I dont know that I agree with this. I suppose you may want to walk the parse tree in other circumstances, but I dont see why simple printing out the exact code that failed would be insufficient, and I think that would be reasonably doable with what we have already. Sure, it may not work on some far out there cases, but then again I wonder what the use of such an insane assertion would be. Anyway, im not knocking the idea if it comes along, but until its here im not going to spend much time in anticipation. As I said IMO, %99.99 of the things that most macro langauges do can be done as well or better by perl itself. All that a defined macro language does is provide consistancy. Which in of itself is probably a good thing (or maybe a bad thing if you read my other node), but my original point was merely that a LFSP need not have macros within it.
<Elian> And I do take a kind of perverse pleasure in having an OO assembly language...