Files
docs-rust/ch20/ch20-05-macros.html
2026-06-22 21:27:36 +05:30

486 lines
31 KiB
HTML
Raw Permalink Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8">
<title>Macros</title>
</head>
<body>
<h2 id="macros"><a class="header" href="#macros">Macros</a></h2>
<p>Weve used macros like <code>println!</code> throughout this book, but we havent fully
explored what a macro is and how it works. The term <em>macro</em> refers to a family
of features in Rust—declarative macros with <code>macro_rules!</code> and three kinds of
procedural macros:</p>
<ul>
<li>Custom <code>#[derive]</code> macros that specify code added with the <code>derive</code> attribute
used on structs and enums</li>
<li>Attribute-like macros that define custom attributes usable on any item</li>
<li>Function-like macros that look like function calls but operate on the tokens
specified as their argument</li>
</ul>
<p>Well talk about each of these in turn, but first, lets look at why we even
need macros when we already have functions.</p>
<h3 id="the-difference-between-macros-and-functions"><a class="header" href="#the-difference-between-macros-and-functions">The Difference Between Macros and Functions</a></h3>
<p>Fundamentally, macros are a way of writing code that writes other code, which
is known as <em>metaprogramming</em>. In Appendix C, we discuss the <code>derive</code>
attribute, which generates an implementation of various traits for you. Weve
also used the <code>println!</code> and <code>vec!</code> macros throughout the book. All of these
macros <em>expand</em> to produce more code than the code youve written manually.</p>
<p>Metaprogramming is useful for reducing the amount of code you have to write and
maintain, which is also one of the roles of functions. However, macros have
some additional powers that functions dont have.</p>
<p>A function signature must declare the number and type of parameters the
function has. Macros, on the other hand, can take a variable number of
parameters: We can call <code>println!("hello")</code> with one argument or
<code>println!("hello {}", name)</code> with two arguments. Also, macros are expanded
before the compiler interprets the meaning of the code, so a macro can, for
example, implement a trait on a given type. A function cant, because it gets
called at runtime and a trait needs to be implemented at compile time.</p>
<p>The downside to implementing a macro instead of a function is that macro
definitions are more complex than function definitions because youre writing
Rust code that writes Rust code. Due to this indirection, macro definitions are
generally more difficult to read, understand, and maintain than function
definitions.</p>
<p>Another important difference between macros and functions is that you must
define macros or bring them into scope <em>before</em> you call them in a file, as
opposed to functions you can define anywhere and call anywhere.</p>
<!-- Old headings. Do not remove or links may break. -->
<p><a id="declarative-macros-with-macro_rules-for-general-metaprogramming"></a></p>
<h3 id="declarative-macros-for-general-metaprogramming"><a class="header" href="#declarative-macros-for-general-metaprogramming">Declarative Macros for General Metaprogramming</a></h3>
<p>The most widely used form of macros in Rust is the <em>declarative macro</em>. These
are also sometimes referred to as “macros by example,” “<code>macro_rules!</code> macros,”
or just plain “macros.” At their core, declarative macros allow you to write
something similar to a Rust <code>match</code> expression. As discussed in Chapter 6,
<code>match</code> expressions are control structures that take an expression, compare the
resultant value of the expression to patterns, and then run the code associated
with the matching pattern. Macros also compare a value to patterns that are
associated with particular code: In this situation, the value is the literal
Rust source code passed to the macro; the patterns are compared with the
