SUnit TestRunner is a nice looking app but it lacks certain key 
functionality:

1. the ability to sort results.
2. the ability to compare results of one test with those of another.
3. To quickly go from a result to the method code that triggered it.

I ran into these because I was going through the sources like a madman 
(I've already succeded in taking a kilobyte off of the binary.) and 
making all kinds of changes. I was able to use GDB to fix some segfaults 
in my code.

I'm beginning to see alot of strangeness in the C compiler that I will 
look at next. I will also explore how recient enhancements to the C 
language can be incorporated into the code generator.

I was using TestRunner as a comprehensive test of my changes. My 
workhorse VM produces 37 failures and  17 errors.

My test VM produces 39 failures and 17 errors...

Some of these are stupid, such as querying wheather a certain class has 
a comment... It's not a terrably useful bit of functionality anyway, 
there's no reason to give me 15  errors about it...

Others seemed to be a result of bad test code. For example, a test of 
the time class seemed to expect that a query of the hour from a class 
set to 1:33 PM ( 13:33) to return 2... I'm not sure wheather it actually 
_SHOULD_ have been 13:33 but the test looked broken.

I have no idea which of the two new failures are a result of my changes 
because there is no way to efficiently find the two tests that are _not_ 
in the refferance list.

I won't be able to continue my hacking untill I correct whatever mistake 
I made...