Auto merge of #118405 - matthiaskrgr:rollup-3a2eevc, r=matthiaskrgr

Rollup of 7 pull requests

Successful merges:

 - #115331 (optimize str::iter::Chars::advance_by)
 - #118236 (Update mod comment)
 - #118299 (Update `OnceLock` documentation to give a concrete 'lazy static' example, and expand on the existing example.)
 - #118314 (Rename `{collections=>alloc}{tests,benches}`)
 - #118341 (Simplify indenting in THIR printing)
 - #118366 (Detect and reject malformed `repr(Rust)` hints)
 - #118397 (Fix comments for unsigned non-zero `checked_add`, `saturating_add`)

r? `@ghost`
`@rustbot` modify labels: rollup

compiler/rustc_attr/src/builtin.rs

@@ -985,7 +985,7 @@ pub fn parse_repr_attr(sess: &Session, attr: &Attribute) -> Vec<ReprAttr> {
                         Ok(literal) => acc.push(ReprPacked(literal)),
                         Err(message) => literal_error = Some(message),
                     };
-                } else if matches!(name, sym::C | sym::simd | sym::transparent)
+                } else if matches!(name, sym::Rust | sym::C | sym::simd | sym::transparent)
                     || int_type_of_word(name).is_some()
                 {
                     recognised = true;
@@ -1018,7 +1018,7 @@ pub fn parse_repr_attr(sess: &Session, attr: &Attribute) -> Vec<ReprAttr> {
                             });
                         } else if matches!(
                             meta_item.name_or_empty(),
-                            sym::C | sym::simd | sym::transparent
+                            sym::Rust | sym::C | sym::simd | sym::transparent
                         ) || int_type_of_word(meta_item.name_or_empty()).is_some()
                         {
                             recognised = true;
@@ -1043,7 +1043,7 @@ pub fn parse_repr_attr(sess: &Session, attr: &Attribute) -> Vec<ReprAttr> {
                             );
                         } else if matches!(
                             meta_item.name_or_empty(),
-                            sym::C | sym::simd | sym::transparent
+                            sym::Rust | sym::C | sym::simd | sym::transparent
                         ) || int_type_of_word(meta_item.name_or_empty()).is_some()
                         {
                             recognised = true;

compiler/rustc_mir_build/src/thir/print.rs

@@ -31,8 +31,8 @@ const INDENT: &str = "    ";
 
 macro_rules! print_indented {
     ($writer:ident, $s:expr, $indent_lvl:expr) => {
-        let indent = (0..$indent_lvl).map(|_| INDENT).collect::<Vec<_>>().concat();
-        writeln!($writer, "{}{}", indent, $s).expect("unable to write to ThirPrinter");
+        $writer.indent($indent_lvl);
+        writeln!($writer, "{}", $s).expect("unable to write to ThirPrinter");
     };
 }
 
@@ -48,6 +48,12 @@ impl<'a, 'tcx> ThirPrinter<'a, 'tcx> {
         Self { thir, fmt: String::new() }
     }
 
+    fn indent(&mut self, level: usize) {
+        for _ in 0..level {
+            self.fmt.push_str(INDENT);
+        }
+    }
+
     fn print(&mut self) {
         print_indented!(self, "params: [", 0);
         for param in self.thir.params.iter() {

library/alloc/Cargo.toml

@@ -17,11 +17,11 @@ rand = { version = "0.8.5", default-features = false, features = ["alloc"] }
 rand_xorshift = "0.3.0"
 
 [[test]]
-name = "collectionstests"
+name = "alloctests"
 path = "tests/lib.rs"
 
 [[bench]]
-name = "collectionsbenches"
+name = "allocbenches"
 path = "benches/lib.rs"
 test = true
 

library/alloc/tests/str.rs

@@ -1170,6 +1170,17 @@ fn test_iterator() {
     assert_eq!(s.chars().count(), v.len());
 }
 
