Rollup merge of #72047 - Julian-Wollersberger:literal_error_reporting…

…_cleanup, r=petrochenkov

Literal error reporting cleanup

While doing some performance work, I noticed some code duplication in `librustc_parser/lexer/mod.rs`, so I cleaned it up.

This PR is probably best reviewed commit by commit.

I'm not sure what the API stability practices for `librustc_lexer` are. Four public methods in `unescape.rs` can be removed, but two are used by clippy, so I left them in for now.
I could open a PR for Rust-Analyzer when this one lands.

But how do I open a PR for clippy? (Git submodules are frustrating to work with)

src/librustc_ast/util/literal.rs

@@ -6,8 +6,7 @@ use crate::tokenstream::TokenTree;
 
 use rustc_data_structures::sync::Lrc;
 use rustc_lexer::unescape::{unescape_byte, unescape_char};
-use rustc_lexer::unescape::{unescape_byte_str, unescape_str};
-use rustc_lexer::unescape::{unescape_raw_byte_str, unescape_raw_str};
+use rustc_lexer::unescape::{unescape_byte_literal, unescape_literal, Mode};
 use rustc_span::symbol::{kw, sym, Symbol};
 use rustc_span::Span;
 
@@ -59,45 +58,53 @@ impl LitKind {
                 // new symbol because the string in the LitKind is different to the
                 // string in the token.
                 let s = symbol.as_str();
-                let symbol = if s.contains(&['\\', '\r'][..]) {
-                    let mut buf = String::with_capacity(s.len());
-                    let mut error = Ok(());
-                    unescape_str(&s, &mut |_, unescaped_char| match unescaped_char {
-                        Ok(c) => buf.push(c),
-                        Err(_) => error = Err(LitError::LexerError),
-                    });
-                    error?;
-                    Symbol::intern(&buf)
-                } else {
-                    symbol
-                };
+                let symbol =
+                    if s.contains(&['\\', '\r'][..]) {
+                        let mut buf = String::with_capacity(s.len());
+                        let mut error = Ok(());
+                        unescape_literal(&s, Mode::Str, &mut |_, unescaped_char| {
+                            match unescaped_char {
+                                Ok(c) => buf.push(c),
+                                Err(_) => error = Err(LitError::LexerError),
+                            }
+                        });
+                        error?;
+                        Symbol::intern(&buf)
+                    } else {
+                        symbol
+                    };
                 LitKind::Str(symbol, ast::StrStyle::Cooked)
             }
             token::StrRaw(n) => {
                 // Ditto.
                 let s = symbol.as_str();
-                let symbol = if s.contains('\r') {
-                    let mut buf = String::with_capacity(s.len());
-                    let mut error = Ok(());
-                    unescape_raw_str(&s, &mut |_, unescaped_char| match unescaped_char {
-                        Ok(c) => buf.push(c),
-                        Err(_) => error = Err(LitError::LexerError),
-                    });
-                    error?;
-                    buf.shrink_to_fit();
-                    Symbol::intern(&buf)
-                } else {
-                    symbol
-                };
+                let symbol =
+                    if s.contains('\r') {
+                        let mut buf = String::with_capacity(s.len());
+                        let mut error = Ok(());
+                        unescape_literal(&s, Mode::RawStr, &mut |_, unescaped_char| {
+                            match unescaped_char {
+                                Ok(c) => buf.push(c),
+                                Err(_) => error = Err(LitError::LexerError),
+                            }
+                        });
+                        error?;
+                        buf.shrink_to_fit();
+                        Symbol::intern(&buf)
+                    } else {
+                        symbol
+                    };
                 LitKind::Str(symbol, ast::StrStyle::Raw(n))
             }
             token::ByteStr => {
                 let s = symbol.as_str();
                 let mut buf = Vec::with_capacity(s.len());
                 let mut error = Ok(());
-                unescape_byte_str(&s, &mut |_, unescaped_byte| match unescaped_byte {
-                    Ok(c) => buf.push(c),
-                    Err(_) => error = Err(LitError::LexerError),
+                unescape_byte_literal(&s, Mode::ByteStr, &mut |_, unescaped_byte| {
+                    match unescaped_byte {
+                        Ok(c) => buf.push(c),
+                        Err(_) => error = Err(LitError::LexerError),
+                    }
                 });
                 error?;
                 buf.shrink_to_fit();
@@ -108,9 +115,11 @@ impl LitKind {
                 let bytes = if s.contains('\r') {
                     let mut buf = Vec::with_capacity(s.len());
                     let mut error = Ok(());
-                    unescape_raw_byte_str(&s, &mut |_, unescaped_byte| match unescaped_byte {
-                        Ok(c) => buf.push(c),
-                        Err(_) => error = Err(LitError::LexerError),
+                    unescape_byte_literal(&s, Mode::RawByteStr, &mut |_, unescaped_byte| {
+                        match unescaped_byte {
+                            Ok(c) => buf.push(c),
+                            Err(_) => error = Err(LitError::LexerError),
+                        }
                     });
                     error?;
                     buf.shrink_to_fit();

src/librustc_lexer/src/lib.rs

@@ -1,5 +1,11 @@
 //! Low-level Rust lexer.
 //!
+//! The idea with `librustc_lexer` is to make a reusable library,
+//! by separating out pure lexing and rustc-specific concerns, like spans,
+//! error reporting an interning.  So, rustc_lexer operates directly on `&str`,
+//! produces simple tokens which are a pair of type-tag and a bit of original text,
+//! and does not report errors, instead storing them as flags on the token.
+//!
 //! Tokens produced by this lexer are not yet ready for parsing the Rust syntax,
 //! for that see `librustc_parse::lexer`, which converts this basic token stream
 //! into wide tokens used by actual parser.
@@ -719,6 +725,9 @@ impl Cursor<'_> {
 
