Trait Arbitrary

pub trait Arbitrary<'a>: Sized {
    // Required method
    fn arbitrary(u: &mut Unstructured<'a>) -> Result<Self, Error>;

    // Provided methods
    fn arbitrary_take_rest(u: Unstructured<'a>) -> Result<Self, Error> { ... }
    fn size_hint(depth: usize) -> (usize, Option<usize>) { ... }
    fn try_size_hint(
        depth: usize,
    ) -> Result<(usize, Option<usize>), MaxRecursionReached> { ... }
}

Expand description

Generate arbitrary structured values from raw, unstructured data.

The Arbitrary trait allows you to generate valid structured values, like HashMaps, or ASTs, or MyTomlConfig, or any other data structure from raw, unstructured bytes provided by a fuzzer.

§Deriving `Arbitrary`

Automatically deriving the Arbitrary trait is the recommended way to implement Arbitrary for your types.

Using the custom derive requires that you enable the "derive" cargo feature in your Cargo.toml:

[dependencies]
arbitrary = { version = "1", features = ["derive"] }

Then, you add the #[derive(Arbitrary)] annotation to your struct or enum type definition:

use arbitrary::Arbitrary;
use std::collections::HashSet;

#[derive(Arbitrary)]
pub struct AddressBook {
    friends: HashSet<Friend>,
}

#[derive(Arbitrary, Hash, Eq, PartialEq)]
pub enum Friend {
    Buddy { name: String },
    Pal { age: usize },
}

Every member of the struct or enum must also implement Arbitrary.

It is also possible to change the default bounds added by the derive:

use arbitrary::Arbitrary;

trait Trait {
    type Assoc: for<'a> Arbitrary<'a>;
}

#[derive(Arbitrary)]
// The bounds are used verbatim, so any existing trait bounds will need to be repeated.
#[arbitrary(bound = "T: Trait")]
struct Point<T: Trait> {
    x: T::Assoc,
}

§Implementing `Arbitrary` By Hand

Implementing Arbitrary mostly involves nested calls to other Arbitrary arbitrary implementations for each of your struct or enum’s members. But sometimes you need some amount of raw data, or you need to generate a variably-sized collection type, or something of that sort. The Unstructured type helps you with these tasks.

use arbitrary::{Arbitrary, Result, Unstructured};

impl<'a, T> Arbitrary<'a> for MyCollection<T>
where
    T: Arbitrary<'a>,
{
    fn arbitrary(u: &mut Unstructured<'a>) -> Result<Self> {
        // Get an iterator of arbitrary `T`s.
        let iter = u.arbitrary_iter::<T>()?;

        // And then create a collection!
        let mut my_collection = MyCollection::new();
        for elem_result in iter {
            let elem = elem_result?;
            my_collection.insert(elem);
        }

        Ok(my_collection)
    }
}

§A Note On Output Distributions

There is no requirement for a particular distribution of the values. For example, it is not required that every value appears with the same probability. That being said, the main use for Arbitrary is for fuzzing, so in many cases a uniform distribution will make the most sense in order to provide the best coverage of the domain. In other cases this is not desirable or even possible, for example when sampling from a uniform distribution is computationally expensive or in the case of collections that may grow indefinitely.

Required Methods§

fn arbitrary(u: &mut Unstructured<'a>) -> Result<Self, Error>

Generate an arbitrary value of Self from the given unstructured data.

Calling Arbitrary::arbitrary requires that you have some raw data, perhaps given to you by a fuzzer like AFL or libFuzzer. You wrap this raw data in an Unstructured, and then you can call <MyType as Arbitrary>::arbitrary to construct an arbitrary instance of MyType from that unstructured data.

Implementations may return an error if there is not enough data to construct a full instance of Self, or they may fill out the rest of Self with dummy values. Using dummy values when the underlying data is exhausted can help avoid accidentally “defeating” some of the fuzzer’s mutations to the underlying byte stream that might otherwise lead to interesting runtime behavior or new code coverage if only we had just a few more bytes. However, it also requires that implementations for recursive types (e.g. struct Foo(Option<Box<Foo>>)) avoid infinite recursion when the underlying data is exhausted.

use arbitrary::{Arbitrary, Unstructured};

#[derive(Arbitrary)]
pub struct MyType {
    // ...
}

// Get the raw data from the fuzzer or wherever else.
let raw_data: &[u8] = get_raw_data_from_fuzzer();

// Wrap that raw data in an `Unstructured`.
let mut unstructured = Unstructured::new(raw_data);

// Generate an arbitrary instance of `MyType` and do stuff with it.
if let Ok(value) = MyType::arbitrary(&mut unstructured) {
    do_stuff(value);
}

See also the documentation for Unstructured.