structure of that source code; and the code associated with each pattern, when
matched, replaces the code passed to the macro. This all happens during
compilation.</p>
<p>To define a macro, you use the <code>macro_rules!</code> construct. Lets explore how to
use <code>macro_rules!</code> by looking at how the <code>vec!</code> macro is defined. Chapter 8
covered how we can use the <code>vec!</code> macro to create a new vector with particular
values. For example, the following macro creates a new vector containing three
integers:</p>
<pre class="playground"><code class="language-rust edition2024"><span class="boring">#![allow(unused)]
</span><span class="boring">fn main() {
</span>let v: Vec&lt;u32&gt; = vec![1, 2, 3];
<span class="boring">}</span></code></pre>
<p>We could also use the <code>vec!</code> macro to make a vector of two integers or a vector
of five string slices. We wouldnt be able to use a function to do the same
because we wouldnt know the number or type of values up front.</p>
<p>Listing 20-35 shows a slightly simplified definition of the <code>vec!</code> macro.</p>
<figure class="listing" id="listing-20-35">
<span class="file-name">Filename: src/lib.rs</span>
<pre><code class="language-rust noplayground">#[macro_export]
macro_rules! vec {
( $( $x:expr ),* ) =&gt; {
{
let mut temp_vec = Vec::new();
$(
temp_vec.push($x);
)*
temp_vec
}
};
}</code></pre>
<figcaption><a href="#listing-20-35">Listing 20-35</a>: A simplified version of the <code>vec!</code> macro definition</figcaption>
</figure>
<section class="note" aria-role="note">
<p>Note: The actual definition of the <code>vec!</code> macro in the standard library
includes code to pre-allocate the correct amount of memory up front. That code
is an optimization that we dont include here, to make the example simpler.</p>
</section>
<p>The <code>#[macro_export]</code> annotation indicates that this macro should be made
available whenever the crate in which the macro is defined is brought into
scope. Without this annotation, the macro cant be brought into scope.</p>
<p>We then start the macro definition with <code>macro_rules!</code> and the name of the
macro were defining <em>without</em> the exclamation mark. The name, in this case
<code>vec</code>, is followed by curly brackets denoting the body of the macro definition.</p>
<p>The structure in the <code>vec!</code> body is similar to the structure of a <code>match</code>
expression. Here we have one arm with the pattern <code>( $( $x:expr ),* )</code>,
followed by <code>=&gt;</code> and the block of code associated with this pattern. If the
pattern matches, the associated block of code will be emitted. Given that this
is the only pattern in this macro, there is only one valid way to match; any
other pattern will result in an error. More complex macros will have more than
one arm.</p>
<p>Valid pattern syntax in macro definitions is different from the pattern syntax
covered in Chapter 19 because macro patterns are matched against Rust code
structure rather than values. Lets walk through what the pattern pieces in
Listing 20-29 mean; for the full macro pattern syntax, see the <a href="../reference/macros-by-example.html">Rust
Reference</a>.</p>
<p>First, we use a set of parentheses to encompass the whole pattern. We use a
dollar sign (<code>$</code>) to declare a variable in the macro system that will contain
the Rust code matching the pattern. The dollar sign makes it clear this is a
macro variable as opposed to a regular Rust variable. Next comes a set of
parentheses that captures values that match the pattern within the parentheses
for use in the replacement code. Within <code>$()</code> is <code>$x:expr</code>, which matches any
Rust expression and gives the expression the name <code>$x</code>.</p>
<p>The comma following <code>$()</code> indicates that a literal comma separator character
must appear between each instance of the code that matches the code in <code>$()</code>.
