cajeta.concurrent.AtomicInt32 — atomic 32-bit signed integer: a lock-free int32 cell that multiple threads can read, write, and update without tearing or data races. The class owns a heap-allocated cell that the compiler emits inline LLVM atomic instructions against. The plain methods (load, store, fetchAdd, compareAndSet) use sequentially consistent ordering; to opt into a weaker ordering, call the variant whose name encodes it — there is no ordering parameter. The 64-bit counterpart is AtomicInt64.

AtomicInt32 counter = heap AtomicInt32(0);
int32 prior = counter.fetchAdd(1);          // 0; cell now holds 1
boolean won = counter.compareAndSet(1, 2);  // true; cell now holds 2
int32 n = counter.load();                   // 2

Methods

Signature
`AtomicInt32(int32 initial)` ⚑	Creates an atomic cell seeded with `initial`
`int32 load()`	Atomic load (seq_cst)
`void store(int32 v)`	Atomic store (seq_cst)
`int32 fetchAdd(int32 delta)`	Atomic fetch-and-add; returns the prior value
`boolean compareAndSet(int32 expected, int32 desired)`	Atomic compare-and-set
`int32 loadRelaxed()`	Load with `RELAXED` ordering — atomic, no cross-thread synchronization
`int32 loadAcquire()`	Load with `ACQUIRE` ordering — pairs with `storeRelease`
`void storeRelaxed(int32 v)`	Store with `RELAXED` ordering — no release semantics
`void storeRelease(int32 v)`	Store with `RELEASE` ordering — publishes prior writes to a paired `loadAcquire`
`int32 fetchAddRelaxed(int32 delta)`	`RELAXED` fetch-and-add; returns the value held before the add
`boolean compareAndSetAcquire(int32 expected, int32 desired)`	Compare-and-set with `ACQUIRE` ordering on success and failure

⚑ = @EntryPoint

AtomicInt32

Methods

See also