Document Sum::sum returns additive identities for []

[JakenHerman]

Because the neutral element of <fNN as iter::Sum> was changed to neg_zero, the documentation needed to be updated, as it was reporting inadequate information about what should be expected from the return.

Relevant Commit: 4908188
Relevant Pull Request: #129321

---

The referenced commit causes unintended side effects on presentation layer applications like using Tera templates, for example. I'm not sure what the motivation was behind the original change, but it seems like more discussion should be put into this issue and potentially have that change reverted.

[rustbot]

r? @Noratrieb

rustbot has assigned @Noratrieb.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

[workingjubilee]

@JakenHerman The reason is simple, as the negative zero is the true neutral element for floats: https://play.rust-lang.org/?version=stable&mode=debug&edition=2021&gist=9d21ada96adf24c4d3f9f2ef360c9630

fn main() {
    assert!(( 0.0f32 +  0.0).is_sign_positive());
    assert!((-0.0f32 +  0.0).is_sign_positive());
    assert!(( 0.0f32 + -0.0).is_sign_positive());
    assert!((-0.0f32 + -0.0).is_sign_negative());
}

[workingjubilee]

You would not expect the following to change the sign of the zero, would you?

[-0.0f32].into_iter().sum()

[workingjubilee]

Because the neutral element of <fNN as iter::Sum> was changed to neg_zero, the documentation needed to be updated, as it was reporting incorrect information about what should be expected from the return.

And I must note, the documentation was not technically incorrect.

...You can say it was uselessly correct, of course.

[workingjubilee]

I'm curious, can you describe the effects on the Tera applications?

[JakenHerman]

I'm curious, can you describe the effects on the Tera applications?

@workingjubilee Sure, I'll give a simple example... Consider an HTML/PDF report that sums up financial transactions then displays the total value of the summed transactions. If there are no transactions, the report would show that you spent $-0.00, whereas it used to be $0.00.

That's a very simple example (and not the one I ran into), but it is the easiest I could come up with for illustrative purposes. I suppose the real solution on the template side would be to abs the value before presenting it, but it is slightly jarring to have numerous reports mysteriously and suddenly reporting negative 0 values in production.

I committed your suggestion, as I think you summed up the expected return value better than I did. Thanks for your input.

[workingjubilee]

That is slightly unfortunate, yes. I guess that would have worked out fine if you just always had a positive value in the summation such that even if it reached zero, it would have been positive.

@bors r+ rollup

[bors]

📌 Commit 1093779 has been approved by workingjubilee

It is now in the queue for this repository.

[workingjubilee]

just a squash

@bors r+ rollup

[bors]

📌 Commit 4457f44 has been approved by workingjubilee

It is now in the queue for this repository.