Implement proper RBF logic satisfying rule 4

Removes `min_fee` constraint in favour of proper rule 4
handling (min_fee was pretty useless).

See for rule 4:
https://github.com/bitcoin/bitcoin/blob/master/doc/policy/mempool-replacements.md#current-replace-by-fee-policy

LowestFee BnB metric updated to bound correctly for this.

A few other coin_selector API changes

src/metrics/lowest_fee.rs

@@ -52,12 +52,13 @@ impl BnbMetric for LowestFee {
 
             let drain_value = cs.drain_value(self.target, self.change_policy);
 
+            // I think this whole if statement could be removed if we made this metric decide the change policy
             if let Some(drain_value) = drain_value {
-                // it's possible that adding another input might reduce your long term fee if it gets
-                // rid of an expensive change output. Our strategy is to take the lowest sat per
-                // value candidate we have and use it as a benchmark. We imagine it has the perfect
-                // value (but the same sats per weight unit) to get rid of the change output by
-                // adding negative effective value (i.e. perfectly reducing excess to the point
+                // it's possible that adding another input might reduce your long term fee if it
+                // gets rid of an expensive change output. Our strategy is to take the lowest sat
+                // per value candidate we have and use it as a benchmark. We imagine it has the
+                // perfect value (but the same sats per weight unit) to get rid of the change output
+                // by adding negative effective value (i.e. perfectly reducing excess to the point
                 // where change wouldn't be added according to the policy).
                 //
                 // TODO: This metric could be tighter by being more complicated but this seems to be
@@ -91,6 +92,19 @@ impl BnbMetric for LowestFee {
                         }
                     }
                 }
+            } else {
+                // Ok but maybe adding change could improve the metric?
+                let cost_of_adding_change = self
+                    .change_policy
+                    .drain_weights
+                    .waste(self.target.fee.rate, self.long_term_feerate);
+                let cost_of_no_change = cs.excess(self.target, Drain::none());
+
+                let best_score_with_change =
+                    Ordf32(current_score.0 - cost_of_no_change as f32 + cost_of_adding_change);
+                if best_score_with_change < current_score {
+                    return Some(best_score_with_change);
+                }
             }
 
             Some(current_score)

tests/lowest_fee.proptest-regressions

@@ -8,3 +8,4 @@ cc 9c841bb85574de2412972df187e7ebd01f7a06a178a67f4d99c0178dd578ac34 # shrinks to
 cc e30499b75a1846759fc9ffd7ee558b08a4795598cf7919f6be6d62cc7a79d4cb # shrinks to n_candidates = 25, target_value = 56697, base_weight = 621, min_fee = 0, feerate = 9.417939, feerate_lt_diff = 0.0, drain_weight = 100, drain_spend_weight = 1, drain_dust = 100
 cc c580ee452624915fc710d5fe724c7a9347472ccd178f66c9db9479cfc6168f48 # shrinks to n_candidates = 25, target_value = 488278, base_weight = 242, min_fee = 0, feerate = 6.952743, feerate_lt_diff = 0.0, drain_weight = 100, drain_spend_weight = 1, drain_dust = 100
 cc 850e0115aeeb7ed50235fdb4b5183eb5bf8309a45874dc261e3d3fd2d8c84660 # shrinks to n_candidates = 8, target_value = 444541, base_weight = 253, min_fee = 0, feerate = 55.98181, feerate_lt_diff = 36.874306, drain_weight = 490, drain_spend_weight = 1779, drain_dust = 100
+cc ca9831dfe4ad27fc705ae4af201b9b50739404bda0e1e130072858634f8d25d9 # added by llfourn from ci: has a case where the lowest fee metric would benefit by adding change