Make non-ASCII errors more consistent.

There are three kinds of "byte" literals: byte literals, byte string
literals, and raw byte string literals. None are allowed to have
non-ASCII chars in them.

Two `EscapeError` variants exist for when that constraint is violated.
- `NonAsciiCharInByte`: used for byte literals and byte string literals.
- `NonAsciiCharInByteString`: used for raw byte string literals.

As a result, the messages for raw byte string literals use different
wording, without good reason. Also, byte string literals are incorrectly
described as "byte constants" in some error messages.

This commit eliminates `NonAsciiCharInByteString` so the three cases are
handled similarly, and described correctly. The `mode` is enough to
distinguish them.

Note: Some existing error messages mention "byte constants" and some
mention "byte literals". I went with the latter here, because it's a
more correct name, as used by the Reference.

Author: nnethercote

compiler/rustc_lexer/src/unescape.rs

@@ -52,10 +52,8 @@ pub enum EscapeError {
 
     /// Unicode escape code in byte literal.
     UnicodeEscapeInByte,
-    /// Non-ascii character in byte literal.
+    /// Non-ascii character in byte literal, byte string literal, or raw byte string literal.
     NonAsciiCharInByte,
-    /// Non-ascii character in byte string literal.
-    NonAsciiCharInByteString,
 
     /// After a line ending with '\', the next line contains whitespace
     /// characters that are not skipped.
@@ -349,8 +347,7 @@ where
         let start = src.len() - chars.as_str().len() - c.len_utf8();
         let result = match c {
             '\r' => Err(EscapeError::BareCarriageReturnInRawString),
-            c if is_byte && !c.is_ascii() => Err(EscapeError::NonAsciiCharInByteString),
-            c => Ok(c),
+            _ => ascii_check(c, is_byte),
         };
         let end = src.len() - chars.as_str().len();
         callback(start..end, result);

compiler/rustc_lexer/src/unescape/tests.rs

@@ -289,9 +289,6 @@ fn test_unescape_raw_byte_str() {
     }
 
     check("\r", &[(0..1, Err(EscapeError::BareCarriageReturnInRawString))]);
-    check("🦀", &[(0..4, Err(EscapeError::NonAsciiCharInByteString))]);
-    check(
-        "🦀a",
-        &[(0..4, Err(EscapeError::NonAsciiCharInByteString)), (4..5, Ok(byte_from_char('a')))],
-    );
+    check("🦀", &[(0..4, Err(EscapeError::NonAsciiCharInByte))]);
+    check("🦀a", &[(0..4, Err(EscapeError::NonAsciiCharInByte)), (4..5, Ok(byte_from_char('a')))]);
 }

compiler/rustc_parse/src/lexer/unescape_error_reporting.rs

@@ -231,16 +231,23 @@ pub(crate) fn emit_unescape_error(
                 .emit();
         }
         EscapeError::NonAsciiCharInByte => {
-            assert!(mode.is_byte());
             let (c, span) = last_char();
-            let mut err = handler.struct_span_err(span, "non-ASCII character in byte constant");
+            let desc = match mode {
+                Mode::Byte => "byte literal",
+                Mode::ByteStr => "byte string literal",
+                Mode::RawByteStr => "raw byte string literal",
+                _ => panic!("non-is_byte literal paired with NonAsciiCharInByte"),
+            };
+            let mut err = handler.struct_span_err(span, format!("non-ASCII character in {}", desc));
             let postfix = if unicode_width::UnicodeWidthChar::width(c).unwrap_or(1) == 0 {
                 format!(" but is {:?}", c)
             } else {
                 String::new()
             };
-            err.span_label(span, &format!("byte constant must be ASCII{}", postfix));
-            if (c as u32) <= 0xFF {
+            err.span_label(span, &format!("must be ASCII{}", postfix));
+            // Note: the \\xHH suggestions are not given for raw byte string
+            // literals, because they are araw and so cannot use any escapes.
+            if (c as u32) <= 0xFF && mode != Mode::RawByteStr {
                 err.span_suggestion(
                     span,
                     &format!(
@@ -250,9 +257,9 @@ pub(crate) fn emit_unescape_error(
                     format!("\\x{:X}", c as u32),
                     Applicability::MaybeIncorrect,
                 );
-            } else if matches!(mode, Mode::Byte) {
+            } else if mode == Mode::Byte {
                 err.span_label(span, "this multibyte character does not fit into a single byte");
-            } else if matches!(mode, Mode::ByteStr) {
+            } else if mode != Mode::RawByteStr {
                 let mut utf8 = String::new();
                 utf8.push(c);
                 err.span_suggestion(
@@ -270,19 +277,6 @@ pub(crate) fn emit_unescape_error(
             }
             err.emit();
         }
-        EscapeError::NonAsciiCharInByteString => {
-            assert!(mode.is_byte());
-            let (c, span) = last_char();
-            let postfix = if unicode_width::UnicodeWidthChar::width(c).unwrap_or(1) == 0 {
-                format!(" but is {:?}", c)
-            } else {
-                String::new()
-            };
-            handler
-                .struct_span_err(span, "raw byte string must be ASCII")
-                .span_label(span, &format!("must be ASCII{}", postfix))
-                .emit();
-        }
         EscapeError::OutOfRangeHexEscape => {
             handler
                 .struct_span_err(span, "out of range hex escape")

