The standard Rust distribution ships with a tool called rustdoc. Its job is to generate documentation for Rust projects. On a fundamental level, Rustdoc takes as an argument either a crate root or a Markdown file, and produces HTML, CSS, and JavaScript.

Let's give it a try! Let's create a new project with Cargo:

$ cargo new docs
$ cd docs

In src/lib.rs, you'll find that Cargo has generated some sample code. Delete it and replace it with this:

# #![allow(unused_variables)]
#fn main() {
/// foo is a function
fn foo() {}
#}

Let's run rustdoc on our code. To do so, we can call it with the path to our crate root like this:

$ rustdoc src/lib.rs

This will create a new directory, doc, with a website inside! In our case, the main page is located in doc/lib/index.html. If you open that up in a web browser, you'll see a page with a search bar, and "Crate lib" at the top, with no contents. There's two problems with this: first, why does it think that our package is named "lib"? Second, why does it not have any contents?

The first problem is due to rustdoc trying to be helpful; like rustc, it assumes that our crate's name is the name of the file for the crate root. To fix this, we can pass in a command-line flag:

$ rustdoc src/lib.rs --crate-name docs

Now, doc/docs/index.html will be generated, and the page says "Crate docs."

For the second issue, it's because our function foo is not public; rustdoc defaults to generating documentation for only public functions. If we change our code...

# #![allow(unused_variables)]
#fn main() {
/// foo is a function
pub fn foo() {}
#}

... and then re-run rustdoc:

$ rustdoc src/lib.rs --crate-name docs

We'll have some generated documentation. Open up doc/docs/index.html and check it out! It should show a link to the foo function's page, which is located at doc/docs/fn.foo.html. On that page, you'll see the "foo is a function" we put inside the documentation comment in our crate.

Cargo also has integration with rustdoc to make it easier to generate docs. Instead of the rustdoc command, we could have done this:

$ cargo doc

Internally, this calls out to rustdoc like this:

$ rustdoc --crate-name docs srclib.rs -o <path>\docs\target\doc -L
dependency=<path>docs\target\debug\deps

You can see this with cargo doc --verbose.

It generates the correct --crate-name for us, as well as pointing to src/lib.rs But what about those other arguments? -o controls the output of our docs. Instead of a top-level doc directory, you'll notice that Cargo puts generated documentation under target. That's the idiomatic place for generated files in Cargo projects. Also, it passes -L, a flag that helps rustdoc find the dependencies your code relies on. If our project used dependencies, we'd get documentation for them as well!

rustdoc can also generate HTML from standalone Markdown files. Let's give it a try: create a README.md file with these contents:

    # Docs

    This is a project to test out `rustdoc`.

    [Here is a link!](https://www.rust-lang.org)

    ## Subheading

    ```rust
    fn foo() -> i32 {
        1 + 1
    }
    ```

And call rustdoc on it:

$ rustdoc README.md

You'll find an HTML file in docs/doc/README.html generated from its Markdown contents.

Cargo currently does not understand standalone Markdown files, unfortunately.

This covers the simplest use-cases of rustdoc. The rest of this book will explain all of the options that rustdoc has, and how to use them.

Here's the list of arguments you can pass to rustdoc:

Using this flag looks like this:

$ rustdoc -h
$ rustdoc --help

This will show rustdoc's built-in help, which largely consists of a list of possible command-line flags.

Some of rustdoc's flags are unstable; this page only shows stable options, --help will show them all.

Using this flag looks like this:

$ rustdoc -V
$ rustdoc --version

This will show rustdoc's version, which will look something like this:

rustdoc 1.17.0 (56124baa9 2017-04-24)

Using this flag looks like this:

$ rustdoc -v src/lib.rs
$ rustdoc --verbose src/lib.rs

This enables "verbose mode", which means that more information will be written to standard out. What is written depends on the other flags you've passed in. For example, with --version:

$ rustdoc --verbose --version
rustdoc 1.17.0 (56124baa9 2017-04-24)
binary: rustdoc
commit-hash: hash
commit-date: date
host: host-triple
release: 1.17.0
LLVM version: 3.9

This flag is currently ignored; the idea is that rustdoc would support various input formats, and you could specify them via this flag.

