feat: replace HashMap with HashTable for WeakMap

[shruti2522]

As discussed in #13, HashTable is a better approach for our WeakMap

• HashTable prevents any redundnat key storage since HashTable<(usize,Ptr)> stores address once, while HashMap<usize, Ptr> stores it twice
• HashTable::insert_unique avoids the lookup HashMap::insert always performs.
• HashTable::find_entry returns an OccupiedEntry so .remove() extracts the value at that same slot without a second lookup.
• HashTable::retain(|&mut T|) operates on the entry tuple directly avoiding separate key/value refs

[nekevss]

Nice! Pretty big fan of the change.

Can you add a unit test for what happens when somebody tries to insert a Gc<K> for the key on two entries into a WeakMap? Do we handle that correctly?

Mostly I'm skeptical about whether key_addr is actually unique ... so we should test it! 😄

[shruti2522]

added three new tests to verify key_addr uniqueness:
two_distinct_keys_wm : two different keys in the same map, to confirm they don’t collide and their values stay separate.
two_maps_same_key_wm : same key used in two different maps, ensuring each map works independently.
drop_map_with_live_key_wm : dropping a WeakMap while its key is still alive, to check that the collector safely ignores the removed map