Test::Unit::Mock Example Usage
Test::Unit::Mock is designed to allow you do make a mocked object that will respond to all the methods of the real class (albeit probably not with correct results) with one line of code:
mockSocket = Test::Unit::MockObject( TCPSocket ).new
You can then specify return values for the methods you wish to test in one of several different ways:
# Make the #addr method return three cycling values (which # will be repeated when it reaches the end mockSocket.setReturnValues( :addr => [ ["AF_INET", 23, "localhost", "127.0.0.1"], ["AF_INET", 80, "slashdot.org", "66.35.250.150"], ["AF_INET", 2401, "helium.ruby-lang.org", "210.251.121.214"], ] ) # Make the #write and #read methods call a Proc and a Method, # respectively mockSocket.setReturnValues( :write => Proc::new {|str| str.length}, :read => method(:fakeRead) ) # Set up the #getsockopt method to return a value based on the # arguments given: mockSocket.setReturnValues :getsockopt => { [Socket::SOL_TCP, Socket::TCP_NODELAY] => [0].pack("i_"), [Socket::SOL_SOCKET, Socket::SO_REUSEADDR] => [1].pack("i_"), }
You can also set the order in which you expect methods to be called, but you don't have to do so if you don't care:
mockSocket.setCallOrder( :addr, :getsockopt, :write, :read, :write, :read )
By default, when testing for call order, other method calls may be interspersed between the calls specified without effect, and only a missing or misordered method call causes the assertions to fail. If you want the call order to be adhered to strictly, you can set that:
mockSocket.strictCallOrder = true
Then, when you're ready to test, just activate the object and send it off to whatever code you're testing:
mockSocket.activate testedObject.setSocket( mockSocket ) ... # Check method call order on the mocked socket (adds assertions) mockSocket.verify
Assertion failures contain a message that specifies exactly what went wrong:
$ ruby misc/readmecode.rb
Loaded suite misc/readmecode
Started
.FF
Finished in 0.31509 seconds.
1) Failure:
test_incorrectorder(MockTestExperiment)
[/usr/lib/ruby/site_ruby/1.8/test/unit/mock.rb:288:in `verify'
misc/readmecode.rb:105:in `test_incorrectorder']:
Call order assertion failed: Expected call to :write, but got call to :read
from misc/readmecode.rb:94:in `test_incorrectorder' at 0.00076 instead
2) Failure:
test_missingcall(MockTestExperiment)
[/usr/lib/ruby/site_ruby/1.8/test/unit/mock.rb:288:in `verify'
misc/readmecode.rb:126:in `test_missingcall']:
Call order assertion failed: Missing call to :read.
3 tests, 3 assertions, 2 failures, 0 errors
If you require more advanced functionality, you can also use the mocked object class as a superclass:
# Create a mock socket class class MockSocket < Test::Unit::MockObject( TCPSocket ) def initialize super setCallOrder( :read, :read, :read, :write, :read ) strictCallOrder = true @io = '' end def read( len ) super # Call the mocked method to record the call rval = @io[0,len] @io[0,len] = '' return rval end def write( str ) super # Call the mocked method to record the call @io += str return str.length end end
You can also add debugging to your tests to give you a timestamped history of each call:
# Call the methods in the correct order mockSocket.addr mockSocket.getsockopt( Socket::SOL_TCP, Socket::TCP_NODELAY ) mockSocket.write( "foo" ) mockSocket.read( 1024 ) mockSocket.write( "bar" ) mockSocket.read( 4096 ) # Check method call order on the mocked socket mockSocket.verify if $DEBUG puts "Call trace:\n\t" + mockSocket.callTrace.join("\n\t") end
This outputs something like:
Call trace:
addr( ) at 0.00015 seconds from misc/readmecode.rb:64:
in `test_correctorder'
getsockopt( 6,1 ) at 0.00030 seconds from misc/readmecode.rb:65:
in `test_correctorder'
write( "foo" ) at 0.00040 seconds from misc/readmecode.rb:66:
in `test_correctorder'
read( 1024 ) at 0.00050 seconds from misc/readmecode.rb:67:
in `test_correctorder'
write( "bar" ) at 0.00063 seconds from misc/readmecode.rb:68:
in `test_correctorder'
read( 4096 ) at 0.00072 seconds from misc/readmecode.rb:69:
in `test_correctorder'
