessing(input) : before input string is preprocessed : return input string for next hook to preprocess
- stubbing(object, method, result) : called when stubbing an object's method, and return value (result). : (no return value)
- requiring(dependency, message) : requiring a dependency : (no return value)
- beforeAssertion(assertion) : before an assertion has been made : (no return value)
- afterAssertion(assertion) : after an assertion has been made : (no return value)
- addingMatcher(name, body) : unprocessed matcher name and body : (no return value)
- addingSuite(suite) : adding Suite instance to JSpec : (no return value)
- beforeSuite(suite) : before running of suite (describe block) : (no return value)
- afterSuite(suite) : after running of suite (describe block) : (no return value)
- beforeSpec(spec) : before running of spec (it block) : (no return value)
- afterSpec(spec) : after running of spec (it block) : (no return value)
- reporting(options) : called before reporting : (no return value)
- evaluatingBody(dsl, matchers, context, contents) : evaluating body contents, with the given context, matchers and dsl. : (no return value)
For example you may wish to proxy files which are being executed, simply implement the
executing method like below. This example will stop execution of any file matching /matchers/.
MyModule = {
executing : function(file) {
if (file.match(/matchers/))
return 'stop'
}
}
JSpec.include(MyModule)
Immutable values may also be passed to hooks using hookImmutable() internally. This allows
for simple numbers, strings, etc to be utilized or altered within a hook implementation. Below
is an example module which adds functionality to the JSpec grammar by converting SomeObject.stub('method')
to stub(SomeObject, 'method'):
JSpec.include({
preprocessing : function(input) {
return input.replace(/(\w+)\.(stub|destub)\((.*?)\)$/gm, '$2($1, $3)')
}
})
== JSpec Command-line Utility
When installed as a Ruby Gem, the `jspec` executable will become available,
allowing you to initialize project templates quickly, as well as auto-testing
specifications when a file is altered.
Initialize JSpec-driven project template in directory 'myproject':
$ jspec init myproject
Once within 'myproject' start testing by executing:
$ jspec
For additional usage execute:
$ jspec help
Or for specific usage:
$ jspec help run
== Rhino
JSpec provides transparent support for Rhino, while using the Terminal formatter.
Simply create a JavaScript file with contents similar to below, and then execute
the command following it:
load('lib/jspec.js')
JSpec
.exec('spec/spec.grammar.js')
.exec('spec/spec.core.js')
.run({ formatter : JSpec.formatters.Terminal, failuresOnly : true })
.report()
Initialize project with:
$ jspec init myproject
Run with:
$ jspec run --rhino
Or bind (automated testing):
$ jspec --rhino
== Server
The Ruby JavaScript testing server included with JSpec simply runs
the spec suites within each browser you specify, while reporting result
back to the terminal. It is essentially the same as using the DOM formatter
and auto-testing each browser, however results are centralized to the terminal,
removing the need to manually view each browser's output.
When utilizing the server if a file named spec/jspec.rb (or jspec/jspec.rb for rails)
is present, then it will be loaded before the server is started. This allows you to
add Sinatra routes, support additional Browsers, etc.
Run with all supported browsers:
$ jspec run --server
Run with specific browsers:
$ jspec run --browsers Safari,Firefox,Chrome,Explorer
Run with alternative browser names:
$ jspec run --browsers safari,ff,chrome,ie
Browsers supported in core:
Browser::Safari
Browser::Chrome
Browser::Opera
Browser::Firefox
Browser::IE
Supplied routes:
/slow/NUMBER
/status/NUMBER
For example $.get('/slow/4', function(){}) will take 4 seconds
to reply, where as $.get('/status/404', function(){}) will respond
with an 404 status code. Add additional Sinatra routes to the jspec.rb
file to add your own functionality.
== Interactive Shell
JSpec provides an interactive shell through Rhino, utilize with:
$ jspec shell
Or to specify additional files to load:
$ jspec shell lib/*.js
Or view additional shell help
$ jspec help shell
== Ruby on Rails
No additional gems are required for JSpec to work with rails, although
http://github.com/bhauman/jspec-rails has been created by 'bhauman'. JSpec
supports Rails out of the box, simply execute:
$ jspec init --rails
Then while still in the root directory of your Rails project, run the following
command which will bind to, and refresh your browsers automatically when any changes
are made to ./public/javascripts/*.js or ./jspec/*.js
$ jspec
Or just like regular JSpec applications, run once:
$ jspec run
Or run via the terminal using Rhino:
$ jspec run --rhino
== Support Browsers
Browsers below are supported and can be found in server/browsers.rb, however
your spec/server.rb file may support additional browsers.
* Safari
* Chrome
* Firefox
* Opera
* Internet Explorer
== Known Issues
* Tabs may cause a parse error. To prevent this use 'soft tabs' (setting in your IDE/Editor)
or use JSpec's grammar-less alternative (mentioned above).
* The preprocessor is not (yet) capable of multiline conversions. For example the following is invalid
object.stub('getContentsOfURL').and_return(function(url){
return 'html'
})
In cases such as this, you may always revert to utilizing JSpec in a grammar-less form as follows:
stub(object, 'getContentsOfURL').and_return(function(url){
return 'html'
})
== Additional JSpec Modules
* JSocka stubbing http://github.com/gisikw/jsocka/tree/master
== More Information
* IRC Channel irc://irc.freenode.net#jspec
* Featured article in JSMag: http://www.jsmag.com/main.issues.description/id=21/
* Syntax comparison with other frameworks http://gist.github.com/92283
* Get the TextMate bundle at https://github.com/visionmedia/jspec.tmbundle/tree
* For more information consult the JSpec source code documentation or visit http://visionmedia.github.com/jspec
* jQuery + HTML fixture example http://gist.github.com/147831
== Contributors
Many ideas and bug reports were contributed by
the following developers, thankyou for making
JSpec more enjoyable, and bug free ;)
* Lawrence Pit
* mpd@jesters-court.ne
* kevin.gisi@gmail.com
* tony_t_tubbs@yahoo.com
* enno84@gmx.net
* fnando
== License
(The MIT License)
Copyright (c) 2008 - 2009 TJ Holowaychuk <tj@vision-media.ca>
Permission is hereby granted, free of charge, to any person obtaining
a copy of this software and associated documentation files (the
'Software'), to deal in the Software without restriction, including
without limitation the rights to use, copy, modify, merge, publish,
distribute, sublicense, and/or sell copies of the Software, and to
permit persons to whom the Software is furnished to do so, subject to
the following conditions:
The above copyright notice and this permission notice shall be
included in all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED 'AS IS', WITHOUT WARRANTY OF ANY KIND,
EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY
CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.