rustdoc: move manual "extern crate" statements outside automatic "fn main"s in doctests
Gated on https://github.com/rust-lang/rust/pull/48095 - I based the branch atop that so i could show off the change in one of its tests, the actual change in this PR is just the last commit
There are a handful of unfortunate assumptions in the way rustdoc processes `extern crate` statements in doctests:
1. In the absence of an `extern crate` statement in the test, if the test also uses the local crate name, it will automatically insert an `extern crate cratename;` statement into the test.
2. If the doctest *does* include an `extern crate` statement, rustdoc will not automatically insert one, on the assumption that doing so would introduce a duplicate import.
3. If a doctest does not have the substring `fn main` outside a comment, rustdoc will wrap the whole doctest in a generated `fn main` so it can be compiled.
In short, whenever you write a doctest like this...
```rust
//! extern crate my_crate;
//! my_crate::some_cool_thing();
```
...rustdoc will turn it into (something like) this:
```rust
fn main() {
extern crate my_crate;
my_crate::some_cool_thing();
}
```
This creates issues when compiled, because now `my_crate` isn't even properly in scope! This forces people who want to have multiple crates in their doctests (or an explicit `extern crate` statement) to also manually include their own `fn main`, so rustdoc doesn't put their imports in the wrong place.
This PR just taps into another processing step rustdoc does to doctests: Whenever you add an `#![inner_attribute]` to the beginning of a doctest, rustdoc will actually splice those out and put it before the generated `fn main`. Now, we can just do the same with `extern crate`s at the beginning, too, and get a much nicer experience.
Now, the above example will be converted into this:
```rust
extern crate my_crate;
fn main() {
my_crate::some_cool_thing();
}
```
Generate documentation for auto-trait impls
A new section is added to both both struct and trait doc pages.
On struct/enum pages, a new 'Auto Trait Implementations' section displays any synthetic implementations for auto traits. Currently, this is only done for Send and Sync.
On trait pages, a new 'Auto Implementors' section displays all types which automatically implement the trait. Effectively, this is a list of all public types in the standard library.
Synthesized impls for a particular auto trait ('synthetic impls') take generic bounds into account. For example, a type
```rust
struct Foo<T>(T)
```
will have 'impl<T> Send for Foo<T> where T: Send' generated for it.
Manual implementations of auto traits are also taken into account. If we have
the following types:
```rust
struct Foo<T>(T)
struct Wrapper<T>(Foo<T>)
unsafe impl<T> Send for Wrapper<T>' // pretend that Wrapper<T> makes this sound somehow
```
Then Wrapper will have the following impl generated:
```rust
impl<T> Send for Wrapper<T>
```
reflecting the fact that 'T: Send' need not hold for 'Wrapper<T>: Send' to hold
Lifetimes, HRTBS, and projections (e.g. '<T as Iterator>::Item') are taken into account by synthetic impls:
However, if a type can *never* implement a particular auto trait (e.g. `struct MyStruct<T>(*const T)`), then a negative impl will be generated (in this case, `impl<T> !Send for MyStruct<T>`)
All of this means that a user should be able to copy-paste a syntheticimpl into their code, without any observable changes in behavior (assuming the rest of the program remains unchanged).
Current document takes 2^4, which is equal to 4^2.
This example is not very helpful for those unfamiliar with math words in English and thus rely on example codes.
A new section is added to both both struct and trait doc pages.
On struct/enum pages, a new 'Auto Trait Implementations' section displays any
synthetic implementations for auto traits. Currently, this is only done
for Send and Sync.
On trait pages, a new 'Auto Implementors' section displays all types
which automatically implement the trait. Effectively, this is a list of
all public types in the standard library.
Synthesized impls for a particular auto trait ('synthetic impls') take
into account generic bounds. For example, a type 'struct Foo<T>(T)' will
have 'impl<T> Send for Foo<T> where T: Send' generated for it.
Manual implementations of auto traits are also taken into account. If we have
the following types:
'struct Foo<T>(T)'
'struct Wrapper<T>(Foo<T>)'
'unsafe impl<T> Send for Wrapper<T>' // pretend that Wrapper<T> makes
this sound somehow
Then Wrapper will have the following impl generated:
'impl<T> Send for Wrapper<T>'
reflecting the fact that 'T: Send' need not hold for 'Wrapper<T>: Send'
to hold
Lifetimes, HRTBS, and projections (e.g. '<T as Iterator>::Item') are
taken into account by synthetic impls
However, if a type can *never* implement a particular auto trait
(e.g. 'struct MyStruct<T>(*const T)'), then a negative impl will be
generated (in this case, 'impl<T> !Send for MyStruct<T>')
All of this means that a user should be able to copy-paste a synthetic
impl into their code, without any observable changes in behavior
(assuming the rest of the program remains unchanged).
