book: make the test succeed by avoiding using non-existent function

[tshepang]

No description provided.

[rust-highfive]

r? @steveklabnik

(rust_highfive has picked a reviewer for you, use r? to override)

[bluss]

I think it should be identical to the preceding it_works, it's an example in the same context, so this change is unnecessary.

[tshepang]

But the test fails at the moment

[Manishearth]

Plop in a definition of it_works commented out with #es, then. Testing is supposed to test something, in this case you're testing add_two.

[tshepang]

I will add the add_two function

[tshepang]

fixed

[Manishearth]

@bors r+ rollup

[bors]

📌 Commit d30ebd1 has been approved by Manishearth

[bors]

💔 Test failed - auto-mac-64-nopt-t

[Manishearth]

---- The__code_ignore__code__attribute_0 stdout ----
    <anon>:2:5: 4:6 error: visibility has no effect inside functions [E0447]
<anon>:2     pub fn add_two(a: i32) -> i32 {
<anon>:3         a + 2
<anon>:4     }
error: aborting due to previous error

[bors]

☔ The latest upstream changes (presumably #29458) made this pull request unmergeable. Please resolve the merge conflicts.

[tshepang]

test ignored, as is the case with one before it

[Manishearth]

We don't want to ignore the test.

We want to not mark the function as pub.

Actually, I thought this PR was to solve a doctest failure/ignore, which is why I suggested using a #'d out function. But the doctest passes, since doctests don't even compile #[test] bits.

It's a pretty common pattern in rust docs to use snippets that need parts from previous snippets. Because of this, I'm not sure if this PR is really necessary.

[tshepang]

I think it's marked pub in order to demo that the function is to be used externally... it's a library. Note that pub is used earlier in the file for the same function, and that case has been ignored as well.

I will close this regardless, given your last point.