+#[test]
+fn test_iterator_advance() {
+    let s = "「赤錆」と呼ばれる鉄錆は、水の存在下での鉄の自然酸化によって生じる、オキシ水酸化鉄(III) 等の（含水）酸化物粒子の疎な凝集膜であるとみなせる。";
+    let chars: Vec<char> = s.chars().collect();
+    let mut it = s.chars();
+    it.advance_by(1).unwrap();
+    assert_eq!(it.next(), Some(chars[1]));
+    it.advance_by(33).unwrap();
+    assert_eq!(it.next(), Some(chars[35]));
+}
+
 #[test]
 fn test_rev_iterator() {
     let s = "ศไทย中华Việt Nam";

library/core/benches/lib.rs

@@ -5,6 +5,7 @@
 #![feature(trusted_random_access)]
 #![feature(iter_array_chunks)]
 #![feature(iter_next_chunk)]
+#![feature(iter_advance_by)]
 
 extern crate test;
 

library/core/benches/str.rs

@@ -3,6 +3,7 @@ use test::{black_box, Bencher};
 
 mod char_count;
 mod corpora;
+mod iter;
 
 #[bench]
 fn str_validate_emoji(b: &mut Bencher) {

library/core/benches/str/iter.rs

@@ -0,0 +1,17 @@
+use super::corpora;
+use test::{black_box, Bencher};
+
+#[bench]
+fn chars_advance_by_1000(b: &mut Bencher) {
+    b.iter(|| black_box(corpora::ru::LARGE).chars().advance_by(1000));
+}
+
+#[bench]
+fn chars_advance_by_0010(b: &mut Bencher) {
+    b.iter(|| black_box(corpora::ru::LARGE).chars().advance_by(10));
+}
+
+#[bench]
+fn chars_advance_by_0001(b: &mut Bencher) {
+    b.iter(|| black_box(corpora::ru::LARGE).chars().advance_by(1));
+}

library/core/src/num/mod.rs

@@ -474,7 +474,7 @@ impl isize {
     }
 }
 
-/// If 6th bit is set ascii is lower case.
+/// If the 6th bit is set ascii is lower case.
 const ASCII_CASE_MASK: u8 = 0b0010_0000;
 
 impl u8 {
@@ -549,7 +549,7 @@ impl u8 {
     #[rustc_const_stable(feature = "const_ascii_methods_on_intrinsics", since = "1.52.0")]
     #[inline]
     pub const fn to_ascii_uppercase(&self) -> u8 {
-        // Toggle the fifth bit if this is a lowercase letter
+        // Toggle the 6th bit if this is a lowercase letter
         *self ^ ((self.is_ascii_lowercase() as u8) * ASCII_CASE_MASK)
     }
 
@@ -574,7 +574,7 @@ impl u8 {
     #[rustc_const_stable(feature = "const_ascii_methods_on_intrinsics", since = "1.52.0")]
     #[inline]
     pub const fn to_ascii_lowercase(&self) -> u8 {
-        // Set the fifth bit if this is an uppercase letter
+        // Set the 6th bit if this is an uppercase letter
         *self | (self.is_ascii_uppercase() as u8 * ASCII_CASE_MASK)
     }
 

library/core/src/num/nonzero.rs

@@ -355,7 +355,7 @@ macro_rules! nonzero_unsigned_operations {
                     if let Some(result) = self.get().checked_add(other) {
                         // SAFETY:
                         // - `checked_add` returns `None` on overflow
-                        // - `self` and `other` are non-zero
+                        // - `self` is non-zero
                         // - the only way to get zero from an addition without overflow is for both
                         //   sides to be zero
                         //
@@ -393,7 +393,7 @@ macro_rules! nonzero_unsigned_operations {
                 pub const fn saturating_add(self, other: $Int) -> $Ty {
                     // SAFETY:
                     // - `saturating_add` returns `u*::MAX` on overflow, which is non-zero
-                    // - `self` and `other` are non-zero
+                    // - `self` is non-zero
                     // - the only way to get zero from an addition without overflow is for both
                     //   sides to be zero
                     //

library/core/src/str/iter.rs

@@ -8,6 +8,7 @@ use crate::iter::{TrustedRandomAccess, TrustedRandomAccessNoCoerce};
 use crate::ops::Try;
 use crate::option;
 use crate::slice::{self, Split as SliceSplit};
+use core::num::NonZeroUsize;
 
 use super::from_utf8_unchecked;
 use super::pattern::Pattern;
@@ -49,6 +50,55 @@ impl<'a> Iterator for Chars<'a> {
         super::count::count_chars(self.as_str())
     }
 
