Bug or optimisation

The following code:

struct A:
    fn __init__(inout self): ...
    
fn main(): 
    print(UnsafePointer.address_of(A()) == UnsafePointer.address_of(A())) # Prints True

struct A:
    fn __init__(inout self): ...
    
fn main(): 
    print(UnsafePointer.address_of(A()) == UnsafePointer.address_of(A())) # Prints True

indicates that both initialised empty struct has the same address. Are empty structs automatically "singleton"? Or is it a bug in the compiler? I am using Mac M1.

4 Replies

ivellapillil•2mo ago

mojo 24.4.0 (2cb57382)

Darkmatter•2mo ago

This looks like the compiler is seeing that the constructor is pure and deciding to de-duplicate it because you never modify either instance.

ivellapillil•2mo ago

Thanks! I was not aware compilers perform this optimisation.

Darkmatter•2mo ago

Compilers like to look for places to de-duplicate work because you can save a lot of processing if you can pull it off. In your case it’s trivial work, but if each constructor needed to calculate the first 200 primes or something like that, you would see noticeable performance increases.

Gaming

Programming

Bug or optimisation