Rollup merge of #95251 - GrishaVar:hashes-u16-to-u8, r=dtolnay

Reduce max hash in raw strings from u16 to u8

[Relevant discussion](https://rust-lang.zulipchat.com/#narrow/stream/237824-t-lang.2Fdoc/topic/Max.20raw.20string.20delimiters)

Author: Dylan-DPC

compiler/rustc_ast/src/ast.rs

@@ -1616,7 +1616,7 @@ pub enum StrStyle {
     /// A raw string, like `r##"foo"##`.
     ///
     /// The value is the number of `#` symbols used.
-    Raw(u16),
+    Raw(u8),
 }
 
 /// An AST literal.

compiler/rustc_ast/src/token.rs

@@ -59,9 +59,9 @@ pub enum LitKind {
     Integer,
     Float,
     Str,
-    StrRaw(u16), // raw string delimited by `n` hash symbols
+    StrRaw(u8), // raw string delimited by `n` hash symbols
     ByteStr,
-    ByteStrRaw(u16), // raw byte string delimited by `n` hash symbols
+    ByteStrRaw(u8), // raw byte string delimited by `n` hash symbols
     Err,
 }
 

compiler/rustc_lexer/src/lib.rs

@@ -161,15 +161,15 @@ pub enum LiteralKind {
     /// "b"abc"", "b"abc"
     ByteStr { terminated: bool },
     /// "r"abc"", "r#"abc"#", "r####"ab"###"c"####", "r#"a"
-    RawStr { n_hashes: u16, err: Option<RawStrError> },
+    RawStr { n_hashes: u8, err: Option<RawStrError> },
     /// "br"abc"", "br#"abc"#", "br####"ab"###"c"####", "br#"a"
-    RawByteStr { n_hashes: u16, err: Option<RawStrError> },
+    RawByteStr { n_hashes: u8, err: Option<RawStrError> },
 }
 
 /// Error produced validating a raw string. Represents cases like:
 /// - `r##~"abcde"##`: `InvalidStarter`
 /// - `r###"abcde"##`: `NoTerminator { expected: 3, found: 2, possible_terminator_offset: Some(11)`
-/// - Too many `#`s (>65535): `TooManyDelimiters`
+/// - Too many `#`s (>255): `TooManyDelimiters`
 // perf note: It doesn't matter that this makes `Token` 36 bytes bigger. See #77629
 #[derive(Clone, Copy, Debug, PartialEq, Eq, PartialOrd, Ord)]
 pub enum RawStrError {
@@ -178,7 +178,7 @@ pub enum RawStrError {
     /// The string was never terminated. `possible_terminator_offset` is the number of characters after `r` or `br` where they
     /// may have intended to terminate it.
     NoTerminator { expected: usize, found: usize, possible_terminator_offset: Option<usize> },
-    /// More than 65535 `#`s exist.
+    /// More than 255 `#`s exist.
     TooManyDelimiters { found: usize },
 }
 
@@ -698,12 +698,12 @@ impl Cursor<'_> {
     }
 
     /// Eats the double-quoted string and returns `n_hashes` and an error if encountered.
-    fn raw_double_quoted_string(&mut self, prefix_len: usize) -> (u16, Option<RawStrError>) {
+    fn raw_double_quoted_string(&mut self, prefix_len: usize) -> (u8, Option<RawStrError>) {
         // Wrap the actual function to handle the error with too many hashes.
         // This way, it eats the whole raw string.
         let (n_hashes, err) = self.raw_string_unvalidated(prefix_len);
-        // Only up to 65535 `#`s are allowed in raw strings
-        match u16::try_from(n_hashes) {
+        // Only up to 255 `#`s are allowed in raw strings
+        match u8::try_from(n_hashes) {
             Ok(num) => (num, err),
             // We lie about the number of hashes here :P
             Err(_) => (0, Some(RawStrError::TooManyDelimiters { found: n_hashes })),

compiler/rustc_lexer/src/tests.rs

@@ -2,7 +2,7 @@ use super::*;
 
 use expect_test::{expect, Expect};
 
-fn check_raw_str(s: &str, expected_hashes: u16, expected_err: Option<RawStrError>) {
+fn check_raw_str(s: &str, expected_hashes: u8, expected_err: Option<RawStrError>) {
     let s = &format!("r{}", s);
     let mut cursor = Cursor::new(s);
     cursor.bump();
@@ -68,13 +68,13 @@ fn test_unterminated_no_pound() {
 
 #[test]
 fn test_too_many_hashes() {
-    let max_count = u16::MAX;
+    let max_count = u8::MAX;
     let mut hashes: String = "#".repeat(max_count.into());
 
-    // Valid number of hashes (65535 = 2^16 - 1), but invalid string.
+    // Valid number of hashes (255 = 2^8 - 1 = u8::MAX), but invalid string.
     check_raw_str(&hashes, max_count, Some(RawStrError::InvalidStarter { bad_char: '\u{0}' }));
 
-    // One more hash sign (65536 = 2^16) becomes too many.
+    // One more hash sign (256 = 2^8) becomes too many.
     hashes.push('#');
     check_raw_str(
         &hashes,

compiler/rustc_parse/src/lexer/mod.rs

@@ -597,15 +597,13 @@ impl<'a> StringReader<'a> {
         }
     }
 
-    /// Note: It was decided to not add a test case, because it would be too big.
-    /// <https://github.com/rust-lang/rust/pull/50296#issuecomment-392135180>
     fn report_too_many_hashes(&self, start: BytePos, found: usize) -> ! {
         self.fatal_span_(
             start,
             self.pos,
             &format!(
                 "too many `#` symbols: raw strings may be delimited \
-                by up to 65535 `#` symbols, but found {}",
+                by up to 255 `#` symbols, but found {}",
                 found
             ),
         )

src/tools/clippy/clippy_lints/src/regex.rs

@@ -81,7 +81,7 @@ impl<'tcx> LateLintPass<'tcx> for Regex {
 
 #[allow(clippy::cast_possible_truncation)] // truncation very unlikely here
 #[must_use]
-fn str_span(base: Span, c: regex_syntax::ast::Span, offset: u16) -> Span {
+fn str_span(base: Span, c: regex_syntax::ast::Span, offset: u8) -> Span {
     let offset = u32::from(offset);
     let end = base.lo() + BytePos(u32::try_from(c.end.offset).expect("offset too large") + offset);
     let start = base.lo() + BytePos(u32::try_from(c.start.offset).expect("offset too large") + offset);