+    #[inline]
+    fn advance_by(&mut self, mut remainder: usize) -> Result<(), NonZeroUsize> {
+        const CHUNK_SIZE: usize = 32;
+
+        if remainder >= CHUNK_SIZE {
+            let mut chunks = self.iter.as_slice().array_chunks::<CHUNK_SIZE>();
+            let mut bytes_skipped: usize = 0;
+
+            while remainder > CHUNK_SIZE
+                && let Some(chunk) = chunks.next()
+            {
+                bytes_skipped += CHUNK_SIZE;
+
+                let mut start_bytes = [false; CHUNK_SIZE];
+
+                for i in 0..CHUNK_SIZE {
+                    start_bytes[i] = !super::validations::utf8_is_cont_byte(chunk[i]);
+                }
+
+                remainder -= start_bytes.into_iter().map(|i| i as u8).sum::<u8>() as usize;
+            }
+
+            // SAFETY: The amount of bytes exists since we just iterated over them,
+            // so advance_by will succeed.
+            unsafe { self.iter.advance_by(bytes_skipped).unwrap_unchecked() };
+
+            // skip trailing continuation bytes
+            while self.iter.len() > 0 {
+                let b = self.iter.as_slice()[0];
+                if !super::validations::utf8_is_cont_byte(b) {
+                    break;
+                }
+                // SAFETY: We just peeked at the byte, therefore it exists
+                unsafe { self.iter.advance_by(1).unwrap_unchecked() };
+            }
+        }
+
+        while (remainder > 0) && (self.iter.len() > 0) {
+            remainder -= 1;
+            let b = self.iter.as_slice()[0];
+            let slurp = super::validations::utf8_char_width(b);
+            // SAFETY: utf8 validity requires that the string must contain
+            // the continuation bytes (if any)
+            unsafe { self.iter.advance_by(slurp).unwrap_unchecked() };
+        }
+
+        NonZeroUsize::new(remainder).map_or(Ok(()), Err)
+    }
+
     #[inline]
     fn size_hint(&self) -> (usize, Option<usize>) {
         let len = self.iter.len();

library/std/src/sync/once_lock.rs

@@ -13,22 +13,54 @@ use crate::sync::Once;
 ///
 /// # Examples
 ///
+/// Using `OnceCell` to store a function’s previously computed value (a.k.a.
+/// ‘lazy static’ or ‘memoizing’):
+///
+/// ```
+/// use std::collections::HashMap;
+/// use std::sync::OnceLock;
+///
+/// fn hash_map() -> &'static HashMap<u32, char> {
+///     static HASHMAP: OnceLock<HashMap<u32, char>> = OnceLock::new();
+///     HASHMAP.get_or_init(|| {
+///         let mut m = HashMap::new();
+///         m.insert(0, 'a');
+///         m.insert(1, 'b');
+///         m.insert(2, 'c');
+///         m
+///     })
+/// }
+///
+/// // The `HashMap` is built, stored in the `OnceLock`, and returned.
+/// let _ = hash_map();
+///
+/// // The `HashMap` is retrieved from the `OnceLock` and returned.
+/// let _ = hash_map();
+/// ```
+///
+/// Writing to a `OnceLock` from a separate thread:
+///
 /// ```
 /// use std::sync::OnceLock;
 ///
-/// static CELL: OnceLock<String> = OnceLock::new();
+/// static CELL: OnceLock<usize> = OnceLock::new();
+///
+/// // `OnceLock` has not been written to yet.
 /// assert!(CELL.get().is_none());
 ///
+/// // Spawn a thread and write to `OnceLock`.
 /// std::thread::spawn(|| {
-///     let value: &String = CELL.get_or_init(|| {
-///         "Hello, World!".to_string()
-///     });
-///     assert_eq!(value, "Hello, World!");
-/// }).join().unwrap();
+///     let value = CELL.get_or_init(|| 12345);
+///     assert_eq!(value, &12345);
+/// })
+/// .join()
+/// .unwrap();
 ///
-/// let value: Option<&String> = CELL.get();
-/// assert!(value.is_some());
-/// assert_eq!(value.unwrap().as_str(), "Hello, World!");
+/// // `OnceLock` now contains the value.
+/// assert_eq!(
+///     CELL.get(),
+///     Some(&12345),
+/// );
 /// ```
 #[stable(feature = "once_cell", since = "1.70.0")]
 pub struct OnceLock<T> {

tests/ui/repr/issue-83921-ice.rs → tests/ui/repr/malformed-repr-hints.rs

@@ -19,6 +19,15 @@ struct S3;
 //~^ ERROR: incorrect `repr(align)` attribute format
 struct S4;
 
