From: Holden Glova <dsafari@...>
Date: 2000-12-30T15:44:14+09:00
Subject: [ruby-talk:8343] Re: open3 woes [sort of long]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Sat, 30 Dec 2000 19:29, I think Dave Thomas wrote:
> Holden Glova <dsafari@paradise.net.nz> writes:
> > Ok, granted my test is set up to fail, but I wrote this simply to
> > make sure that the correct data was coming to stdout. Why does it
> > seem to append the entire contents of the file I pass to
> > RuntimeInterpreter#invoke to itself?
>
> Well, I can't reproduce it here:
>
>
> useless.rb
>
>      class Useless
>        def meth
>          puts "Useless method in a useless class"
>        end
>      end
>      u = Useless.new()
>      u.meth()
>
>
> t.rb
>
>      require 'open3'
>
>      class RuntimeInterpreter
>        def invoke(file)
>          streams = Open3.popen3("ruby -w #{file}")
>          streams.at(1).each() { |line|
>            puts line
>          }
>       end
>      end
>
>      require 'runit/testcase'
>
>      class Test_RuntimeInterpreter < RUNIT::TestCase
>        def testInvoke
>          interpreter = RuntimeInterpreter.new()
>          result = interpreter.invoke("../useless.rb")
>          assert(result == "Useless method in a useless class")
>        end
>      end
>
>      if __FILE__ == $0
>        require 'runit/cui/testrunner'
>        suite = RUNIT::TestSuite.new
>
>        suite.add_test(Test_RuntimeInterpreter.suite)
>        RUNIT::CUI::TestRunner.run(suite)
>      end
>
>
> dave[~/tmp 0:26:40] ruby t.rb
> Test_RuntimeInterpreter#testInvoke F.
> Time: 0.024983
> FAILURES!!!
> Test Results:
>  Run: 1/1(1 asserts) Failures: 1 Errors: 0
> Failures: 1
> t.rb:21:in `testInvoke': The condition is <false:FalseClass>
> (RUNIT::AssertionFailedError) from t.rb:30
> dave[~/tmp 0:26:44] cat useless.rb
> class Useless
>         def meth
>                 puts "Useless method in a useless class"
>         end
> end
> u = Useless.new()
> u.meth()
>
>
> I ran it several times with the same result.
>
> Are you using rubywin?

Nope this is on a Debian GNU Linux. Strange that your output doesn't produe 
the String from Useless#meth. That is what was happening to me when i was 
using backquotes `ruby -w #{file} 2>&1`. I'm totally stumped. Would it make 
any difference the fact that  my testcase is in a separate file? I noticed 
that for some reason this makes a difference for me when i'm using backquotes 
- - I don't know why, but my it will work just fine if I use backquotes and the 
test is in the same source file, but move the test into a separate source 
file and a separate directory and it fails. 

I'm already bald, yet I find myself constantly reaching for hair to pull out. 
Please help a desparate bald man :)

FWIW: Thanks for taking the time to try and simulate my problem.
- -- 
Signed,
Holden Glova
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.4 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE6TYOczrxa+Gy/b/4RAsEhAJ4qAQxRT0BdRTIY3ncgArcYiWFklwCfcdUR
FxWp8Xt+z1QlO7F5CvGnHeI=
=s9v2
-----END PGP SIGNATURE-----