Rustdoc only supports Rust source code and Markdown input formats. If the file ends in .md or .markdown, rustdoc treats it as a Markdown file. Otherwise, it assumes that the input file is Rust.

This flag is currently ignored; the idea is that rustdoc would support various output formats, and you could specify them via this flag.

Rustdoc only supports HTML output, and so this flag is redundant today.

Using this flag looks like this:

$ rustdoc src/lib.rs -o target\\doc
$ rustdoc src/lib.rs --output target\\doc

By default, rustdoc's output appears in a directory named doc in the current working directory. With this flag, it will place all output into the directory you specify.

Using this flag looks like this:

$ rustdoc src/lib.rs --crate-name mycrate

By default, rustdoc assumes that the name of your crate is the same name as the .rs file. --crate-name lets you override this assumption with whatever name you choose.

Using this flag looks like this:

$ rustdoc src/lib.rs -L target/debug/deps
$ rustdoc src/lib.rs --library-path target/debug/deps

If your crate has dependencies, rustdoc needs to know where to find them. Passing --library-path gives rustdoc a list of places to look for these dependencies.

This flag takes any number of directories as its argument, and will use all of them when searching.

Using this flag looks like this:

$ rustdoc src/lib.rs --cfg feature="foo"

This flag accepts the same values as rustc --cfg, and uses it to configure compilation. The example above uses feature, but any of the cfg values are acceptable.

Using this flag looks like this:

$ rustdoc src/lib.rs --extern lazy-static=/path/to/lazy-static

Similar to --library-path, --extern is about specifying the location of a dependency. --library-path provides directories to search in, --extern instead lets you specify exactly which dependency is located where.

Using this flag looks like this:

$ rustdoc --passes list
$ rustdoc src/lib.rs --passes strip-priv-imports

An argument of "list" will print a list of possible "rustdoc passes", and other arguments will be the name of which passes to run in addition to the defaults.

For more details on passes, see the chapter on them.

See also --no-defaults.

Using this flag looks like this:

$ rustdoc src/lib.rs --no-defaults

By default, rustdoc will run several passes over your code. This removes those defaults, allowing you to use --passes to specify exactly which passes you want.

For more details on passes, see the chapter on them.

See also --passes.

Using this flag looks like this:

$ rustdoc src/lib.rs --test

This flag will run your code examples as tests. For more, see the chapter on documentation tests.

See also --test-args.

Using this flag looks like this:

$ rustdoc src/lib.rs --test --test-args ignored

This flag will pass options to the test runner when running documentation tests. For more, see the chapter on documentation tests.

See also --test.

Using this flag looks like this:

$ rustdoc src/lib.rs --target x86_64-pc-windows-gnu

Similar to the --target flag for rustc, this generates documentation for a target triple that's different than your host triple.

All of the usual caveats of cross-compiling code apply.

Using this flag looks like this:

$ rustdoc README.md --markdown-css foo.css

When rendering Markdown files, this will create a <link> element in the <head> section of the generated HTML. For example, with the invocation above,

<link rel="stylesheet" type="text/css" href="foo.css">

will be added.

When rendering Rust files, this flag is ignored.

Using this flag looks like this:

$ rustdoc src/lib.rs --html-in-header header.html
$ rustdoc README.md --html-in-header header.html

This flag takes a list of files, and inserts them into the <head> section of the rendered documentation.

Using this flag looks like this:

$ rustdoc src/lib.rs --html-before-content extra.html
$ rustdoc README.md --html-before-content extra.html

This flag takes a list of files, and inserts them inside the <body> tag but before the other content rustdoc would normally produce in the rendered documentation.

Using this flag looks like this:

$ rustdoc src/lib.rs --html-after-content extra.html
$ rustdoc README.md --html-after-content extra.html

This flag takes a list of files, and inserts them before the </body> tag but after the other content rustdoc would normally produce in the rendered documentation.

Using this flag looks like this:

$ rustdoc README.md --markdown-playground-url https://play.rust-lang.org/

When rendering a Markdown file, this flag gives the base URL of the Rust Playground, to use for generating Run buttons.

