Struct memchr::memmem::FinderRev

pub struct FinderRev<'n> {
    needle: CowBytes<'n>,
    searcher: SearcherRev,
}

A single substring reverse searcher fixed to a particular needle.

The purpose of this type is to permit callers to construct a substring searcher that can be used to search haystacks without the overhead of constructing the searcher in the first place. This is a somewhat niche concern when it’s necessary to re-use the same needle to search multiple different haystacks with as little overhead as possible. In general, using rfind is good enough, but FinderRev is useful when you can meaningfully observe searcher construction time in a profile.

When the std feature is enabled, then this type has an into_owned version which permits building a FinderRev that is not connected to the lifetime of its needle.

Fields

needle: CowBytes<'n>

searcher: SearcherRev

Implementations

impl<'n> FinderRev<'n>

pub fn new<B: ?Sized + AsRef<[u8]>>(needle: &'n B) -> FinderRev<'n>

Create a new reverse finder for the given needle.

pub fn rfind<B: AsRef<[u8]>>(&self, haystack: B) -> Option<usize>

Returns the index of the last occurrence of this needle in the given haystack.

The haystack may be any type that can be cheaply converted into a &[u8]. This includes, but is not limited to, &str and &[u8].

Complexity

This routine is guaranteed to have worst case linear time complexity with respect to both the needle and the haystack. That is, this runs in O(needle.len() + haystack.len()) time.

This routine is also guaranteed to have worst case constant space complexity.

Examples

Basic usage:

use memchr::memmem::FinderRev;

let haystack = b"foo bar baz";
assert_eq!(Some(0), FinderRev::new("foo").rfind(haystack));
assert_eq!(Some(4), FinderRev::new("bar").rfind(haystack));
assert_eq!(None, FinderRev::new("quux").rfind(haystack));

pub fn rfind_iter<'a, 'h>(&'a self, haystack: &'h [u8]) -> FindRevIter<'h, 'a>

Returns a reverse iterator over all occurrences of a substring in a haystack.

Complexity

This routine is guaranteed to have worst case linear time complexity with respect to both the needle and the haystack. That is, this runs in O(needle.len() + haystack.len()) time.

This routine is also guaranteed to have worst case constant space complexity.

Examples

Basic usage:

use memchr::memmem::FinderRev;

let haystack = b"foo bar foo baz foo";
let finder = FinderRev::new(b"foo");
let mut it = finder.rfind_iter(haystack);
assert_eq!(Some(16), it.next());
assert_eq!(Some(8), it.next());
assert_eq!(Some(0), it.next());
assert_eq!(None, it.next());

pub fn into_owned(self) -> FinderRev<'static>

Convert this finder into its owned variant, such that it no longer borrows the needle.

If this is already an owned finder, then this is a no-op. Otherwise, this copies the needle.

This is only available when the std feature is enabled.

pub fn as_ref(&self) -> FinderRev<'_>

Convert this finder into its borrowed variant.

This is primarily useful if your finder is owned and you’d like to store its borrowed variant in some intermediate data structure.

Note that the lifetime parameter of the returned finder is tied to the lifetime of self, and may be shorter than the 'n lifetime of the needle itself. Namely, a finder’s needle can be either borrowed or owned, so the lifetime of the needle returned must necessarily be the shorter of the two.

pub fn needle(&self) -> &[u8]

Returns the needle that this finder searches for.

Note that the lifetime of the needle returned is tied to the lifetime of the finder, and may be shorter than the 'n lifetime. Namely, a finder’s needle can be either borrowed or owned, so the lifetime of the needle returned must necessarily be the shorter of the two.

Trait Implementations

impl<'n> Clone for FinderRev<'n>

fn clone(&self) -> FinderRev<'n>

Returns a copy of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl<'n> Debug for FinderRev<'n>

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.