From: Eric Hodel <drbrain@...7.net>
Date: 2007-08-08T16:08:10+09:00
Subject: Re: capturing stdout during unit tests

On Aug 7, 2007, at 20:59, Mark Slater wrote:
> I've written a script for use with Nagios (a monitoring tool) that  
> I'd like to test. The requirements for the script is that it output  
> a single line of text to STDOUT and then exit with a status code (0  
> - 3). My script is designed to monitor a few different services...  
> each invocation includes a required command line argument  
> specifying the service to monitor in the current invocation. To do  
> that, I created a class for each type of service monitor.
>
> I'm now writing unit tests for my script, and I'd like to check  
> that the output written to STDOUT by each service monitor class is  
> correct. But I'm new to Ruby and I'm not sure how to do that. In  
> other languages, I'd simply redefine STDOUT as a stream that goes  
> to a large in-memory buffer, but I haven't seen anything that  
> suggests that is possible in Ruby. The best I've come up with so  
> far is creating a temporary file and calling $stdout.reopen() with  
> the path to that temporary file. However, I'd much rather do this  
> in memory because then I never have to worry about what file system  
> and permissions the user executing the unit test has.

Install the ZenTest gem, then:

require 'test/zentest_assertions'

class TestBlah < Test::Unit::TestCase

   def test_my_stuff
     out, err = util_capture do the_thing end
     assert_equal "...", out.string
     assert_equal "...", err.string
   end
end

--
Poor workers blame their tools. Good workers build better tools. The
best workers get their tools to do the work for them. -- Syndicate Wars