Function explore_balance_states

fn explore_balance_states<T, M, G, F, H>(
    users: &[M::User],
    deposits: &[T::Asset],
    adjustments: &[T::Asset],
    subjects: &[T::Subject],
    max_depth: u32,
    allowed_bps: u32,
    allowed_diff: u32,
    overrides: Option<BalanceTraps<G, F>>,
    hasher: Option<H>,
    results: &mut BalanceModelResults<T, M>,
)
where
    T: LazyBalanceMarker + Clone,
    M: ManualBalanceModel<T> + BalanceStateHasher<T, M> + BalanceGuards<T, M> + Clone,
    T::Id: Default,
    T::Asset: From<T::Asset> + Into<T::Asset>,
    G: Fn(&BalanceState<T, M>, &BalanceOp<T, M>) -> bool,
    F: Fn(&BalanceState<T, M>, &BalanceOp<T, M>) -> bool,
    H: Fn(&BalanceState<T, M>) -> u64,

Core state exploration engine for LazyBalance.

Performs breadth-first exploration over operation sequences, starting from an initial empty state.

-> [Deposit]   -> [Deposit, Mint] -> [Deposit, Mint, Withdraw]
-> [Mint]      -> [Mint, Reap]
-> [Reap]      -> ...
-> [Withdraw]  -> ...
-> [Drain]     -> ...

Exploration proceeds level by level:

• all sequences of length n are explored before n+1
• each step extends the current trace by one BalanceOp
• guards / traps / flow determine whether a branch continues

At each step:

• checks BalanceGuards::invariant on the current state
• generates candidate operations (vector of BalanceOp)
• filters using:
  • guard_{op}(...) || trap(...)
  • guard_flow_{op}(...) && trap_flow(...)

For each valid operation:

• executes lazy model
• applies manual model
• updates state (receipts, trace)
• evaluates withdrawal drift (compare_balances_values)

Uses:

• visited set with BalanceStateHasher to avoid revisiting states
• optional custom hasher for additional deduplication

Records:

• errors (failures)
• exits (drift beyond tolerance)
• successful paths

This function defines the full execution semantics for:

• guarded exploration
• trap-based overrides
• state-space traversal