Class: Buildr::TestTask

Inherits:
Rake::Task show all
Defined in:
lib/buildr/core/test.rb

Overview

The test task controls the entire test lifecycle.

You can use the test task in three ways. You can access and configure specific test tasks, e.g. enhance the #compile task, or run code during #setup/#teardown.

You can use convenient methods that handle the most common settings. For example, add dependencies using #with, or include only specific tests using #include.

You can also enhance this task directly. This task will first execute the #compile task, followed by the #setup task, run the unit tests, any other enhancements, and end by executing #teardown.

The test framework is determined based on the available test files, for example, if the test cases are written in Java, then JUnit is selected as the test framework. You can also select a specific test framework, for example, to use TestNG instead of JUnit:

test.using :testng

Instance Attribute Summary collapse

Class Method Summary collapse

Instance Method Summary collapse

Methods inherited from Rake::Task

#invoke, #invoke_with_call_chain

Constructor Details

#initialize(*args) ⇒ TestTask

:nodoc:



193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
 
# File 'lib/buildr/core/test.rb', line 193

def initialize(*args) #:nodoc:
  super
  @dependencies = FileList[]
  @include = []
  @exclude = []
  @forced_need = false
  parent_task = Project.parent_task(name)
  if parent_task.respond_to?(:options)
    @options = OpenObject.new { |hash, key| hash[key] = parent_task.options[key].clone rescue hash[key] = parent_task.options[key] }
  else
    @options = OpenObject.new(default_options)
  end

  unless ENV["IGNORE_BUILDFILE"] =~ /(true)|(yes)/i
    enhance [ application.buildfile.name ]
    enhance application.buildfile.prerequisites
  end
  enhance do
    run_tests if framework
  end
end

Instance Attribute Details

#dependencies ⇒ Object

The dependencies used for running the tests. Includes the compiled files (compile.target) and their dependencies. Will also include anything you pass to #with, shared between the testing compile and run dependencies.



218
219
220
 
# File 'lib/buildr/core/test.rb', line 218

def dependencies
  @dependencies
end

#failed_tests ⇒ Object (readonly)

After running the task, returns all the tests that failed, empty array if all tests passed.



379
380
381
 
# File 'lib/buildr/core/test.rb', line 379

def failed_tests
  @failed_tests
end

#options ⇒ Object (readonly)

Returns various test options.



303
304
305
 
# File 'lib/buildr/core/test.rb', line 303

def options
  @options
end

#passed_tests ⇒ Object (readonly)

After running the task, returns all the tests that passed, empty array if no tests passed.



381
382
383
 
# File 'lib/buildr/core/test.rb', line 381

def passed_tests
  @passed_tests
end

#project ⇒ Object (readonly)

The project this task belongs to.



440
441
442
 
# File 'lib/buildr/core/test.rb', line 440

def project
  @project
end

#tests ⇒ Object (readonly)

After running the task, returns all tests selected to run, based on availability and include/exclude pattern.



377
378
379
 
# File 'lib/buildr/core/test.rb', line 377

def tests
  @tests
end

Class Method Details

.only_run(tests) ⇒ Object

Used by the test/integration rule to only run tests that match the specified names.



173
174
175
176
177
178
 
# File 'lib/buildr/core/test.rb', line 173

def only_run(tests) #:nodoc:
  tests = tests.map { |name| name =~ /\*/ ? name : "*#{name}*" }
  # Since the tests may reside in a sub-project, we need to set the include/exclude pattern on
  # all sub-projects, but only invoke test on the local project.
  Project.projects.each { |project| project.test.send :only_run, tests }
end

.only_run_failed ⇒ Object

Used by the test/integration rule to only run tests that failed the last time.



181
182
183
184
185
 
# File 'lib/buildr/core/test.rb', line 181

def only_run_failed() #:nodoc:
  # Since the tests may reside in a sub-project, we need to set the include/exclude pattern on
  # all sub-projects, but only invoke test on the local project.
  Project.projects.each { |project| project.test.send :only_run_failed }
end

.run_local_tests(integration) ⇒ Object

Used by the local test and integration tasks to a) Find the local project(s), b) Find all its sub-projects and narrow down to those that have either unit or integration tests, c) Run all the (either unit or integration) tests, and d) Ignore failure if necessary.