The <code>*</code> specifies that the pattern matches zero or more of whatever precedes
the <code>*</code>.</p>
<p>When we call this macro with <code>vec![1, 2, 3];</code>, the <code>$x</code> pattern matches three
times with the three expressions <code>1</code>, <code>2</code>, and <code>3</code>.</p>
<p>Now lets look at the pattern in the body of the code associated with this arm:
<code>temp_vec.push()</code> within <code>$()*</code> is generated for each part that matches <code>$()</code>
in the pattern zero or more times depending on how many times the pattern
matches. The <code>$x</code> is replaced with each expression matched. When we call this
macro with <code>vec![1, 2, 3];</code>, the code generated that replaces this macro call
will be the following:</p>
<pre><code class="language-rust ignore">{
let mut temp_vec = Vec::new();
temp_vec.push(1);
temp_vec.push(2);
temp_vec.push(3);
temp_vec
}</code></pre>
<p>Weve defined a macro that can take any number of arguments of any type and can
generate code to create a vector containing the specified elements.</p>
<p>To learn more about how to write macros, consult the online documentation or
other resources, such as <a href="https://veykril.github.io/tlborm/">“The Little Book of Rust Macros”</a> started by
Daniel Keep and continued by Lukas Wirth.</p>
<h3 id="procedural-macros-for-generating-code-from-attributes"><a class="header" href="#procedural-macros-for-generating-code-from-attributes">Procedural Macros for Generating Code from Attributes</a></h3>
<p>The second form of macros is the procedural macro, which acts more like a
function (and is a type of procedure). <em>Procedural macros</em> accept some code as
an input, operate on that code, and produce some code as an output rather than
matching against patterns and replacing the code with other code as declarative
macros do. The three kinds of procedural macros are custom <code>derive</code>,
attribute-like, and function-like, and all work in a similar fashion.</p>
<p>When creating procedural macros, the definitions must reside in their own crate
with a special crate type. This is for complex technical reasons that we hope
to eliminate in the future. In Listing 20-36, we show how to define a
procedural macro, where <code>some_attribute</code> is a placeholder for using a specific
macro variety.</p>
<figure class="listing" id="listing-20-36">
<span class="file-name">Filename: src/lib.rs</span>
<pre><code class="language-rust ignore">use proc_macro::TokenStream;
#[some_attribute]
pub fn some_name(input: TokenStream) -&gt; TokenStream {
}</code></pre>
<figcaption><a href="#listing-20-36">Listing 20-36</a>: An example of defining a procedural macro</figcaption>
</figure>
<p>The function that defines a procedural macro takes a <code>TokenStream</code> as an input
and produces a <code>TokenStream</code> as an output. The <code>TokenStream</code> type is defined by
the <code>proc_macro</code> crate that is included with Rust and represents a sequence of
tokens. This is the core of the macro: The source code that the macro is
operating on makes up the input <code>TokenStream</code>, and the code the macro produces
is the output <code>TokenStream</code>. The function also has an attribute attached to it
that specifies which kind of procedural macro were creating. We can have
multiple kinds of procedural macros in the same crate.</p>
<p>Lets look at the different kinds of procedural macros. Well start with a
custom <code>derive</code> macro and then explain the small dissimilarities that make the
other forms different.</p>
<!-- Old headings. Do not remove or links may break. -->
<p><a id="how-to-write-a-custom-derive-macro"></a></p>
<h3 id="custom-derive-macros"><a class="header" href="#custom-derive-macros">Custom <code>derive</code> Macros</a></h3>
<p>Lets create a crate named <code>hello_macro</code> that defines a trait named
<code>HelloMacro</code> with one associated function named <code>hello_macro</code>. Rather than
making our users implement the <code>HelloMacro</code> trait for each of their types,
well provide a procedural macro so that users can annotate their type with
<code>#[derive(HelloMacro)]</code> to get a default implementation of the <code>hello_macro</code>
function. The default implementation will print <code>Hello, Macro! My name is TypeName!</code> where <code>TypeName</code> is the name of the type on which this trait has
been defined. In other words, well write a crate that enables another
programmer to write code like Listing 20-37 using our crate.</p>
<figure class="listing" id="listing-20-37">
<span class="file-name">Filename: src/main.rs</span>
<pre><code class="language-rust ignore does_not_compile">use hello_macro::HelloMacro;
use hello_macro_derive::HelloMacro;