Using this flag looks like this:

$ rustdoc README.md --markdown-no-toc

When generating documentation from a Markdown file, by default, rustdoc will generate a table of contents. This flag suppresses that, and no TOC will be generated.

Using this flag looks like this:

$ rustdoc src/lib.rs -e extra.css
$ rustdoc src/lib.rs --extend-css extra.css

With this flag, the contents of the files you pass are included at the bottom of Rustdoc's theme.css file.

While this flag is stable, the contents of theme.css are not, so be careful! Updates may break your theme extensions.

Using this flag looks like this:

$ rustdoc src/lib.rs --sysroot /path/to/sysroot

Similar to rustc --sysroot, this lets you change the sysroot rustdoc uses when compiling your code.

The #[doc] attribute lets you control various aspects of how rustdoc does its job.

The most basic function of #[doc] is to handle the actual documentation text. That is, /// is syntax sugar for #[doc]. This means that these two are the same:

/// This is a doc comment.
#[doc = " This is a doc comment."]

(Note the leading space in the attribute version.)

In most cases, /// is easier to use than #[doc]. One case where the latter is easier is when generating documentation in macros; the collapse-docs pass will combine multiple #[doc] attributes into a single doc comment, letting you generate code like this:

#[doc = "This is"]
#[doc = " a "]
#[doc = "doc comment"]

Which can feel more flexible. Note that this would generate this:

#[doc = "This is\n a \ndoc comment"]

but given that docs are rendered via Markdown, it will remove these newlines.

The doc attribute has more options though! These don't involve the text of the output, but instead, various aspects of the presentation of the output. We've split them into two kinds below: attributes that are useful at the crate level, and ones that are useful at the item level.

These options control how the docs look at a macro level.

This form of the doc attribute lets you control the favicon of your docs.

#![doc(html_favicon_url = "https://example.com/favicon.ico")]

This will put <link rel="shortcut icon" href="{}"> into your docs, where the string for the attribute goes into the {}.

If you don't use this attribute, there will be no favicon.

This form of the doc attribute lets you control the logo in the upper left hand side of the docs.

#![doc(html_logo_url = "https://example.com/logo.jpg")]

This will put <a href='index.html'><img src='{}' alt='logo' width='100'></a> into your docs, where the string for the attribute goes into the {}.

If you don't use this attribute, there will be no logo.

This form of the doc attribute lets you control where the "run" buttons on your documentation examples make requests to.

#![doc(html_playground_url = "https://playground.example.com/")]

Now, when you press "run", the button will make a request to this domain.

If you don't use this attribute, there will be no run buttons.

This form of the doc attribute is mostly only useful for the standard library; When a feature is unstable, an issue number for tracking the feature must be given. rustdoc uses this number, plus the base URL given here, to link to the tracking issue.

#![doc(issue_tracker_base_url = "https://github.com/rust-lang/rust/issues/")]

By default, rustdoc will include the source code of your program, with links to it in the docs. But if you include this:

#![doc(html_no_source)]

it will not.

By default, rustdoc will automatically add a line with extern crate my_crate; into each doctest. But if you include this:

#![doc(test(no_crate_inject))]

it will not.

This form of the doc attribute allows you to add arbitrary attributes to all your doctests. For example, if you want your doctests to fail if they produce any warnings, you could add this:

#![doc(test(attr(deny(warnings))))]

These forms of the #[doc] attribute are used on individual items, to control how they are documented.

These attributes are used on use statements, and control where the documentation shows up. For example, consider this Rust code:

pub use bar::Bar;

/// bar docs
pub mod bar {
    /// the docs for Bar
    pub struct Bar;
}

The documentation will generate a "Re-exports" section, and say pub use bar::Bar;, where Bar is a link to its page.

If we change the use line like this:

#[doc(inline)]
pub use bar::Bar;

Instead, Bar will appear in a Structs section, just like Bar was defined at the top level, rather than pub use'd.

