Closed oscbyspro closed 3 weeks ago
I believe pointer-count is a more appropriate representation because systems integers (and other fixed-size types) know their count at compile time, but not their end pointer. In other words, the pointer-pointer is necessarily derived from a pointer-count in such cases. Some algorithms may benefit from a pointer-pointer, but it would be better to derive it from a pointer-count than to round-trip it. I may consider a separate pointer-pointer iterator model, however.
It should be straightforward to rewrite
DataInt.Body
as a pair of pointers rather than a pointer and a count. I don't know whether it's better or worse, but I'll try it out when I get the time to do so. I'll update this post when I have the results.