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

@@ -1,4 +1,4 @@
 #![feature(rustc_attrs)]
 
-#[rustc_dummy = b"ffi.rs"] //~ ERROR non-ASCII character in byte constant
+#[rustc_dummy = b"ffi.rs"] //~ ERROR non-ASCII character in byte string literal
 fn main() {}

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

@@ -1,8 +1,8 @@
-error: non-ASCII character in byte constant
+error: non-ASCII character in byte string literal
   --> $DIR/key-value-non-ascii.rs:3:19
    |
 LL | #[rustc_dummy = b"ffi.rs"]
-   |                   ^ byte constant must be ASCII
+   |                   ^ must be ASCII
    |
 help: if you meant to use the UTF-8 encoding of 'ffi', use \xHH escapes
    |

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

@@ -7,6 +7,6 @@ pub fn main() {
     b'\x0Z';  //~ ERROR invalid character in numeric character escape: `Z`
     b'	';  //~ ERROR byte constant must be escaped
     b''';  //~ ERROR byte constant must be escaped
-    b'é';  //~ ERROR non-ASCII character in byte constant
+    b'é';  //~ ERROR non-ASCII character in byte literal
     b'a  //~ ERROR unterminated byte constant [E0763]
 }

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

@@ -32,11 +32,11 @@ error: byte constant must be escaped: `'`
 LL |     b''';
    |       ^ help: escape the character: `\'`
 
-error: non-ASCII character in byte constant
+error: non-ASCII character in byte literal
   --> $DIR/byte-literals.rs:10:7
    |
 LL |     b'é';
-   |       ^ byte constant must be ASCII
+   |       ^ must be ASCII
    |
 help: if you meant to use the unicode code point for 'é', use a \xHH escape
    |

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

@@ -3,7 +3,7 @@ static FOO: &'static [u8] = b"\f";  //~ ERROR unknown byte escape
 pub fn main() {
     b"\f";  //~ ERROR unknown byte escape
     b"\x0Z";  //~ ERROR invalid character in numeric character escape: `Z`
-    b"é";  //~ ERROR non-ASCII character in byte constant
-    br##"é"##;  //~ ERROR raw byte string must be ASCII
+    b"é";  //~ ERROR non-ASCII character in byte string literal
+    br##"é"##;  //~ ERROR non-ASCII character in raw byte string literal
     b"a  //~ ERROR unterminated double quote byte string
 }

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

@@ -20,18 +20,18 @@ error: invalid character in numeric character escape: `Z`
 LL |     b"\x0Z";
    |          ^ invalid character in numeric character escape
 
-error: non-ASCII character in byte constant
+error: non-ASCII character in byte string literal
   --> $DIR/byte-string-literals.rs:6:7
    |
 LL |     b"é";
-   |       ^ byte constant must be ASCII
+   |       ^ 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
+error: non-ASCII character in raw byte string literal
   --> $DIR/byte-string-literals.rs:7:10
    |
 LL |     br##"é"##;

src/test/ui/parser/raw/raw-byte-string-literals.rs

@@ -2,6 +2,6 @@
 
 pub fn main() {
     br"a"; //~ ERROR bare CR not allowed in raw string
-    br"é";  //~ ERROR raw byte string must be ASCII
+    br"é";  //~ ERROR non-ASCII character in raw byte string literal
     br##~"a"~##;  //~ ERROR only `#` is allowed in raw string delimitation
 }