Rollup merge of #131520 - zachs18:const-str-split, r=Noratrieb

Mark `str::is_char_boundary` and `str::split_at*` unstably `const`.

Tracking issues: #131516, #131518

First commit implements `const_is_char_boundary`, second commit implements `const_str_split_at` (which depends on `const_is_char_boundary`)

~~I used `const_eval_select` for `is_char_boundary` since there is a comment about optimizations that would theoretically not happen with the simple `const`-compatible version (since `slice::get` is not `const`ifiable) cc #84751. I have not checked if this code difference is still required for the optimization, so it might not be worth the code complication, but 🤷.~~

This changes `str::split_at_checked` to use a new private helper function `split_at_unchecked` (copied from `split_at_mut_unchecked`) that does pointer stuff instead of `get_unchecked`, since that is not currently `const`ifiable due to using the `SliceIndex` trait.
This commit is contained in:
Jubilee 2024-10-29 03:11:39 -07:00 committed by GitHub
commit b496974c53
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
2 changed files with 38 additions and 12 deletions

View File

@ -185,7 +185,9 @@
#![feature(cfg_target_has_atomic_equal_alignment)]
#![feature(cfg_ub_checks)]
#![feature(const_for)]
#![feature(const_is_char_boundary)]
#![feature(const_precise_live_drops)]
#![feature(const_str_split_at)]
#![feature(decl_macro)]
#![feature(deprecated_suggestion)]
#![feature(doc_cfg)]

View File

@ -185,8 +185,9 @@ pub const fn is_empty(&self) -> bool {
/// ```
#[must_use]
#[stable(feature = "is_char_boundary", since = "1.9.0")]
#[rustc_const_unstable(feature = "const_is_char_boundary", issue = "131516")]
#[inline]
pub fn is_char_boundary(&self, index: usize) -> bool {
pub const fn is_char_boundary(&self, index: usize) -> bool {
// 0 is always ok.
// Test for 0 explicitly so that it can optimize out the check
// easily and skip reading string data for that case.
@ -195,8 +196,8 @@ pub fn is_char_boundary(&self, index: usize) -> bool {
return true;
}
match self.as_bytes().get(index) {
// For `None` we have two options:
if index >= self.len() {
// For `true` we have two options:
//
// - index == self.len()
// Empty strings are valid, so return true
@ -205,9 +206,9 @@ pub fn is_char_boundary(&self, index: usize) -> bool {
//
// The check is placed exactly here, because it improves generated
// code on higher opt-levels. See PR #84751 for more details.
None => index == self.len(),
Some(&b) => b.is_utf8_char_boundary(),
index == self.len()
} else {
self.as_bytes()[index].is_utf8_char_boundary()
}
}
@ -637,7 +638,8 @@ pub unsafe fn slice_mut_unchecked(&mut self, begin: usize, end: usize) -> &mut s
#[inline]
#[must_use]
#[stable(feature = "str_split_at", since = "1.4.0")]
pub fn split_at(&self, mid: usize) -> (&str, &str) {
#[rustc_const_unstable(feature = "const_str_split_at", issue = "131518")]
pub const fn split_at(&self, mid: usize) -> (&str, &str) {
match self.split_at_checked(mid) {
None => slice_error_fail(self, 0, mid),
Some(pair) => pair,
@ -677,7 +679,8 @@ pub fn split_at(&self, mid: usize) -> (&str, &str) {
#[inline]
#[must_use]
#[stable(feature = "str_split_at", since = "1.4.0")]
pub fn split_at_mut(&mut self, mid: usize) -> (&mut str, &mut str) {
#[rustc_const_unstable(feature = "const_str_split_at", issue = "131518")]
pub const fn split_at_mut(&mut self, mid: usize) -> (&mut str, &mut str) {
// is_char_boundary checks that the index is in [0, .len()]
if self.is_char_boundary(mid) {
// SAFETY: just checked that `mid` is on a char boundary.
@ -716,11 +719,12 @@ pub fn split_at_mut(&mut self, mid: usize) -> (&mut str, &mut str) {
#[inline]
#[must_use]
#[stable(feature = "split_at_checked", since = "1.80.0")]
pub fn split_at_checked(&self, mid: usize) -> Option<(&str, &str)> {
#[rustc_const_unstable(feature = "const_str_split_at", issue = "131518")]
pub const fn split_at_checked(&self, mid: usize) -> Option<(&str, &str)> {
// is_char_boundary checks that the index is in [0, .len()]
if self.is_char_boundary(mid) {
// SAFETY: just checked that `mid` is on a char boundary.
Some(unsafe { (self.get_unchecked(0..mid), self.get_unchecked(mid..self.len())) })
Some(unsafe { self.split_at_unchecked(mid) })
} else {
None
}
@ -756,7 +760,9 @@ pub fn split_at_checked(&self, mid: usize) -> Option<(&str, &str)> {
#[inline]
#[must_use]
#[stable(feature = "split_at_checked", since = "1.80.0")]
pub fn split_at_mut_checked(&mut self, mid: usize) -> Option<(&mut str, &mut str)> {
#[rustc_const_unstable(feature = "const_str_split_at", issue = "131518")]
#[rustc_allow_const_fn_unstable(const_is_char_boundary)]
pub const fn split_at_mut_checked(&mut self, mid: usize) -> Option<(&mut str, &mut str)> {
// is_char_boundary checks that the index is in [0, .len()]
if self.is_char_boundary(mid) {
// SAFETY: just checked that `mid` is on a char boundary.
@ -772,7 +778,25 @@ pub fn split_at_mut_checked(&mut self, mid: usize) -> Option<(&mut str, &mut str
///
/// The caller must ensure that `mid` is a valid byte offset from the start
/// of the string and falls on the boundary of a UTF-8 code point.
unsafe fn split_at_mut_unchecked(&mut self, mid: usize) -> (&mut str, &mut str) {
const unsafe fn split_at_unchecked(&self, mid: usize) -> (&str, &str) {
let len = self.len();
let ptr = self.as_ptr();
// SAFETY: caller guarantees `mid` is on a char boundary.
unsafe {
(
from_utf8_unchecked(slice::from_raw_parts(ptr, mid)),
from_utf8_unchecked(slice::from_raw_parts(ptr.add(mid), len - mid)),
)
}
}
/// Divides one string slice into two at an index.
///
/// # Safety
///
/// The caller must ensure that `mid` is a valid byte offset from the start
/// of the string and falls on the boundary of a UTF-8 code point.
const unsafe fn split_at_mut_unchecked(&mut self, mid: usize) -> (&mut str, &mut str) {
let len = self.len();
let ptr = self.as_mut_ptr();
// SAFETY: caller guarantees `mid` is on a char boundary.