Document overflow analysis for magnitude weight factor as fraction

src/gridcoin/voting/registry.cpp

@@ -568,6 +568,11 @@ std::optional<CAmount> PollReference::GetActiveVoteWeight(const PollResultOption
             const arith_uint256 scaled_pool_magnitude,
             const arith_uint256 scaled_network_magnitude)
     {
+        // Unlike the complementary calculation in EnableMagnitudeWeight for vote weights, this calculation is not
+        // subject to overflow because it uses 256 bit integers. In the end it the RESULT is cast to a 64 bit integer, but
+        // by then all of the multiplication and division are done. m_magnitude_weight_factor as a fraction something that is
+        // not expected to stray any farther than the interval [1/10, 1/1], where the current default value of 100 / 567 is
+        // in that interval. So we should never have an overflow problem here.
         active_vote_weight_tally += net_weight + money_supply * (scaled_network_magnitude - scaled_pool_magnitude)
                                                               * arith_uint256(m_magnitude_weight_factor.GetNumerator())
                                                               / arith_uint256(m_magnitude_weight_factor.GetDenominator())

src/gridcoin/voting/result.cpp

@@ -797,6 +797,19 @@ class VoteCounter
         }
 
         // Use integer arithmetic to avoid floating-point discrepancies:
+
+        // Overflow analysis:
+        //
+        // Current supply as of block 3404576 = 493424753.
+        // network magnitude = 115000
+        // max value for uint64_t = 2^64 - 1
+        //
+        // 2^64 - 1 >= (493424753 / 115000) * mwf numerator
+        //
+        // This means mwf numerator <= (2^64 - 1) * (115000 / 493424753)
+        //                         ~<= 4.3E15
+        // Even if money supply is approximately doubled we can safely handle 2E15, which means 15 decimal
+        // places in the fraction numerator
         m_magnitude_factor = supply / total_mag
                              * (uint64_t) m_poll.m_magnitude_weight_factor.GetNumerator()
                              / (uint64_t) m_poll.m_magnitude_weight_factor.GetDenominator();