After running automated tests, check for runtime problems

[mciul]

I was thinking about adding this to Automated Testing, but I think it would slow the tests down a lot. It would probably be much faster to run it afterwards.

What I'd like to do is scan the testtranscript file after finishing all tests for a few patterns that would indicate bugs. I'd like to look for "*** Run-time Problem" and "You can't reach into"... but there may be more.