In JS, you can't control the hashing, but you don't have to.
Two things are the same if they're equal. The hash is not part of the definition, it's just an implementation detail. Under the covers, two different objects may have the same hash, but they're still different, and the implementation has to deal with that magically (e.g., by using a chaining hash table).
Also, the keys of an object are always strings—the interpreter will stringify the values inside the hash constructor, inside the [], or after the ., rather than storing the actual values, which means that this rarely comes up in the first place.
To give some examples:
function X() {}
x = new X()
y = new Y()
h = {x: 2, y: 3} // h has 2 members, named "x" and "y"
a = (x, y)
b = (x, y)
h[a] = 4
h[b] = 5 // h has 3 members, named "x", "y", and "[object Object]"
Put in Python terms, it's as if dict called __repr__ on keys instead of __hash__ (although this isn't quite 100% accurate), which means you can provide a custom toString method to control equal-ness of different instances of your class.