Congratulations on discovering some Ruby internals. There’s a certain peculiar joy that comes from hacking around the innards of your favorite runtime, seeing how it all works under the hood.

If Ruby is anything like Lisp (from whose tradition the names “fixnum” and “bignum” come), then the object id employs tag bits to determine the type of an object; and the rightmost bit is such a tag bit to represent small integer constants (fixnums). Various other tag bits would mark an object as true, false, nil, or a character constant.

Furthermore, odds are pretty good that once you mask the tag bits that identify it as an aggregate object, the object ids of various other objects in the Ruby system indeed become pointers; because such objects are usually aligned to machine-word boundaries, it becomes useful to assume that the last four bits of the pointed-to address are zero, and use those bits as tag bits.

Again, I don’t know that much about Ruby; I’m only extrapolating from what I’ve learned from Lisp. Nevertheless I hope I’ve opened up new avenues of exploration.

Very interesting post. I would like to link back to it….