Provided Methods§

fn arbitrary_take_rest(u: Unstructured<'a>) -> Result<Self, Error>

Generate an arbitrary value of Self from the entirety of the given unstructured data.

This is similar to Arbitrary::arbitrary, however it assumes that it is the last consumer of the given data, and is thus able to consume it all if it needs. See also the documentation for Unstructured.

fn size_hint(depth: usize) -> (usize, Option<usize>)

Get a size hint for how many bytes out of an Unstructured this type needs to construct itself.

This is useful for determining how many elements we should insert when creating an arbitrary collection.

The return value is similar to Iterator::size_hint: it returns a tuple where the first element is a lower bound on the number of bytes required, and the second element is an optional upper bound.

The default implementation return (0, None) which is correct for any type, but not ultimately that useful. Using #[derive(Arbitrary)] will create a better implementation. If you are writing an Arbitrary implementation by hand, and your type can be part of a dynamically sized collection (such as Vec), you are strongly encouraged to override this default with a better implementation, and also override try_size_hint.

§How to implement this

If the size hint calculation is a trivial constant and does not recurse into any other size_hint call, you should implement it in size_hint:

use arbitrary::{size_hint, Arbitrary, Result, Unstructured};

struct SomeStruct(u8);

impl<'a> Arbitrary<'a> for SomeStruct {
    fn arbitrary(u: &mut Unstructured<'a>) -> Result<Self> {
        let buf = &mut [0];
        u.fill_buffer(buf)?;
        Ok(SomeStruct(buf[0]))
    }

    #[inline]
    fn size_hint(depth: usize) -> (usize, Option<usize>) {
        let _ = depth;
        (1, Some(1))
    }
}

Otherwise, it should instead be implemented in try_size_hint, and the size_hint implementation should forward to it:

use arbitrary::{size_hint, Arbitrary, MaxRecursionReached, Result, Unstructured};

struct SomeStruct<A, B> {
    a: A,
    b: B,
}

impl<'a, A: Arbitrary<'a>, B: Arbitrary<'a>> Arbitrary<'a> for SomeStruct<A, B> {
    fn arbitrary(u: &mut Unstructured<'a>) -> Result<Self> {
        // ...
    }

    fn size_hint(depth: usize) -> (usize, Option<usize>) {
        // Return the value of try_size_hint
        //
        // If the recursion fails, return the default, always valid `(0, None)`
        Self::try_size_hint(depth).unwrap_or_default()
    }

    fn try_size_hint(depth: usize) -> Result<(usize, Option<usize>), MaxRecursionReached> {
        // Protect against potential infinite recursion with
        // `try_recursion_guard`.
        size_hint::try_recursion_guard(depth, |depth| {
            // If we aren't too deep, then `recursion_guard` calls
            // this closure, which implements the natural size hint.
            // Don't forget to use the new `depth` in all nested
            // `try_size_hint` calls! We recommend shadowing the
            // parameter, like what is done here, so that you can't
            // accidentally use the wrong depth.
            Ok(size_hint::and(
                <A as Arbitrary>::try_size_hint(depth)?,
                <B as Arbitrary>::try_size_hint(depth)?,
            ))
        })
    }
}

§Invariant

It must be possible to construct every possible output using only inputs of lengths bounded by these parameters. This applies to both Arbitrary::arbitrary and Arbitrary::arbitrary_take_rest.

This is trivially true for (0, None). To restrict this further, it must be proven that all inputs that are now excluded produced redundant outputs which are still possible to produce using the reduced input space.

fn try_size_hint( depth: usize, ) -> Result<(usize, Option<usize>), MaxRecursionReached>

Get a size hint for how many bytes out of an Unstructured this type needs to construct itself.

Unlike size_hint, this function keeps the information that the recursion limit was reached. This is required to “short circuit” the calculation and avoid exponential blowup with recursive structures.