#[derive(HelloMacro)]
struct Pancakes;
fn main() {
Pancakes::hello_macro();
}</code></pre>
<figcaption><a href="#listing-20-37">Listing 20-37</a>: The code a user of our crate will be able to write when using our procedural macro</figcaption>
</figure>
<p>This code will print <code>Hello, Macro! My name is Pancakes!</code> when were done. The
first step is to make a new library crate, like this:</p>
<pre><code class="language-console">$ cargo new hello_macro --lib
</code></pre>
<p>Next, in Listing 20-38, well define the <code>HelloMacro</code> trait and its associated
function.</p>
<figure class="listing" id="listing-20-38">
<span class="file-name">Filename: src/lib.rs</span>
<pre><code class="language-rust noplayground">pub trait HelloMacro {
fn hello_macro();
}</code></pre>
<figcaption><a href="#listing-20-38">Listing 20-38</a>: A simple trait that we will use with the <code>derive</code> macro</figcaption>
</figure>
<p>We have a trait and its function. At this point, our crate user could implement
the trait to achieve the desired functionality, as in Listing 20-39.</p>
<figure class="listing" id="listing-20-39">
<span class="file-name">Filename: src/main.rs</span>
<pre><code class="language-rust ignore">use hello_macro::HelloMacro;
struct Pancakes;
impl HelloMacro for Pancakes {
fn hello_macro() {
println!("Hello, Macro! My name is Pancakes!");
}
}
fn main() {
Pancakes::hello_macro();
}</code></pre>
<figcaption><a href="#listing-20-39">Listing 20-39</a>: How it would look if users wrote a manual implementation of the <code>HelloMacro</code> trait</figcaption>
</figure>
<p>However, they would need to write the implementation block for each type they
wanted to use with <code>hello_macro</code>; we want to spare them from having to do this
work.</p>
<p>Additionally, we cant yet provide the <code>hello_macro</code> function with default
implementation that will print the name of the type the trait is implemented
on: Rust doesnt have reflection capabilities, so it cant look up the types
name at runtime. We need a macro to generate code at compile time.</p>
<p>The next step is to define the procedural macro. At the time of this writing,
procedural macros need to be in their own crate. Eventually, this restriction
might be lifted. The convention for structuring crates and macro crates is as
follows: For a crate named <code>foo</code>, a custom <code>derive</code> procedural macro crate is
called <code>foo_derive</code>. Lets start a new crate called <code>hello_macro_derive</code> inside
our <code>hello_macro</code> project:</p>
<pre><code class="language-console">$ cargo new hello_macro_derive --lib
</code></pre>
<p>Our two crates are tightly related, so we create the procedural macro crate
within the directory of our <code>hello_macro</code> crate. If we change the trait
definition in <code>hello_macro</code>, well have to change the implementation of the
procedural macro in <code>hello_macro_derive</code> as well. The two crates will need to
be published separately, and programmers using these crates will need to add
both as dependencies and bring them both into scope. We could instead have the
<code>hello_macro</code> crate use <code>hello_macro_derive</code> as a dependency and re-export the
procedural macro code. However, the way weve structured the project makes it
possible for programmers to use <code>hello_macro</code> even if they dont want the
<code>derive</code> functionality.</p>
<p>We need to declare the <code>hello_macro_derive</code> crate as a procedural macro crate.
Well also need functionality from the <code>syn</code> and <code>quote</code> crates, as youll see
in a moment, so we need to add them as dependencies. Add the following to the
<em>Cargo.toml</em> file for <code>hello_macro_derive</code>:</p>
<figure class="listing">
<span class="file-name">Filename: hello_macro_derive/Cargo.toml</span>
<pre><code class="language-toml">[lib]
proc-macro = true
[dependencies]
syn = "2.0"
quote = "1.0"
</code></pre>
</figure>
<p>To start defining the procedural macro, place the code in Listing 20-40 into
your <em>src/lib.rs</em> file for the <code>hello_macro_derive</code> crate. Note that this code
wont compile until we add a definition for the <code>impl_hello_macro</code> function.</p>
<figure class="listing" id="listing-20-40">
<span class="file-name">Filename: hello_macro_derive/src/lib.rs</span>
<pre><code class="language-rust ignore does_not_compile">use proc_macro::TokenStream;
use quote::quote;
#[proc_macro_derive(HelloMacro)]
pub fn hello_macro_derive(input: TokenStream) -&gt; TokenStream {
// Construct a representation of Rust code as a syntax tree
// that we can manipulate.
let ast = syn::parse(input).unwrap();
// Build the trait implementation.
