On Sunday, 7 February 2021 at 21:55:34 UTC, Adam D. Ruppe wrote:
On Sunday, 7 February 2021 at 21:40:12 UTC, Jack wrote:
I think it would be fine except it assumes the number of items
of the array to doesn't grow, it rather overwritten new
elements
from docs:
"Use this only when it is certain there are no elements in use
beyond the array in the memory block. If there are, those
elements will be overwritten by appending to this array."
That's referring to a case like this:
int[] a = [1, 2, 3, 4, 5];
int[] b = a[0 .. 3];
Normally, if you were to do
b ~= 6;
it would allocate a new array for b, leaving a alone.
a == [1, 2, 3, 4, 5];
b == [1, 2, 3, 6];
a.ptr != b.ptr because b got reallocated.
If you assumeSafeAppended there though, the b ~= 6 would reuse
the remainder of the block.
b.assumeSafeAppend();
b ~= 6;
a == [1, 2, 3, 6, 5];
b == [1, 2, 3, 6];
a.ptr == b.ptr; // assumeSafeAppend meant no reallocation
So the "elements in use beyond the array in the memory block"
are referring to the variable a in the example still having [4,
5] at the end. Since that 4 gets overwritten by the 6, if you
weren't prepared for this, it can be a surprising bug.
The docs also say this is "undefined behavior" simply because
the 4 won't *always* get overwritten by the 6. well, in this
case the 4 is always overwritten by the 6, but say I wanted to
append [6, 7, 8], then it might reallocate anyway because the
memory block is only big enough for two new elements, not
three. If that happened, that 4 would stay put.
So the array is allowed to grow as much as it wants, just in
some cases it will overwrite existing data and in other cases
it will realloc a new slice to make room.
got it, thank you very much!
