Continuing the theme of short notes on my current computing project, and how I don’t have time to work on it, I can now report that I’ve achieved, in this iteration, what I’d achieved in the previous, which is to say the current code base passes the (slightly modified) auto-regression tests I’ve been writing. Because I used the supplied EBNF to build the parser, it also means I’ve surpassed (theoretically – not verified) the parsing capability of the previous iteration. It’d be nice if I could find a publicly available test suite.
The debugging approach described in my previous missive has also worked well.
The auto-regression test suite, however, needs a complete re-think. The lesson there is one all coders should know: test suites need to be thought out as well as what they are testing. And I didn’t do that. Bad Hue, go lay down by your dish.
Next steps: Continue working on the “well-formedness” and validation parts of the processor; expand the auto-regression test suite.
All this for a compiler which may never be completed, nor ever be widely used. Tilting at windmills entered a new age when open source software started up, all those years ago.
