Fix invalid suggestions for non-ASCII characters in byte constants

Author: FabianWolff

compiler/rustc_parse/src/lexer/unescape_error_reporting.rs

@@ -153,16 +153,37 @@ pub(crate) fn emit_unescape_error(
         EscapeError::NonAsciiCharInByte => {
             assert!(mode.is_bytes());
             let (c, span) = last_char();
-            handler
-                .struct_span_err(span, "non-ASCII character in byte constant")
-                .span_label(span, "byte constant must be ASCII")
-                .span_suggestion(
+            let mut err = handler.struct_span_err(span, "non-ASCII character in byte constant");
+            err.span_label(span, "byte constant must be ASCII");
+            if (c as u32) <= 0xFF {
+                err.span_suggestion(
                     span,
-                    "use a \\xHH escape for a non-ASCII byte",
+                    &format!(
+                        "if you meant to use the unicode code point for '{}', use a \\xHH escape",
+                        c
+                    ),
                     format!("\\x{:X}", c as u32),
-                    Applicability::MachineApplicable,
-                )
-                .emit();
+                    Applicability::MaybeIncorrect,
+                );
+            } else if matches!(mode, Mode::Byte) {
+                err.span_label(span, "this multibyte character does not fit into a single byte");
+            } else if matches!(mode, Mode::ByteStr) {
+                let mut utf8 = String::new();
+                utf8.push(c);
+                err.span_suggestion(
+                    span,
+                    &format!(
+                        "if you meant to use the UTF-8 encoding of '{}', use \\xHH escapes",
+                        c
+                    ),
+                    utf8.as_bytes()
+                        .iter()
+                        .map(|b: &u8| format!("\\x{:X}", *b))
+                        .fold("".to_string(), |a, c| a + &c),
+                    Applicability::MaybeIncorrect,
+                );
+            }
+            err.emit();
         }
         EscapeError::NonAsciiCharInByteString => {
             assert!(mode.is_bytes());

src/test/ui/attributes/key-value-non-ascii.stderr

@@ -2,10 +2,12 @@ error: non-ASCII character in byte constant
   --> $DIR/key-value-non-ascii.rs:3:19
    |
 LL | #[rustc_dummy = b"ffi.rs"]
-   |                   ^
-   |                   |
-   |                   byte constant must be ASCII
-   |                   help: use a \xHH escape for a non-ASCII byte: `\xFB03`
+   |                   ^ byte constant must be ASCII
+   |
+help: if you meant to use the UTF-8 encoding of 'ffi', use \xHH escapes
+   |
+LL | #[rustc_dummy = b"/xEF/xAC/x83.rs"]
+   |                   ^^^^^^^^^^^^
 
 error: aborting due to previous error
 

src/test/ui/parser/byte-literals.stderr

@@ -36,10 +36,12 @@ error: non-ASCII character in byte constant
   --> $DIR/byte-literals.rs:10:7
    |
 LL |     b'é';
-   |       ^
-   |       |
-   |       byte constant must be ASCII
-   |       help: use a \xHH escape for a non-ASCII byte: `\xE9`
+   |       ^ byte constant must be ASCII
+   |
+help: if you meant to use the unicode code point for 'é', use a \xHH escape
+   |
+LL |     b'\xE9';
+   |       ^^^^
 
 error[E0763]: unterminated byte constant
   --> $DIR/byte-literals.rs:11:6

src/test/ui/parser/byte-string-literals.stderr

@@ -24,10 +24,12 @@ error: non-ASCII character in byte constant
   --> $DIR/byte-string-literals.rs:6:7
    |
 LL |     b"é";
-   |       ^
-   |       |
-   |       byte constant must be ASCII
-   |       help: use a \xHH escape for a non-ASCII byte: `\xE9`
+   |       ^ byte constant must be ASCII
+   |
+help: if you meant to use the unicode code point for 'é', use a \xHH escape
+   |
+LL |     b"\xE9";
+   |       ^^^^
 
 error: raw byte string must be ASCII
   --> $DIR/byte-string-literals.rs:7:10

src/test/ui/suggestions/multibyte-escapes.rs

@@ -0,0 +1,18 @@
+// Regression test for #87397.
+
+fn main() {
+    b'µ';
+    //~^ ERROR: non-ASCII character in byte constant
+    //~| HELP: if you meant to use the unicode code point for 'µ', use a \xHH escape
+    //~| NOTE: byte constant must be ASCII
+
+    b'字';
+    //~^ ERROR: non-ASCII character in byte constant
+    //~| NOTE: this multibyte character does not fit into a single byte
+    //~| NOTE: byte constant must be ASCII
+
+    b"字";
+    //~^ ERROR: non-ASCII character in byte constant
+    //~| HELP: if you meant to use the UTF-8 encoding of '字', use \xHH escapes
+    //~| NOTE: byte constant must be ASCII
+}

src/test/ui/suggestions/multibyte-escapes.stderr

@@ -0,0 +1,33 @@
+error: non-ASCII character in byte constant
+  --> $DIR/multibyte-escapes.rs:4:7
+   |
+LL |     b'µ';
+   |       ^ byte constant must be ASCII
+   |
+help: if you meant to use the unicode code point for 'µ', use a \xHH escape
+   |
+LL |     b'\xB5';
+   |       ^^^^
+
+error: non-ASCII character in byte constant
+  --> $DIR/multibyte-escapes.rs:9:7
+   |
+LL |     b'字';
+   |       ^^
+   |       |
+   |       byte constant must be ASCII
+   |       this multibyte character does not fit into a single byte
+
+error: non-ASCII character in byte constant
+  --> $DIR/multibyte-escapes.rs:14:7
+   |
+LL |     b"字";
+   |       ^^ byte constant must be ASCII
+   |
+help: if you meant to use the UTF-8 encoding of '字', use \xHH escapes
+   |
+LL |     b"\xE5\xAD\x97";
+   |       ^^^^^^^^^^^^
+
+error: aborting due to 3 previous errors
+