If you are implementing size_hint for a struct that could be recursive, you should implement try_size_hint and call the try_size_hint when recursing

The return value is similar to core::iter::Iterator::size_hint: it returns a tuple where the first element is a lower bound on the number of bytes required, and the second element is an optional upper bound.

The default implementation returns the value of size_hint which is correct for any type, but might lead to exponential blowup when dealing with recursive types.

§Invariant

It must be possible to construct every possible output using only inputs of lengths bounded by these parameters. This applies to both Arbitrary::arbitrary and Arbitrary::arbitrary_take_rest.

This is trivially true for (0, None). To restrict this further, it must be proven that all inputs that are now excluded produced redundant outputs which are still possible to produce using the reduced input space.

§When to implement `try_size_hint`

If you 100% know that the type you are implementing Arbitrary for is not a recursive type, or your implementation is not transitively calling any other size_hint methods, you may implement size_hint, and the default try_size_hint implementation will use it.

Note that if you are implementing Arbitrary for a generic type, you cannot guarantee the lack of type recursion!

Otherwise, when there is possible type recursion, you should implement try_size_hint instead.

§The `depth` parameter

When implementing try_size_hint, you need to use arbitrary::size_hint::try_recursion_guard(depth) to prevent potential infinite recursion when calculating size hints for potentially recursive types:

use arbitrary::{size_hint, Arbitrary, MaxRecursionReached, Unstructured};

// This can potentially be a recursive type if `L` or `R` contain
// something like `Box<Option<MyEither<L, R>>>`!
enum MyEither<L, R> {
    Left(L),
    Right(R),
}

impl<'a, L, R> Arbitrary<'a> for MyEither<L, R>
where
    L: Arbitrary<'a>,
    R: Arbitrary<'a>,
{
    fn arbitrary(u: &mut Unstructured) -> arbitrary::Result<Self> {
        // ...
    }

    fn size_hint(depth: usize) -> (usize, Option<usize>) {
        // Return the value of `try_size_hint`
        //
        // If the recursion fails, return the default `(0, None)` range,
        // which is always valid.
        Self::try_size_hint(depth).unwrap_or_default()
    }

    fn try_size_hint(depth: usize) -> Result<(usize, Option<usize>), MaxRecursionReached> {
        // Protect against potential infinite recursion with
        // `try_recursion_guard`.
        size_hint::try_recursion_guard(depth, |depth| {
            // If we aren't too deep, then `recursion_guard` calls
            // this closure, which implements the natural size hint.
            // Don't forget to use the new `depth` in all nested
            // `try_size_hint` calls! We recommend shadowing the
            // parameter, like what is done here, so that you can't
            // accidentally use the wrong depth.
            Ok(size_hint::or(
                <L as Arbitrary>::try_size_hint(depth)?,
                <R as Arbitrary>::try_size_hint(depth)?,
            ))
        })
    }
}

Dyn Compatibility§

This trait is not dyn compatible.

In older versions of Rust, dyn compatibility was called "object safety", so this trait is not object safe.

Implementations on Foreign Types§

§

impl Arbitrary<'_> for OpTxType

Available on crate feature arbitrary only.

§

fn arbitrary(u: &mut Unstructured<'_>) -> Result<OpTxType, Error>

§

impl Arbitrary<'_> for TxType

Available on crate feature arbitrary only.

§

impl<'a> Arbitrary<'a> for usize

Trait ArbitraryCopy item path

§Deriving Arbitrary

§Implementing Arbitrary By Hand

§A Note On Output Distributions

Required Methods§

fn arbitrary(u: &mut Unstructured<'a>) -> Result<Self, Error>

Provided Methods§

fn arbitrary_take_rest(u: Unstructured<'a>) -> Result<Self, Error>

fn size_hint(depth: usize) -> (usize, Option<usize>)

§How to implement this

§Invariant

fn try_size_hint( depth: usize, ) -> Result<(usize, Option<usize>), MaxRecursionReached>