impl_hello_macro(&amp;ast)
}</code></pre>
<figcaption><a href="#listing-20-40">Listing 20-40</a>: Code that most procedural macro crates will require in order to process Rust code</figcaption>
</figure>
<p>Notice that weve split the code into the <code>hello_macro_derive</code> function, which
is responsible for parsing the <code>TokenStream</code>, and the <code>impl_hello_macro</code>
function, which is responsible for transforming the syntax tree: This makes
writing a procedural macro more convenient. The code in the outer function
(<code>hello_macro_derive</code> in this case) will be the same for almost every
procedural macro crate you see or create. The code you specify in the body of
the inner function (<code>impl_hello_macro</code> in this case) will be different
depending on your procedural macros purpose.</p>
<p>Weve introduced three new crates: <code>proc_macro</code>, <a href="https://crates.io/crates/syn"><code>syn</code></a><!-- ignore -->,
and <a href="https://crates.io/crates/quote"><code>quote</code></a><!-- ignore -->. The <code>proc_macro</code> crate comes with Rust,
so we didnt need to add that to the dependencies in <em>Cargo.toml</em>. The
<code>proc_macro</code> crate is the compilers API that allows us to read and manipulate
Rust code from our code.</p>
<p>The <code>syn</code> crate parses Rust code from a string into a data structure that we
can perform operations on. The <code>quote</code> crate turns <code>syn</code> data structures back
into Rust code. These crates make it much simpler to parse any sort of Rust
code we might want to handle: Writing a full parser for Rust code is no simple
task.</p>
<p>The <code>hello_macro_derive</code> function will be called when a user of our library
specifies <code>#[derive(HelloMacro)]</code> on a type. This is possible because weve
annotated the <code>hello_macro_derive</code> function here with <code>proc_macro_derive</code> and
specified the name <code>HelloMacro</code>, which matches our trait name; this is the
convention most procedural macros follow.</p>
<p>The <code>hello_macro_derive</code> function first converts the <code>input</code> from a
<code>TokenStream</code> to a data structure that we can then interpret and perform
operations on. This is where <code>syn</code> comes into play. The <code>parse</code> function in
<code>syn</code> takes a <code>TokenStream</code> and returns a <code>DeriveInput</code> struct representing the
parsed Rust code. Listing 20-41 shows the relevant parts of the <code>DeriveInput</code>
struct we get from parsing the <code>struct Pancakes;</code> string.</p>
<figure class="listing" id="listing-20-41">
<pre><code class="language-rust ignore">DeriveInput {
// --snip--
ident: Ident {
ident: "Pancakes",
span: #0 bytes(95..103)
},
data: Struct(
DataStruct {
struct_token: Struct,
fields: Unit,
semi_token: Some(
Semi
)
}
)
}</code></pre>
<figcaption><a href="#listing-20-41">Listing 20-41</a>: The <code>DeriveInput</code> instance we get when parsing the code that has the macros attribute in Listing 20-37</figcaption>
</figure>
<p>The fields of this struct show that the Rust code weve parsed is a unit struct
with the <code>ident</code> (<em>identifier</em>, meaning the name) of <code>Pancakes</code>. There are more
fields on this struct for describing all sorts of Rust code; check the <a href="https://docs.rs/syn/2.0/syn/struct.DeriveInput.html"><code>syn</code>
documentation for <code>DeriveInput</code></a> for more information.</p>
<p>Soon well define the <code>impl_hello_macro</code> function, which is where well build
the new Rust code we want to include. But before we do, note that the output
for our <code>derive</code> macro is also a <code>TokenStream</code>. The returned <code>TokenStream</code> is
added to the code that our crate users write, so when they compile their crate,
theyll get the extra functionality that we provide in the modified
<code>TokenStream</code>.</p>
<p>You might have noticed that were calling <code>unwrap</code> to cause the
<code>hello_macro_derive</code> function to panic if the call to the <code>syn::parse</code> function
fails here. Its necessary for our procedural macro to panic on errors because
<code>proc_macro_derive</code> functions must return <code>TokenStream</code> rather than <code>Result</code> to
conform to the procedural macro API. Weve simplified this example by using
<code>unwrap</code>; in production code, you should provide more specific error messages
about what went wrong by using <code>panic!</code> or <code>expect</code>.</p>
<p>Now that we have the code to turn the annotated Rust code from a <code>TokenStream</code>
into a <code>DeriveInput</code> instance, lets generate the code that implements the
<code>HelloMacro</code> trait on the annotated type, as shown in Listing 20-42.</p>
<figure class="listing" id="listing-20-42">
<span class="file-name">Filename: hello_macro_derive/src/lib.rs</span>
<pre><code class="language-rust ignore"><span class="boring">use proc_macro::TokenStream;
</span><span class="boring">use quote::quote;
</span><span class="boring">
</span><span class="boring">#[proc_macro_derive(HelloMacro)]
</span><span class="boring">pub fn hello_macro_derive(input: TokenStream) -&gt; TokenStream {
