On 6 Aug 2002, Ken Causey wrote:

> I think I must be missing something.  I'm finding examining errors and
> failure with SUnit to be very painful.  If I click on the failure or
> error in the appropriate section of the TestRunner (and by the way, if
> said error is already selected it has to be clicked again to reselect
> before anything happens) I get a debug window.  This is great and fine
> but I have to step through, over, and into several times before I get
> anywhere near the problem and even then I have to step past it several
> times before I get an error which is really quite unclear because it
> only appears replacing the debug window title well after the code that
> the error applies to has left the debug window.

This shouldn't be the case for errors - the debug window should be right
where the error is.  I used to do exactly what you describe for failures,
but if you hit Proceed instead it'll pop up another debugger when it hits
the failing assertion.  This is much easier.

I agree that the TestRunner could benefit from showing the error/failure
line, though.  I just ran into this about 10 minutes ago with a
time-sensitive test that only sometimes failed, and *always* passed when
the debugger or transcript introduced some extra delay...