             // Check that amount of closing '#' symbols
             // is equal to the amount of opening ones.
+            // Note that this will not consume extra trailing `#` characters:
+            // `r###"abcde"####` is lexed as a `LexedRawString { n_hashes: 3 }`
+            // followed by a `#` token.
             let mut hashes_left = n_start_hashes;
             let is_closing_hash = |c| {
                 if c == '#' && hashes_left != 0 {
@@ -739,8 +748,8 @@ impl Cursor<'_> {
                     possible_terminator_offset: None,
                 };
             } else if n_end_hashes > max_hashes {
-                // Keep track of possible terminators to give a hint about where there might be
-                // a missing terminator
+                // Keep track of possible terminators to give a hint about
+                // where there might be a missing terminator
                 possible_terminator_offset =
                     Some(self.len_consumed() - start_pos - n_end_hashes + prefix_len);
                 max_hashes = n_end_hashes;

src/librustc_lexer/src/unescape.rs

@@ -58,69 +58,57 @@ pub enum EscapeError {
     NonAsciiCharInByteString,
 }
 
-/// Takes a contents of a char literal (without quotes), and returns an
-/// unescaped char or an error
-pub fn unescape_char(literal_text: &str) -> Result<char, (usize, EscapeError)> {
-    let mut chars = literal_text.chars();
-    unescape_char_or_byte(&mut chars, Mode::Char)
-        .map_err(|err| (literal_text.len() - chars.as_str().len(), err))
-}
-
-/// Takes a contents of a byte literal (without quotes), and returns an
-/// unescaped byte or an error.
-pub fn unescape_byte(literal_text: &str) -> Result<u8, (usize, EscapeError)> {
-    let mut chars = literal_text.chars();
-    unescape_char_or_byte(&mut chars, Mode::Byte)
-        .map(byte_from_char)
-        .map_err(|err| (literal_text.len() - chars.as_str().len(), err))
-}
-
-/// Takes a contents of a string literal (without quotes) and produces a
+/// Takes a contents of a literal (without quotes) and produces a
 /// sequence of escaped characters or errors.
 /// Values are returned through invoking of the provided callback.
-pub fn unescape_str<F>(literal_text: &str, callback: &mut F)
+pub fn unescape_literal<F>(literal_text: &str, mode: Mode, callback: &mut F)
 where
     F: FnMut(Range<usize>, Result<char, EscapeError>),
 {
-    unescape_str_or_byte_str(literal_text, Mode::Str, callback)
+    match mode {
+        Mode::Char | Mode::Byte => {
+            let mut chars = literal_text.chars();
+            let result = unescape_char_or_byte(&mut chars, mode);
+            // The Chars iterator moved forward.
+            callback(0..(literal_text.len() - chars.as_str().len()), result);
+        }
+        Mode::Str | Mode::ByteStr => unescape_str_or_byte_str(literal_text, mode, callback),
+        // NOTE: Raw strings do not perform any explicit character escaping, here we
+        // only translate CRLF to LF and produce errors on bare CR.
+        Mode::RawStr | Mode::RawByteStr => {
+            unescape_raw_str_or_byte_str(literal_text, mode, callback)
+        }
+    }
 }
 
-/// Takes a contents of a byte string literal (without quotes) and produces a
-/// sequence of bytes or errors.
+/// Takes a contents of a byte, byte string or raw byte string (without quotes)
+/// and produces a sequence of bytes or errors.
 /// Values are returned through invoking of the provided callback.
-pub fn unescape_byte_str<F>(literal_text: &str, callback: &mut F)
+pub fn unescape_byte_literal<F>(literal_text: &str, mode: Mode, callback: &mut F)
 where
     F: FnMut(Range<usize>, Result<u8, EscapeError>),
 {
-    unescape_str_or_byte_str(literal_text, Mode::ByteStr, &mut |range, char| {
-        callback(range, char.map(byte_from_char))
+    assert!(mode.is_bytes());
+    unescape_literal(literal_text, mode, &mut |range, result| {
+        callback(range, result.map(byte_from_char));
     })
 }
 
