I'm trying to translate some pseudocode that starts indexing at -1. Will the following code work like I expect it to?
Code:int main() {
int* a = new int[10];
a = &a[1];
a[-1] = 4;
delete (a - 1);
}
Printable View
I'm trying to translate some pseudocode that starts indexing at -1. Will the following code work like I expect it to?
Code:int main() {
int* a = new int[10];
a = &a[1];
a[-1] = 4;
delete (a - 1);
}
Assuming you don't use the same letter twice. (I.e., you could do int *b = &a[1] and go from there.)
(ETA: I should say, you might get warnings from your compiler (I think for example gcc warns on indices it can guarantee are always negative), but since your pointer, if used correctly, will always refer to valid memory everything is defined.)
More over, why would you want to do that? There might be data there that you don't want to mess with and could end up with errors or corruption.
Look again at what he's doing here:Quote:
Originally Posted by computerquip
Now his array indexes go from -1 to 8 instead of 0 to 9.Code:a = &a[1];
Don't know why anyone would want to do that though.
If you are implementing a lookup table that maps values in the range [x, y], then you don't want to subtract x from each lookup -- it's more efficient to just adjust the base pointer such that the x'th element coincides with the zero'th element of the actual array.Quote:
Originally Posted by cpjust
Of course, merely possessing an out-of-bounds pointer is technically undefined, but not something I'd worry about in practice.