Ron Jeffries wrote:

> Hi, I'm doing some stuff with RubyUnit and have some questions:
> 
> 0. Is there a writeup somewhere about how RubyUnit works? I'm finding
> it a bit opaque, between its zillion files and my inexperience with
> Ruby.
> 


Seems like I saw one somewhere.  I'll browse around for it and post it 
later if nobody beats me to it.  (Oh, are you looking for internals or 
normal usage information?)


> 1. Is there a way to tell after running the suite, whether there have
> been errors? I'm playing with backing up the file every time the tests
> run at 100%



Take a look at runit/cui/testrunner.rb.  That's where this kind of thing 
would happen.  To me, it looks like you would have to write your own 
TestRunner to do what you want (should be trivial), although if you 
wanted to hack it, you could pass in your own mock io object into the 
TestRunner constructor and then parse the output.


> 
> 2. Chet and I wrote an addition to RubyUnit to record all the
> subclasses of TestCase to create a suite and run it. Was there already
> a way to do that? Would the thing we did be useful?
> 


You mean you automated the creation of a test suite by discovering all 
of the TestCases in the system?  Sounds like a problem people are 
solving over and over again in their own way.  I don't think there's a 
builtin way to do this.  At least with Ruby the solutions are generally 
less nauseating than with JUnit. :)


Chad