Auto merge of rust-lang#111222 - scottmcm:constify-is_ascii, r=thomcc

Constify `[u8]::is_ascii` (unstably)

UTF-8 checking in `const fn`-stabilized back in 1.63 (rust-lang#97367), but apparently somehow ASCII checking was never const-ified, despite being simpler.

New constness-tracking issue for `is_ascii`: rust-lang#111090

I noticed this working on `ascii::Char`: rust-lang#110998

library/core/src/array/ascii.rs

@@ -4,10 +4,23 @@ use crate::ascii;
 impl<const N: usize> [u8; N] {
     /// Converts this array of bytes into a array of ASCII characters,
     /// or returns `None` if any of the characters is non-ASCII.
+    ///
+    /// # Examples
+    ///
+    /// ```
+    /// #![feature(ascii_char)]
+    /// #![feature(const_option)]
+    ///
+    /// const HEX_DIGITS: [std::ascii::Char; 16] =
+    ///     *b"0123456789abcdef".as_ascii().unwrap();
+    ///
+    /// assert_eq!(HEX_DIGITS[1].as_str(), "1");
+    /// assert_eq!(HEX_DIGITS[10].as_str(), "a");
+    /// ```
     #[unstable(feature = "ascii_char", issue = "110998")]
     #[must_use]
     #[inline]
-    pub fn as_ascii(&self) -> Option<&[ascii::Char; N]> {
+    pub const fn as_ascii(&self) -> Option<&[ascii::Char; N]> {
         if self.is_ascii() {
             // SAFETY: Just checked that it's ASCII
             Some(unsafe { self.as_ascii_unchecked() })

library/core/src/lib.rs

@@ -150,6 +150,7 @@
 #![feature(const_slice_from_raw_parts_mut)]
 #![feature(const_slice_from_ref)]
 #![feature(const_slice_index)]
+#![feature(const_slice_is_ascii)]
 #![feature(const_slice_ptr_len)]
 #![feature(const_slice_split_at_mut)]
 #![feature(const_str_from_utf8_unchecked_mut)]

library/core/src/slice/ascii.rs

@@ -10,9 +10,10 @@ use crate::ops;
 impl [u8] {
     /// Checks if all bytes in this slice are within the ASCII range.
     #[stable(feature = "ascii_methods_on_intrinsics", since = "1.23.0")]
+    #[rustc_const_unstable(feature = "const_slice_is_ascii", issue = "111090")]
     #[must_use]
     #[inline]
-    pub fn is_ascii(&self) -> bool {
+    pub const fn is_ascii(&self) -> bool {
         is_ascii(self)
     }
 
@@ -21,7 +22,7 @@ impl [u8] {
     #[unstable(feature = "ascii_char", issue = "110998")]
     #[must_use]
     #[inline]
-    pub fn as_ascii(&self) -> Option<&[ascii::Char]> {
+    pub const fn as_ascii(&self) -> Option<&[ascii::Char]> {
         if self.is_ascii() {
             // SAFETY: Just checked that it's ASCII
             Some(unsafe { self.as_ascii_unchecked() })
@@ -262,11 +263,29 @@ impl<'a> fmt::Debug for EscapeAscii<'a> {
 /// Returns `true` if any byte in the word `v` is nonascii (>= 128). Snarfed
 /// from `../str/mod.rs`, which does something similar for utf8 validation.
 #[inline]
-fn contains_nonascii(v: usize) -> bool {
+const fn contains_nonascii(v: usize) -> bool {
     const NONASCII_MASK: usize = usize::repeat_u8(0x80);
     (NONASCII_MASK & v) != 0
 }
 
+/// ASCII test *without* the chunk-at-a-time optimizations.
+///
+/// This is carefully structured to produce nice small code -- it's smaller in
+/// `-O` than what the "obvious" ways produces under `-C opt-level=s`.  If you
+/// touch it, be sure to run (and update if needed) the assembly test.
+#[unstable(feature = "str_internals", issue = "none")]
+#[doc(hidden)]
+#[inline]
+pub const fn is_ascii_simple(mut bytes: &[u8]) -> bool {
+    while let [rest @ .., last] = bytes {
+        if !last.is_ascii() {
+            break;
+        }
+        bytes = rest;
+    }
+    bytes.is_empty()
+}
+
 /// Optimized ASCII test that will use usize-at-a-time operations instead of
 /// byte-at-a-time operations (when possible).
 ///
@@ -280,7 +299,7 @@ fn contains_nonascii(v: usize) -> bool {
 /// If any of these loads produces something for which `contains_nonascii`
 /// (above) returns true, then we know the answer is false.
 #[inline]
-fn is_ascii(s: &[u8]) -> bool {
+const fn is_ascii(s: &[u8]) -> bool {
     const USIZE_SIZE: usize = mem::size_of::<usize>();
 
     let len = s.len();
@@ -292,7 +311,7 @@ fn is_ascii(s: &[u8]) -> bool {
     // We also do this for architectures where `size_of::<usize>()` isn't
     // sufficient alignment for `usize`, because it's a weird edge case.
     if len < USIZE_SIZE || len < align_offset || USIZE_SIZE < mem::align_of::<usize>() {
-        return s.iter().all(|b| b.is_ascii());
+        return is_ascii_simple(s);
     }
 
