The point tomalak16 and argv0 are trying to make is that just because an error is generated does not mean it is a bug - provided the error is handled. And likewise just because some code is bugged does not mean it will generate an error. The try/catch method is a legitimate and often practical method of writing valid, bug-free code. As someone who seems to code in Java I would expect you'd be very familiar with that practice. Although not quite as sophisticated, the :error method can be used to trap and handle unforeseeable problems in mIRC script. That isn't debugging.
Now to try and get back on track with the original suggestion:
A full blow-by-blow of each and every command and event and identifier's behaviour as it happens wouldn't be much good in most cases. You'd just have 100,000 lines of data and no idea where to start. What would make sense is the ability to make mIRC display a complete traceback of the commands called that lead to an error being generated. Beyond that though, all you really need to do with debugging is to add echo's (or use the -s switch in some newer commands) in key places in your code so you can accurately track where the code is going. This is more useful than something dumping all command flow at once because you only get data on the relevant parts of code.
