Trait regex_automata::StateID

pub unsafe trait StateID: Clone + Copy + Debug + Eq + Hash + PartialEq + PartialOrd + Ord {
    // Required methods
    fn from_usize(n: usize) -> Self;
    fn to_usize(self) -> usize;
    fn max_id() -> usize;
    fn read_bytes(slice: &[u8]) -> Self;
    fn write_bytes(self, slice: &mut [u8]);
}

A trait describing the representation of a DFA’s state identifier.

The purpose of this trait is to safely express both the possible state identifier representations that can be used in a DFA and to convert between state identifier representations and types that can be used to efficiently index memory (such as usize).

In general, one should not need to implement this trait explicitly. In particular, this crate provides implementations for u8, u16, u32, u64 and usize. (u32 and u64 are only provided for targets that can represent all corresponding values in a usize.)

Safety

This trait is unsafe because the correctness of its implementations may be relied upon by other unsafe code. For example, one possible way to implement this trait incorrectly would be to return a maximum identifier in max_id that is greater than the real maximum identifier. This will likely result in wrap-on-overflow semantics in release mode, which can in turn produce incorrect state identifiers. Those state identifiers may then in turn access out-of-bounds memory in a DFA’s search routine, where bounds checks are explicitly elided for performance reasons.

Required Methods

fn from_usize(n: usize) -> Self

Convert from a usize to this implementation’s representation.

Implementors may assume that n <= Self::max_id. That is, implementors do not need to check whether n can fit inside this implementation’s representation.

fn to_usize(self) -> usize

Convert this implementation’s representation to a usize.

Implementors must not return a usize value greater than Self::max_id and must not permit overflow when converting between the implementor’s representation and usize. In general, the preferred way for implementors to achieve this is to simply not provide implementations of StateID that cannot fit into the target platform’s usize.

fn max_id() -> usize

Return the maximum state identifier supported by this representation.

Implementors must return a correct bound. Doing otherwise may result in memory unsafety.

fn read_bytes(slice: &[u8]) -> Self

Read a single state identifier from the given slice of bytes in native endian format.

Implementors may assume that the given slice has length at least size_of::<Self>().

fn write_bytes(self, slice: &mut [u8])

Write this state identifier to the given slice of bytes in native endian format.

Implementors may assume that the given slice has length at least size_of::<Self>().

Implementations on Foreign Types

impl StateID for u32

fn from_usize(n: usize) -> u32

fn to_usize(self) -> usize

fn max_id() -> usize

fn read_bytes(slice: &[u8]) -> Self

fn write_bytes(self, slice: &mut [u8])

impl StateID for u16

fn from_usize(n: usize) -> u16

fn to_usize(self) -> usize

fn max_id() -> usize

fn read_bytes(slice: &[u8]) -> Self

fn write_bytes(self, slice: &mut [u8])

impl StateID for usize

fn from_usize(n: usize) -> usize

fn to_usize(self) -> usize

fn max_id() -> usize

fn read_bytes(slice: &[u8]) -> Self

fn write_bytes(self, slice: &mut [u8])

impl StateID for u8

fn from_usize(n: usize) -> u8

fn to_usize(self) -> usize

fn max_id() -> usize

fn read_bytes(slice: &[u8]) -> Self

fn write_bytes(self, slice: &mut [u8])

impl StateID for u64

fn from_usize(n: usize) -> u64

fn to_usize(self) -> usize

fn max_id() -> usize

fn read_bytes(slice: &[u8]) -> Self

fn write_bytes(self, slice: &mut [u8])

Implementors