     // We always read the first word unaligned, which means `align_offset` is
@@ -321,18 +340,26 @@ fn is_ascii(s: &[u8]) -> bool {
     // Paranoia check about alignment, since we're about to do a bunch of
     // unaligned loads. In practice this should be impossible barring a bug in
     // `align_offset` though.
-    debug_assert_eq!(word_ptr.addr() % mem::align_of::<usize>(), 0);
+    // While this method is allowed to spuriously fail in CTFE, if it doesn't
+    // have alignment information it should have given a `usize::MAX` for
+    // `align_offset` earlier, sending things through the scalar path instead of
+    // this one, so this check should pass if it's reachable.
+    debug_assert!(word_ptr.is_aligned_to(mem::align_of::<usize>()));
 
     // Read subsequent words until the last aligned word, excluding the last
     // aligned word by itself to be done in tail check later, to ensure that
     // tail is always one `usize` at most to extra branch `byte_pos == len`.
     while byte_pos < len - USIZE_SIZE {
-        debug_assert!(
-            // Sanity check that the read is in bounds
-            (word_ptr.addr() + USIZE_SIZE) <= start.addr().wrapping_add(len) &&
-            // And that our assumptions about `byte_pos` hold.
-            (word_ptr.addr() - start.addr()) == byte_pos
-        );
+        // Sanity check that the read is in bounds
+        debug_assert!(byte_pos + USIZE_SIZE <= len);
+        // And that our assumptions about `byte_pos` hold.
+        debug_assert!(matches!(
+            word_ptr.cast::<u8>().guaranteed_eq(start.wrapping_add(byte_pos)),
+            // These are from the same allocation, so will hopefully always be
+            // known to match even in CTFE, but if it refuses to compare them
+            // that's ok since it's just a debug check anyway.
+            None | Some(true),
+        ));
 
         // SAFETY: We know `word_ptr` is properly aligned (because of
         // `align_offset`), and we know that we have enough bytes between `word_ptr` and the end

library/core/src/slice/mod.rs

@@ -44,6 +44,10 @@ mod raw;
 mod rotate;
 mod specialize;
 
+#[unstable(feature = "str_internals", issue = "none")]
+#[doc(hidden)]
+pub use ascii::is_ascii_simple;
+
 #[stable(feature = "rust1", since = "1.0.0")]
 pub use iter::{Chunks, ChunksMut, Windows};
 #[stable(feature = "rust1", since = "1.0.0")]

library/core/src/str/mod.rs

@@ -2358,9 +2358,10 @@ impl str {
     /// assert!(!non_ascii.is_ascii());
     /// ```
     #[stable(feature = "ascii_methods_on_intrinsics", since = "1.23.0")]
+    #[rustc_const_unstable(feature = "const_slice_is_ascii", issue = "111090")]
     #[must_use]
     #[inline]
-    pub fn is_ascii(&self) -> bool {
+    pub const fn is_ascii(&self) -> bool {
         // We can treat each byte as character here: all multibyte characters
         // start with a byte that is not in the ASCII range, so we will stop
         // there already.
@@ -2372,7 +2373,7 @@ impl str {
     #[unstable(feature = "ascii_char", issue = "110998")]
     #[must_use]
     #[inline]
-    pub fn as_ascii(&self) -> Option<&[ascii::Char]> {
+    pub const fn as_ascii(&self) -> Option<&[ascii::Char]> {
         // Like in `is_ascii`, we can work on the bytes directly.
         self.as_bytes().as_ascii()
     }

tests/assembly/slice-is_ascii.rs

@@ -0,0 +1,35 @@
+// revisions: WIN LIN
+// [WIN] only-windows
+// [LIN] only-linux
+// assembly-output: emit-asm
+// compile-flags: --crate-type=lib -O -C llvm-args=-x86-asm-syntax=intel
+// min-llvm-version: 14
+// only-x86_64
+// ignore-sgx
+// ignore-debug
+
+#![feature(str_internals)]
+
+// CHECK-LABEL: is_ascii_simple_demo:
+#[no_mangle]
+pub fn is_ascii_simple_demo(bytes: &[u8]) -> bool {
+    // Linux (System V): pointer is rdi; length is rsi
+    // Windows: pointer is rcx; length is rdx.
+
+    // CHECK-NOT: mov
+    // CHECK-NOT: test
+    // CHECK-NOT: cmp
+
+    // CHECK: .[[LOOPHEAD:.+]]:
+    // CHECK-NEXT: mov [[TEMP:.+]], [[LEN:rsi|rdx]]
+    // CHECK-NEXT: sub [[LEN]], 1
+    // CHECK-NEXT: jb .[[LOOPEXIT:.+]]
+    // CHECK-NEXT: cmp byte ptr [{{rdi|rcx}} + [[TEMP]] - 1], 0
+    // CHECK-NEXT: jns .[[LOOPHEAD]]
+
+    // CHECK-NEXT: .[[LOOPEXIT]]:
+    // CHECK-NEXT: test [[TEMP]], [[TEMP]]
+    // CHECK-NEXT: sete al
+    // CHECK-NEXT: ret
+    core::slice::is_ascii_simple(bytes)
+}