Fix any precision issues caused by the usage of double (arithmetic)

[shrnkld]

Scenario: The XBridge protocol currently uses a mix of (float), double, int64_t and uint64_t and as such we are seeing precision issues eg when partial orders are created (ref: https://github.com/blocknetdx/blocknet/issues/570)

Proposed solution: Replace all uses of (float), double, int64_t and uint64_t with a 'signed integer' type with 256 bits for all amounts (int256_t), enabling XBridge support for tokens with 18 decimal places (being similar to what ETH uses (uint256)). Further, using a signed type, as opposed to unsigned, would also prevent underflows in the event of a negative amount that results from a calculation.

To do:

• [ ] remove/fix any precision issues caused by the usage of double (arithmetic) and replace with int256_t throughout the XBridge code

---

QA and Acceptance info to follow...

[atcsecure]

@shrnkld LGTM

[shrnkld]

WIP and ideas: step 1: replace all uint64_t with int64_t and test. ~~step 2: replace all float with double and test calculations and performance (increasing precision points for erc20 support)~~ cannot find any instances of float related to xbridge in the src. step 3:replace all int64_t with int256_t and retest