Let's change our original example, by making bar private:

pub use bar::Bar;

/// bar docs
mod bar {
    /// the docs for Bar
    pub struct Bar;
}

Here, because bar is not public, Bar wouldn't have its own page, so there's nowhere to link to. rustdoc will inline these definitions, and so we end up in the same case as the #[doc(inline)] above; Bar is in a Structs section, as if it were defined at the top level. If we add the no_inline form of the attribute:

#[doc(no_inline)]
pub use bar::Bar;

/// bar docs
mod bar {
    /// the docs for Bar
    pub struct Bar;
}

Now we'll have a Re-exports line, and Bar will not link to anywhere.

Any item annotated with #[doc(hidden)] will not appear in the documentation, unless the strip-hidden pass is removed.

Since primitive types are defined in the compiler, there's no place to attach documentation attributes. This attribute is used by the standard library to provide a way to generate documentation for primitive types.

rustdoc supports executing your documentation examples as tests. This makes sure that your tests are up to date and working.

The basic idea is this:

/// # Examples
///
/// ```
/// let x = 5;
/// ```

The triple backticks start and end code blocks. If this were in a file named foo.rs, running rustdoc --test foo.rs will extract this example, and then run it as a test.

Please note that by default, if no language is set for the block code, rustdoc assumes it is Rust code. So the following:

# #![allow(unused_variables)]
#fn main() {
let x = 5;
#}

is strictly equivalent to:

let x = 5;

There's some subtlety though! Read on for more details.

In the example above, you'll note something strange: there's no main function! Forcing you to write main for every example, no matter how small, adds friction. So rustdoc processes your examples slightly before running them. Here's the full algorithm rustdoc uses to preprocess examples:

1. Some common allow attributes are inserted, including unused_variables, unused_assignments, unused_mut, unused_attributes, and dead_code. Small examples often trigger these lints.
2. Any attributes specified with #![doc(test(attr(...)))] are added.
3. Any leading #![foo] attributes are left intact as crate attributes.
4. If the example does not contain extern crate, and #![doc(test(no_crate_inject))] was not specified, then extern crate <mycrate>; is inserted (note the lack of #[macro_use]).
5. Finally, if the example does not contain fn main, the remainder of the text is wrapped in fn main() { your_code }.

For more about that caveat in rule 4, see "Documenting Macros" below.

Sometimes, you need some setup code, or other things that would distract from your example, but are important to make the tests work. Consider an example block that looks like this:

/// Some documentation.
# fn foo() {}

It will render like this:

# #![allow(unused_variables)]
#fn main() {
/// Some documentation.
# fn foo() {}
#}

Yes, that's right: you can add lines that start with #, and they will be hidden from the output, but will be used when compiling your code. You can use this to your advantage. In this case, documentation comments need to apply to some kind of function, so if I want to show you just a documentation comment, I need to add a little function definition below it. At the same time, it's only there to satisfy the compiler, so hiding it makes the example more clear. You can use this technique to explain longer examples in detail, while still preserving the testability of your documentation.

For example, imagine that we wanted to document this code:

# #![allow(unused_variables)]
#fn main() {
let x = 5;
let y = 6;
println!("{}", x + y);
#}

We might want the documentation to end up looking like this:

First, we set x to five:

# #![allow(unused_variables)]
#fn main() {
let x = 5;
# let y = 6;
# println!("{}", x + y);
#}

Next, we set y to six:

# #![allow(unused_variables)]
#fn main() {
# let x = 5;
let y = 6;
# println!("{}", x + y);
#}

Finally, we print the sum of x and y:

# #![allow(unused_variables)]
#fn main() {
# let x = 5;
# let y = 6;
println!("{}", x + y);
#}

To keep each code block testable, we want the whole program in each block, but we don't want the reader to see every line every time. Here's what we put in our source code:

    First, we set `x` to five:

    ```
    let x = 5;
    # let y = 6;
    # println!("{}", x + y);
    ```

    Next, we set `y` to six:

    ```
    # let x = 5;
    let y = 6;
    # println!("{}", x + y);
    ```

    Finally, we print the sum of `x` and `y`:

    ```
    # let x = 5;
    # let y = 6;
    println!("{}", x + y);
    ```

By repeating all parts of the example, you can ensure that your example still compiles, while only showing the parts that are relevant to that part of your explanation.