§Invariant

§When to implement try_size_hint

§The depth parameter

Dyn Compatibility§

Implementations on Foreign Types§

impl Arbitrary<'_> for OpTxType

fn arbitrary(u: &mut Unstructured<'_>) -> Result<OpTxType, Error>

impl Arbitrary<'_> for TxType

fn arbitrary(u: &mut Unstructured<'_>) -> Result<TxType, Error>

impl<'a> Arbitrary<'a> for &'a str

fn arbitrary(u: &mut Unstructured<'a>) -> Result<&'a str, Error>

fn arbitrary_take_rest(u: Unstructured<'a>) -> Result<&'a str, Error>

fn size_hint(_depth: usize) -> (usize, Option<usize>)

impl<'a> Arbitrary<'a> for &'a [u8]

fn arbitrary(u: &mut Unstructured<'a>) -> Result<&'a [u8], Error>

fn arbitrary_take_rest(u: Unstructured<'a>) -> Result<&'a [u8], Error>

fn size_hint(_depth: usize) -> (usize, Option<usize>)

impl<'a> Arbitrary<'a> for IpAddr

fn arbitrary(u: &mut Unstructured<'a>) -> Result<IpAddr, Error>

fn size_hint(depth: usize) -> (usize, Option<usize>)

impl<'a> Arbitrary<'a> for SocketAddr

fn arbitrary(u: &mut Unstructured<'a>) -> Result<SocketAddr, Error>

fn size_hint(depth: usize) -> (usize, Option<usize>)

impl<'a> Arbitrary<'a> for bool

fn arbitrary(u: &mut Unstructured<'a>) -> Result<bool, Error>

fn size_hint(depth: usize) -> (usize, Option<usize>)

impl<'a> Arbitrary<'a> for char

fn arbitrary(u: &mut Unstructured<'a>) -> Result<char, Error>

fn size_hint(depth: usize) -> (usize, Option<usize>)

impl<'a> Arbitrary<'a> for f32

fn arbitrary(u: &mut Unstructured<'a>) -> Result<f32, Error>

fn size_hint(depth: usize) -> (usize, Option<usize>)

impl<'a> Arbitrary<'a> for f64

fn arbitrary(u: &mut Unstructured<'a>) -> Result<f64, Error>

fn size_hint(depth: usize) -> (usize, Option<usize>)

impl<'a> Arbitrary<'a> for i8

fn arbitrary(u: &mut Unstructured<'a>) -> Result<i8, Error>

fn size_hint(_depth: usize) -> (usize, Option<usize>)

impl<'a> Arbitrary<'a> for i16

fn arbitrary(u: &mut Unstructured<'a>) -> Result<i16, Error>

fn size_hint(_depth: usize) -> (usize, Option<usize>)

impl<'a> Arbitrary<'a> for i32

fn arbitrary(u: &mut Unstructured<'a>) -> Result<i32, Error>

fn size_hint(_depth: usize) -> (usize, Option<usize>)

impl<'a> Arbitrary<'a> for i64

fn arbitrary(u: &mut Unstructured<'a>) -> Result<i64, Error>

fn size_hint(_depth: usize) -> (usize, Option<usize>)

impl<'a> Arbitrary<'a> for i128

fn arbitrary(u: &mut Unstructured<'a>) -> Result<i128, Error>

fn size_hint(_depth: usize) -> (usize, Option<usize>)

impl<'a> Arbitrary<'a> for isize

fn arbitrary(u: &mut Unstructured<'a>) -> Result<isize, Error>

fn size_hint(depth: usize) -> (usize, Option<usize>)

impl<'a> Arbitrary<'a> for u8

fn arbitrary(u: &mut Unstructured<'a>) -> Result<u8, Error>

fn size_hint(_depth: usize) -> (usize, Option<usize>)

impl<'a> Arbitrary<'a> for u16

fn arbitrary(u: &mut Unstructured<'a>) -> Result<u16, Error>

fn size_hint(_depth: usize) -> (usize, Option<usize>)

impl<'a> Arbitrary<'a> for u32

fn arbitrary(u: &mut Unstructured<'a>) -> Result<u32, Error>

fn size_hint(_depth: usize) -> (usize, Option<usize>)

impl<'a> Arbitrary<'a> for u64

fn arbitrary(u: &mut Unstructured<'a>) -> Result<u64, Error>

fn size_hint(_depth: usize) -> (usize, Option<usize>)

impl<'a> Arbitrary<'a> for u128

fn arbitrary(u: &mut Unstructured<'a>) -> Result<u128, Error>

fn size_hint(_depth: usize) -> (usize, Option<usize>)

Trait Arbitrary

§Deriving `Arbitrary`

§Implementing `Arbitrary` By Hand

§When to implement `try_size_hint`

§The `depth` parameter