-/// Takes a contents of a raw string literal (without quotes) and produces a
-/// sequence of characters or errors.
-/// Values are returned through invoking of the provided callback.
-/// NOTE: Raw strings do not perform any explicit character escaping, here we
-/// only translate CRLF to LF and produce errors on bare CR.
-pub fn unescape_raw_str<F>(literal_text: &str, callback: &mut F)
-where
-    F: FnMut(Range<usize>, Result<char, EscapeError>),
-{
-    unescape_raw_str_or_byte_str(literal_text, Mode::Str, callback)
+/// Takes a contents of a char literal (without quotes), and returns an
+/// unescaped char or an error
+pub fn unescape_char(literal_text: &str) -> Result<char, (usize, EscapeError)> {
+    let mut chars = literal_text.chars();
+    unescape_char_or_byte(&mut chars, Mode::Char)
+        .map_err(|err| (literal_text.len() - chars.as_str().len(), err))
 }
 
-/// Takes a contents of a raw byte string literal (without quotes) and produces a
-/// sequence of bytes or errors.
-/// Values are returned through invoking of the provided callback.
-/// NOTE: Raw strings do not perform any explicit character escaping, here we
-/// only translate CRLF to LF and produce errors on bare CR.
-pub fn unescape_raw_byte_str<F>(literal_text: &str, callback: &mut F)
-where
-    F: FnMut(Range<usize>, Result<u8, EscapeError>),
-{
-    unescape_raw_str_or_byte_str(literal_text, Mode::ByteStr, &mut |range, char| {
-        callback(range, char.map(byte_from_char))
-    })
+/// Takes a contents of a byte literal (without quotes), and returns an
+/// unescaped byte or an error.
+pub fn unescape_byte(literal_text: &str) -> Result<u8, (usize, EscapeError)> {
+    let mut chars = literal_text.chars();
+    unescape_char_or_byte(&mut chars, Mode::Byte)
+        .map(byte_from_char)
+        .map_err(|err| (literal_text.len() - chars.as_str().len(), err))
 }
 
 /// What kind of literal do we parse.
@@ -130,13 +118,15 @@ pub enum Mode {
     Str,
     Byte,
     ByteStr,
+    RawStr,
+    RawByteStr,
 }
 
 impl Mode {
     pub fn in_single_quotes(self) -> bool {
         match self {
             Mode::Char | Mode::Byte => true,
-            Mode::Str | Mode::ByteStr => false,
+            Mode::Str | Mode::ByteStr | Mode::RawStr | Mode::RawByteStr => false,
         }
     }
 
@@ -146,8 +136,8 @@ impl Mode {
 
     pub fn is_bytes(self) -> bool {
         match self {
-            Mode::Byte | Mode::ByteStr => true,
-            Mode::Char | Mode::Str => false,
+            Mode::Byte | Mode::ByteStr | Mode::RawByteStr => true,
+            Mode::Char | Mode::Str | Mode::RawStr => false,
         }
     }
 }
@@ -345,7 +335,7 @@ where
 
 fn byte_from_char(c: char) -> u8 {
     let res = c as u32;
-    assert!(res <= u8::max_value() as u32, "guaranteed because of Mode::Byte(Str)");
+    assert!(res <= u8::max_value() as u32, "guaranteed because of Mode::ByteStr");
     res as u8
 }
 

src/librustc_lexer/src/unescape/tests.rs

@@ -102,7 +102,7 @@ fn test_unescape_char_good() {
 fn test_unescape_str_good() {
     fn check(literal_text: &str, expected: &str) {
         let mut buf = Ok(String::with_capacity(literal_text.len()));
-        unescape_str(literal_text, &mut |range, c| {
+        unescape_literal(literal_text, Mode::Str, &mut |range, c| {
             if let Ok(b) = &mut buf {
                 match c {
                     Ok(c) => b.push(c),
@@ -222,7 +222,7 @@ fn test_unescape_byte_good() {
 fn test_unescape_byte_str_good() {
     fn check(literal_text: &str, expected: &[u8]) {
         let mut buf = Ok(Vec::with_capacity(literal_text.len()));
-        unescape_byte_str(literal_text, &mut |range, c| {
+        unescape_byte_literal(literal_text, Mode::ByteStr, &mut |range, c| {
             if let Ok(b) = &mut buf {
                 match c {
                     Ok(c) => b.push(c),
@@ -246,7 +246,7 @@ fn test_unescape_byte_str_good() {
 fn test_unescape_raw_str() {
     fn check(literal: &str, expected: &[(Range<usize>, Result<char, EscapeError>)]) {
         let mut unescaped = Vec::with_capacity(literal.len());
-        unescape_raw_str(literal, &mut |range, res| unescaped.push((range, res)));
+        unescape_literal(literal, Mode::RawStr, &mut |range, res| unescaped.push((range, res)));
         assert_eq!(unescaped, expected);
     }
 
@@ -258,7 +258,9 @@ fn test_unescape_raw_str() {
 fn test_unescape_raw_byte_str() {
     fn check(literal: &str, expected: &[(Range<usize>, Result<u8, EscapeError>)]) {
         let mut unescaped = Vec::with_capacity(literal.len());
-        unescape_raw_byte_str(literal, &mut |range, res| unescaped.push((range, res)));
+        unescape_byte_literal(literal, Mode::RawByteStr, &mut |range, res| {
+            unescaped.push((range, res))
+        });
         assert_eq!(unescaped, expected);
     }