</span><span class="boring"> // Construct a representation of Rust code as a syntax tree
</span><span class="boring"> // that we can manipulate
</span><span class="boring"> let ast = syn::parse(input).unwrap();
</span><span class="boring">
</span><span class="boring"> // Build the trait implementation
</span><span class="boring"> impl_hello_macro(&amp;ast)
</span><span class="boring">}
</span><span class="boring">
</span>fn impl_hello_macro(ast: &amp;syn::DeriveInput) -&gt; TokenStream {
let name = &amp;ast.ident;
let generated = quote! {
impl HelloMacro for #name {
fn hello_macro() {
println!("Hello, Macro! My name is {}!", stringify!(#name));
}
}
};
generated.into()
}</code></pre>
<figcaption><a href="#listing-20-42">Listing 20-42</a>: Implementing the <code>HelloMacro</code> trait using the parsed Rust code</figcaption>
</figure>
<p>We get an <code>Ident</code> struct instance containing the name (identifier) of the
annotated type using <code>ast.ident</code>. The struct in Listing 20-41 shows that when
we run the <code>impl_hello_macro</code> function on the code in Listing 20-37, the
<code>ident</code> we get will have the <code>ident</code> field with a value of <code>"Pancakes"</code>. Thus,
the <code>name</code> variable in Listing 20-42 will contain an <code>Ident</code> struct instance
that, when printed, will be the string <code>"Pancakes"</code>, the name of the struct in
Listing 20-37.</p>
<p>The <code>quote!</code> macro lets us define the Rust code that we want to return. The
compiler expects something different from the direct result of the <code>quote!</code>
macros execution, so we need to convert it to a <code>TokenStream</code>. We do this by
calling the <code>into</code> method, which consumes this intermediate representation and
returns a value of the required <code>TokenStream</code> type.</p>
<p>The <code>quote!</code> macro also provides some very cool templating mechanics: We can
enter <code>#name</code>, and <code>quote!</code> will replace it with the value in the variable
<code>name</code>. You can even do some repetition similar to the way regular macros work.
Check out <a href="https://docs.rs/quote">the <code>quote</code> crates docs</a> for a thorough introduction.</p>
<p>We want our procedural macro to generate an implementation of our <code>HelloMacro</code>
trait for the type the user annotated, which we can get by using <code>#name</code>. The
trait implementation has the one function <code>hello_macro</code>, whose body contains the
functionality we want to provide: printing <code>Hello, Macro! My name is</code> and then
the name of the annotated type.</p>
<p>The <code>stringify!</code> macro used here is built into Rust. It takes a Rust
expression, such as <code>1 + 2</code>, and at compile time turns the expression into a
string literal, such as <code>"1 + 2"</code>. This is different from <code>format!</code> or
<code>println!</code>, which are macros that evaluate the expression and then turn the
result into a <code>String</code>. There is a possibility that the <code>#name</code> input might be
an expression to print literally, so we use <code>stringify!</code>. Using <code>stringify!</code>
also saves an allocation by converting <code>#name</code> to a string literal at compile
time.</p>
<p>At this point, <code>cargo build</code> should complete successfully in both <code>hello_macro</code>
and <code>hello_macro_derive</code>. Lets hook up these crates to the code in Listing
20-37 to see the procedural macro in action! Create a new binary project in
your <em>projects</em> directory using <code>cargo new pancakes</code>. We need to add
<code>hello_macro</code> and <code>hello_macro_derive</code> as dependencies in the <code>pancakes</code>
crates <em>Cargo.toml</em>. If youre publishing your versions of <code>hello_macro</code> and
<code>hello_macro_derive</code> to <a href="https://crates.io/">crates.io</a><!-- ignore -->, they
would be regular dependencies; if not, you can specify them as <code>path</code>
dependencies as follows:</p>
<pre><code class="language-toml">[dependencies]
hello_macro = { path = "../hello_macro" }
hello_macro_derive = { path = "../hello_macro/hello_macro_derive" }
</code></pre>
<p>Put the code in Listing 20-37 into <em>src/main.rs</em>, and run <code>cargo run</code>: It
should print <code>Hello, Macro! My name is Pancakes!</code>. The implementation of the
<code>HelloMacro</code> trait from the procedural macro was included without the
<code>pancakes</code> crate needing to implement it; the <code>#[derive(HelloMacro)]</code> added the
trait implementation.</p>
<p>Next, lets explore how the other kinds of procedural macros differ from custom
<code>derive</code> macros.</p>
<h3 id="attribute-like-macros"><a class="header" href="#attribute-like-macros">Attribute-Like Macros</a></h3>
<p>Attribute-like macros are similar to custom <code>derive</code> macros, but instead of
generating code for the <code>derive</code> attribute, they allow you to create new
attributes. Theyre also more flexible: <code>derive</code> only works for structs and
enums; attributes can be applied to other items as well, such as functions.