157
158
159
160
161
162
163
164
165
166
167
168
169
170
 
# File 'lib/buildr/core/test.rb', line 157

def run_local_tests(integration) #:nodoc:
  Project.local_projects do |project|
    # !(foo ^ bar) tests for equality and accepts nil as false (and select is less obfuscated than reject on ^).
    projects = ([project] + project.projects).select { |project| !(project.test.options[:integration] ^ integration) }
    projects.each do |project|
      info "Testing #{project.name}"
      begin
        project.test.invoke
      rescue
        raise unless Buildr.options.test == :all
      end
    end
  end
end

Instance Method Details

#classes ⇒ Object

Deprecated: Use tests instead.



371
372
373
374
 
# File 'lib/buildr/core/test.rb', line 371

def classes
  Buildr.application.deprecated 'Call tests instead of classes'
  tests
end

#classpath ⇒ Object

Deprecated: Use dependencies instead.



221
222
223
224
 
# File 'lib/buildr/core/test.rb', line 221

def classpath
  Buildr.application.deprecated 'Use dependencies instead.'
  @dependencies
end

#classpath=(artifacts) ⇒ Object

Deprecated: Use dependencies= instead.



227
228
229
230
 
# File 'lib/buildr/core/test.rb', line 227

def classpath=(artifacts)
  Buildr.application.deprecated 'Use dependencies= instead.'
  @dependencies = artifacts
end

#compile(*sources, &block) ⇒ Object

:call-seq:

compile(*sources) => CompileTask
compile(*sources) { |task| .. } => CompileTask

The compile task is similar to the Project’s compile task. However, it compiles all files found in the src/test/source directory into the target/test/code directory. This task is executed by the test task before running any tests.

Once the project definition is complete, all dependencies from the regular compile task are copied over, so you only need to specify dependencies specific to your tests. You can do so by calling #with on the test task. The dependencies used here are also copied over to the junit task.



259
260
261
 
# File 'lib/buildr/core/test.rb', line 259

def compile(*sources, &block)
  @project.task('test:compile').from(sources).enhance &block
end

#default_options ⇒ Object

Default options already set on each test task.



189
190
191
 
# File 'lib/buildr/core/test.rb', line 189

def default_options
  { :fail_on_failure=>true, :fork=>:once, :properties=>{}, :environment=>{} }
end

#exclude(*names) ⇒ Object

:call-seq:

exclude(*names) => self

Exclude the specified tests. This method accepts multiple arguments and returns self. See #include for the type of arguments you can use.



365
366
367
368
 
# File 'lib/buildr/core/test.rb', line 365

def exclude(*names)
  @exclude += names
  self
end

#execute(args) ⇒ Object

:nodoc:



232
233
234
235
236
237
238
239
240
241
242
243
244
245
 
# File 'lib/buildr/core/test.rb', line 232

def execute(args) #:nodoc:
  if Buildr.options.test == false
    info "Skipping tests for #{project.name}"
    return
  end
  setup.invoke
  begin
    super
  rescue RuntimeError
    raise if options[:fail_on_failure] && Buildr.options.test != :all
  ensure
    teardown.invoke
  end
end

#failures_to ⇒ Object

:call-seq:

failures_to => file

We record the list of failed tests for the current framework in this file.



416
417
418
 
# File 'lib/buildr/core/test.rb', line 416

def failures_to
  @failures_to ||= file(@project.path_to(:target, "#{framework}-failed")=>self)
end

#framework ⇒ Object

:call-seq:

framework => symbol

Returns the test framework, e.g. :junit, :testng.



387
388
389
390
391
392
393
394
395
396
397
 
# File 'lib/buildr/core/test.rb', line 387

def framework
  unless @framework
    # Start with all frameworks that apply (e.g. JUnit and TestNG for Java),
    # and pick the first (default) one, unless already specified in parent project.
    candidates = TestFramework.frameworks.select { |cls| cls.applies_to?(@project) }
    candidate = @project.parent && candidates.detect { |framework| framework.to_sym == @project.parent.test.framework } ||
      candidates.first
    self.framework = candidate if candidate
  end
  @framework && @framework.class.to_sym
end

#include(*names) ⇒ Object

:call-seq:

include(*names) => self

Include only the specified tests. Unless specified, the default is to include all tests identified by the test framework. This method accepts multiple arguments and returns self.

