Fix 8167 (BigInt identity constructor) and 9061 (BigInt bitwise ops)

[Biotronic]

No description provided.

[monarchdodra]

Fix 8167 (identity constructor) and 9061 (bitwise ops): LGTM.

Bitwise op behavior: This looks good to me, and is in line with the discussion. However, it would be nice for it to also be publicly documented, so as to avoid any future confusion: eg something like "Bitwise operators behave as if the BigInt was implemented as an infinite length 2's complement number".

[monarchdodra]

Nitpciks: The current implementations with integrals looks correct. However, converting the integral to a BigInt, followed by doing a binary operation of the length of the biggest of both BigInts seems inefficient, especially for operators '|' and '&'.

It's correct, but I think it leaves room for a future improvement.

Anyways, anybody else want to proof read the math? It looks legit to me, but a second opinion never hurts.

[Biotronic]

Yeah, I've also been bothered by the extra allocations and stuff, but with no comments I didn't quite bother. Should be better now.

[monarchdodra]

Logical operations (|, &, ^, ~) are supported

I thought "logical" was ||, &&, !, ==, < etc? Aren't those "binary" operations?

---

I searched wikipedia, and according to them, the correct terminology would be: "Bitwise operation" (|, &, ^, ~) and "Boolean expression" (||, &&, !, ==, <, ...).

Other than this final nitpick, I'm ready to merge.

[Biotronic]

I used 'logical' because that's what the original said, and I'm willing to argue both names are correct, but I prefer the name 'bitwise' myself. :p

[monarchdodra]

And merged! TYVM.

[Biotronic]

Wheee!!!

... sorry.