compiler_add_o() uses an heuristic to compare and merge duplicated constants. It has special cases for float and complex numbers, but it's not designed to handle more types.
Funny, I had the same isue last week why I added support for tuple and frozenset "constants" in AST. I had to explicitly support these types in compiler_add_o().
I see two options:
(1) share code between compiler_add_o() and code_richcompare() to ensure that 1 and 1.0 constants are not seen as equal
(2) modify compiler_add_o() to never merge code objects, always considere them as unequal
For (2), there is a minor technical issue: you have to generate an unique key for the dictionary.
I prefer option (1) for consistency.