CBFalconer wrote:
As I said, a debugging tool ...
IMNSHO a bad attitude. Not sure if it was Djikstra, Hoare, Knuth who had some pithy comments on the practice of removing runtime checks - comparisons with discarding life preservers when going to sea.
I didn't say that one should remove the check at runtime. I was merely pointing out that it doesn't make things so much better, since for the normal a runtime error message is not much better than some obscure behaviour, without a developer sitting next to him so debug the problem.
At any rate, the decision is up to the developer, and thus it should be possible to be very specific as to areas where the checks are removed.
This is planned, and it will be the easiest thing. Adding a compiler directive (of course, {$R+/-}, for compatibility with BP and other compilers since there are no official standards for it) is rather easy.
Nice, but wrong (e.g.: i: -4 .. 4)!
Pointing out that the check has to be performed at index time.
That's what I meant.
Again, this is the wrong perspective. The default should be to check things.
I think we're misunderstanding each other. I'm not arguing that range-checking is a bad thing. It's just a matter of priorities. We can't do all at once, and every bug and every missing feature is very urgent to someone. Of course, any contributions (code or money) might help to shift priorities ... :-)
Frank