Another case where the use of # is handy is when you want to ignore error handling. Lets say you want the following,

/// use std::io;
/// let mut input = String::new();
/// io::stdin().read_line(&mut input)?;

The problem is that ? returns a Result<T, E> and test functions don't return anything so this will give a mismatched types error.

/// A doc test using ?
///
/// ```
/// use std::io;
/// # fn foo() -> io::Result<()> {
/// let mut input = String::new();
/// io::stdin().read_line(&mut input)?;
/// # Ok(())
/// # }
/// ```
# fn foo() {}

You can get around this by wrapping the code in a function. This catches and swallows the Result<T, E> when running tests on the docs. This pattern appears regularly in the standard library.

Here’s an example of documenting a macro:

/// Panic with a given message unless an expression evaluates to true.
///
/// # Examples
///
/// ```
/// # #[macro_use] extern crate foo;
/// # fn main() {
/// panic_unless!(1 + 1 == 2, “Math is broken.”);
/// # }
/// ```
///
/// ```should_panic
/// # #[macro_use] extern crate foo;
/// # fn main() {
/// panic_unless!(true == false, “I’m broken.”);
/// # }
/// ```
#[macro_export]
macro_rules! panic_unless {
    ($condition:expr, $($rest:expr),+) => ({ if ! $condition { panic!($($rest),+); } });
}
# fn main() {}

You’ll note three things: we need to add our own extern crate line, so that we can add the #[macro_use] attribute. Second, we’ll need to add our own main() as well (for reasons discussed above). Finally, a judicious use of # to comment out those two things, so they don’t show up in the output.

There are a few annotations that are useful to help rustdoc do the right thing when testing your code:

# #![allow(unused_variables)]
#fn main() {
/// ```ignore
/// fn foo() {
/// ```
# fn foo() {}
#}

The ignore directive tells Rust to ignore your code. This is almost never what you want, as it's the most generic. Instead, consider annotating it with text if it's not code, or using #s to get a working example that only shows the part you care about.

# #![allow(unused_variables)]
#fn main() {
/// ```should_panic
/// assert!(false);
/// ```
# fn foo() {}
#}

should_panic tells rustdoc that the code should compile correctly, but not actually pass as a test.

/// ```no_run
/// loop {
///     println!("Hello, world");
/// }
/// ```
# fn foo() {}

compile_fail tells rustdoc that the compilation should fail. If it compiles, then the test will fail. However please note that code failing with the current Rust release may work in a future release, as new features are added.

/// ```compile_fail
/// let x = 5;
/// x += 2; // shouldn't compile!
/// ```

The no_run attribute will compile your code, but not run it. This is important for examples such as "Here's how to retrieve a web page," which you would want to ensure compiles, but might be run in a test environment that has no network access.

Rustdoc has a concept called "passes". These are transformations that rustdoc runs on your documentation before producing its final output.

In addition to the passes below, check out the docs for these flags:

• --passes
• --no-defaults

By default, rustdoc will run some passes, namely:

• strip-hidden
• strip-private
• collapse-docs
• unindent-comments

However, strip-private implies strip-private-imports, and so effectively, all passes are run by default.

This pass implements the #[doc(hidden)] attribute. When this pass runs, it checks each item, and if it is annotated with this attribute, it removes it from rustdoc's output.

Without this pass, these items will remain in the output.

When you write a doc comment like this:

/// This is a documentation comment.

There's a space between the /// and that T. That spacing isn't intended to be a part of the output; it's there for humans, to help separate the doc comment syntax from the text of the comment. This pass is what removes that space.

The exact rules are left under-specified so that we can fix issues that we find.

Without this pass, the exact number of spaces is preserved.

With this pass, multiple #[doc] attributes are converted into one single documentation string.

For example:

#[doc = "This is the first line."]
#[doc = "This is the second line."]

Gets collapsed into a single doc string of

This is the first line.
This is the second line.

This removes documentation for any non-public items, so for example:

/// These are private docs.
struct Private;

/// These are public docs.
pub struct Public;

This pass removes the docs for Private, since they're not public.

This pass implies strip-priv-imports.

This is the same as strip-private, but for extern crate and use statements instead of items.