Struct RepeatCounters 

pub struct RepeatCounters { /* private fields */ }

Live decrementing counters for every (ProcessId, term_index) pair currently mid-repeat(N)@(...) wait. Mutex-guarded so Engine (which reaches this only through a shared &self inside EngineNotifySink) can call Self::on_term_fired/Self::clear from a &self-only context — the same pattern every other sched shared-mutable-state field already uses (NbaLog, RegionQueues, …).

Implementations

impl RepeatCounters

pub fn new() -> Self

pub fn on_term_fired( &self, proc: ProcessId, term_index: usize, declared_total: u64, ) -> RepeatOutcome

Looks up (proc, term_index)’s live remaining count, initializing it to declared_total on first sight (the first firing after the process suspended), decrements by 1, and reports the outcome. At 0 post-decrement the entry is removed — so a later, unrelated repeat(N)@(...) wait re-installed at the same term_index starts fresh rather than inheriting a stale zero (see clear_removes_only_the_named_processs_entries for the companion “no stale residue after an explicit clear” proof).

pub fn clear(&self, proc: ProcessId)

Removes every counter entry belonging to proc, regardless of term_index — called from the same place SensitivityTable:: remove_dynamic already is (sched::control’s dispatch_ready_set), so a process that resumes via any path (an ordinary event match, Disabled, …) never leaves a stale partial count behind for a future, unrelated dynamic wait reusing the same term_index.

Trait Implementations

impl Default for RepeatCounters

fn default() -> Self

Returns the “default value” for a type.