Tests are specified by their full name, but you can use glob patterns to select multiple tests, for example:

test.include 'com.example.FirstTest'  # FirstTest only
test.include 'com.example.*'          # All tests under com/example
test.include 'com.example.Module*'    # All tests starting with Module
test.include '*.{First,Second}Test'   # FirstTest, SecondTest



355
356
357
358
 
# File 'lib/buildr/core/test.rb', line 355

def include(*names)
  @include += names
  self
end

#last_failures ⇒ Object

:call-seq:

last_failures => array

We read the last test failures if any and return them.



425
426
427
 
# File 'lib/buildr/core/test.rb', line 425

def last_failures
  @last_failures ||= failures_to.exist? ? File.read(failures_to.to_s).split('\n') : []
end

#last_successful_run_file ⇒ Object

The path to the file that stores the time stamp of the last successful test run.



430
431
432
 
# File 'lib/buildr/core/test.rb', line 430

def last_successful_run_file #:nodoc:
  File.join(report_to.to_s, 'last_successful_run')
end

#report_to ⇒ Object

:call-seq:

report_to => file

Test frameworks that can produce reports, will write them to this directory.

This is framework dependent, so unless you use the default test framework, call this method after setting the test framework.



406
407
408
 
# File 'lib/buildr/core/test.rb', line 406

def report_to
  @report_to ||= file(@project.path_to(:reports, framework)=>self)
end

#resources(*prereqs, &block) ⇒ Object

:call-seq:

resources(*prereqs) => ResourcesTask
resources(*prereqs) { |task| .. } => ResourcesTask

Executes by the #compile task to copy resource files over. See Project#resources.



268
269
270
 
# File 'lib/buildr/core/test.rb', line 268

def resources(*prereqs, &block)
  @project.task('test:resources').enhance prereqs, &block
end

#setup(*prereqs, &block) ⇒ Object

:call-seq:

setup(*prereqs) => task
setup(*prereqs) { |task| .. } => task

Returns the setup task. The setup task is executed at the beginning of the test task, after compiling the test files.



278
279
280
 
# File 'lib/buildr/core/test.rb', line 278

def setup(*prereqs, &block)
  @project.task('test:setup').enhance prereqs, &block
end

#teardown(*prereqs, &block) ⇒ Object

:call-seq:

teardown(*prereqs) => task
teardown(*prereqs) { |task| .. } => task

Returns the teardown task. The teardown task is executed at the end of the test task.



287
288
289
 
# File 'lib/buildr/core/test.rb', line 287

def teardown(*prereqs, &block)
  @project.task('test:teardown').enhance prereqs, &block
end

#timestamp ⇒ Object

The time stamp of the last successful test run. Or Rake::EARLY if no successful test run recorded.



435
436
437
 
# File 'lib/buildr/core/test.rb', line 435

def timestamp #:nodoc:
  File.exist?(last_successful_run_file) ? File.mtime(last_successful_run_file) : Rake::EARLY
end

#using(*args) ⇒ Object

:call-seq:

using(options) => self

Sets various test options from a hash and returns self. For example:

test.using :fork=>:each, :properties=>{ 'url'=>'http://localhost:8080' }

Can also be used to select the test framework, or to run these tests as integration tests. For example:

test.using :testng
test.using :integration

The :fail_on_failure option specifies whether the task should fail if any of the tests fail (default), or should report the failures but continue running the build (when set to false).

All other options depend on the capability of the test framework. These options should be used the same way across all frameworks that support them:

  • :fork – Fork once for each project (:once, default), for each test in each

    project (:each), or don't fork at all (false).

  • :properties – Properties pass to the test, e.g. in Java as system properties.

  • :environment – Environment variables. This hash is made available in the

    form of environment variables.



327
328
329
330
331
332
333
334
335
336
337
338
339
340
 
# File 'lib/buildr/core/test.rb', line 327

def using(*args)
  args.pop.each { |key, value| options[key.to_sym] = value } if Hash === args.last
  args.each do |name|
    if TestFramework.has?(name)
      self.framework = name
    elsif name == :integration
      options[:integration] = true
    else
      Buildr.application.deprecated "Please replace with using(:#{name}=>true)"
      options[name.to_sym] = true
    end
  end
  self
end

#with(*artifacts) ⇒ Object

:call-seq:

with(*specs) => self

Specify artifacts (specs, tasks, files, etc) to include in the dependencies list when compiling and running tests.



296
297
298
299
300
 
# File 'lib/buildr/core/test.rb', line 296

def with(*artifacts)
  @dependencies |= Buildr.artifacts(artifacts.flatten).uniq
  compile.with artifacts
  self
end