Heres an example of using an attribute-like macro. Say you have an attribute
named <code>route</code> that annotates functions when using a web application framework:</p>
<pre><code class="language-rust ignore">#[route(GET, "/")]
fn index() {</code></pre>
<p>This <code>#[route]</code> attribute would be defined by the framework as a procedural
macro. The signature of the macro definition function would look like this:</p>
<pre><code class="language-rust ignore">#[proc_macro_attribute]
pub fn route(attr: TokenStream, item: TokenStream) -&gt; TokenStream {</code></pre>
<p>Here, we have two parameters of type <code>TokenStream</code>. The first is for the
contents of the attribute: the <code>GET, "/"</code> part. The second is the body of the
item the attribute is attached to: in this case, <code>fn index() {}</code> and the rest
of the functions body.</p>
<p>Other than that, attribute-like macros work the same way as custom <code>derive</code>
macros: You create a crate with the <code>proc-macro</code> crate type and implement a
function that generates the code you want!</p>
<h3 id="function-like-macros"><a class="header" href="#function-like-macros">Function-Like Macros</a></h3>
<p>Function-like macros define macros that look like function calls. Similarly to
<code>macro_rules!</code> macros, theyre more flexible than functions; for example, they
can take an unknown number of arguments. However, <code>macro_rules!</code> macros can
only be defined using the match-like syntax we discussed in the <a href="#declarative-macros-with-macro_rules-for-general-metaprogramming">“Declarative
Macros for General Metaprogramming”</a><!-- ignore --> section earlier.
Function-like macros take a <code>TokenStream</code> parameter, and their definition
manipulates that <code>TokenStream</code> using Rust code as the other two types of
procedural macros do. An example of a function-like macro is an <code>sql!</code> macro
that might be called like so:</p>
<pre><code class="language-rust ignore">let sql = sql!(SELECT * FROM posts WHERE id=1);</code></pre>
<p>This macro would parse the SQL statement inside it and check that its
syntactically correct, which is much more complex processing than a
<code>macro_rules!</code> macro can do. The <code>sql!</code> macro would be defined like this:</p>
<pre><code class="language-rust ignore">#[proc_macro]
pub fn sql(input: TokenStream) -&gt; TokenStream {</code></pre>
<p>This definition is similar to the custom <code>derive</code> macros signature: We receive
the tokens that are inside the parentheses and return the code we wanted to
generate.</p>
<h2 id="summary"><a class="header" href="#summary">Summary</a></h2>
<p>Whew! Now you have some Rust features in your toolbox that you likely wont use
often, but youll know theyre available in very particular circumstances.
Weve introduced several complex topics so that when you encounter them in
error message suggestions or in other peoples code, youll be able to
recognize these concepts and syntax. Use this chapter as a reference to guide
you to solutions.</p>
<p>Next, well put everything weve discussed throughout the book into practice
and do one more project!</p>
</body>
</html>