+// Regression test for issue #118334:
+#[repr(Rust(u8))]
+//~^ ERROR: invalid representation hint
+#[repr(Rust(0))]
+//~^ ERROR: invalid representation hint
+#[repr(Rust = 0)]
+//~^ ERROR: invalid representation hint
+struct S5;
+
 #[repr(i8())]
 //~^ ERROR: invalid representation hint
 enum E1 { A, B }

tests/ui/repr/issue-83921-ice.stderr → tests/ui/repr/malformed-repr-hints.stderr

@@ -1,46 +1,64 @@
 error[E0552]: incorrect `repr(packed)` attribute format: `packed` takes exactly one parenthesized argument, or no parentheses at all
-  --> $DIR/issue-83921-ice.rs:6:8
+  --> $DIR/malformed-repr-hints.rs:6:8
    |
 LL | #[repr(packed())]
    |        ^^^^^^^^
 
 error[E0589]: invalid `repr(align)` attribute: `align` needs an argument
-  --> $DIR/issue-83921-ice.rs:10:8
+  --> $DIR/malformed-repr-hints.rs:10:8
    |
 LL | #[repr(align)]
    |        ^^^^^ help: supply an argument here: `align(...)`
 
 error[E0693]: incorrect `repr(align)` attribute format: `align` takes exactly one argument in parentheses
-  --> $DIR/issue-83921-ice.rs:14:8
+  --> $DIR/malformed-repr-hints.rs:14:8
    |
 LL | #[repr(align(2, 4))]
    |        ^^^^^^^^^^^
 
 error[E0693]: incorrect `repr(align)` attribute format: `align` takes exactly one argument in parentheses
-  --> $DIR/issue-83921-ice.rs:18:8
+  --> $DIR/malformed-repr-hints.rs:18:8
    |
 LL | #[repr(align())]
    |        ^^^^^^^
 
+error[E0552]: invalid representation hint: `Rust` does not take a parenthesized argument list
+  --> $DIR/malformed-repr-hints.rs:23:8
+   |
+LL | #[repr(Rust(u8))]
+   |        ^^^^^^^^
+
+error[E0552]: invalid representation hint: `Rust` does not take a parenthesized argument list
+  --> $DIR/malformed-repr-hints.rs:25:8
+   |
+LL | #[repr(Rust(0))]
+   |        ^^^^^^^
+
+error[E0552]: invalid representation hint: `Rust` does not take a value
+  --> $DIR/malformed-repr-hints.rs:27:8
+   |
+LL | #[repr(Rust = 0)]
+   |        ^^^^^^^^
+
 error[E0552]: invalid representation hint: `i8` does not take a parenthesized argument list
-  --> $DIR/issue-83921-ice.rs:22:8
+  --> $DIR/malformed-repr-hints.rs:31:8
    |
 LL | #[repr(i8())]
    |        ^^^^
 
 error[E0552]: invalid representation hint: `u32` does not take a parenthesized argument list
-  --> $DIR/issue-83921-ice.rs:26:8
+  --> $DIR/malformed-repr-hints.rs:35:8
    |
 LL | #[repr(u32(42))]
    |        ^^^^^^^
 
 error[E0552]: invalid representation hint: `i64` does not take a value
-  --> $DIR/issue-83921-ice.rs:30:8
+  --> $DIR/malformed-repr-hints.rs:39:8
    |
 LL | #[repr(i64 = 2)]
    |        ^^^^^^^
 
-error: aborting due to 7 previous errors
+error: aborting due to 10 previous errors
 
 Some errors have detailed explanations: E0552, E0589, E0693.
 For more information about an error, try `rustc --explain E0552`.