Allocate memory for HashMap with given capacity Initializes all control bytes to EMPTY

Create a new empty HashMap with default capacity

Create a HashMap with a specific initial capacity Capacity will be rounded up to next power of 2

Get ctrl pointer (unwrapped for internal use)

Get data pointer (unwrapped for internal use)

Find bucket index for a given key using quadratic probing Returns Some(index) if key is found, None otherwise

Find first available bucket (EMPTY or DELETED) for insertion Returns bucket index using quadratic probing

Check if HashMap needs resizing based on load factor

Resize and rehash the HashMap to a new capacity

Insert or update a key-value pair Returns Ok(Some(old_value)) if key existed, Ok(None) if new key

Get a value by key Returns Some(value) if found, None otherwise

Check if a key exists in the map

Remove a key-value pair by key Returns Some(value) if the key was found and removed, None otherwise

Get the number of key-value pairs in the map

Check if the map is empty

Clear all key-value pairs Resets all control bytes to EMPTY

RAII destructor - automatically called when HashMap goes out of scope

Pointer iterator — yields *(Bucket(K, V)) for each occupied bucket. For privileged/unsafe internal use; value iteration goes through into_iter().

Cycle-GC tracing. Buckets live in a malloc'd open-addressing buffer the compiler's auto-derived field walk cannot reach, so trace each FULL slot (its control byte is neither EMPTY nor DELETED). tracer.visit takes the Bucket(K, V) slot POINTER; the per-value traversal descends inline through the value struct, tracing BOTH the key and the value, WITHOUT touching any reference count (a by-value managed handle would be dup'd then dropped, freeing a live element mid-collection). Mirrors the ArrayList Trace impl + the SwissTable slot scan.