Merge #6485

6485: Remove RAW literals r=matklad a=matklad

bors r+
🤖

closes https://github.com/rust-analyzer/rust-analyzer/issues/6308

Co-authored-by: Aleksey Kladov <[email protected]>

7 files changed, 75 insertions, 91 deletions

diff --git a/crates/syntax/src/ast/expr_ext.rs b/crates/syntax/src/ast/expr_ext.rs
index 3d33cd1cf..eb44bb2ab 100644
--- a/crates/syntax/src/ast/expr_ext.rs
+++ b/crates/syntax/src/ast/expr_ext.rs
@@ -320,6 +320,13 @@ impl ast::Literal {
         ast::IntNumber::cast(self.token())
     }
 
+    pub fn as_string(&self) -> Option<ast::String> {
+        ast::String::cast(self.token())
+    }
+    pub fn as_byte_string(&self) -> Option<ast::ByteString> {
+        ast::ByteString::cast(self.token())
+    }
+
     fn find_suffix(text: &str, possible_suffixes: &[&str]) -> Option<SmolStr> {
         possible_suffixes
             .iter()
@@ -351,10 +358,10 @@ impl ast::Literal {
                     suffix: Self::find_suffix(&text, &ast::FloatNumber::SUFFIXES),
                 }
             }
-            STRING | RAW_STRING => LiteralKind::String,
+            STRING => LiteralKind::String,
             T![true] => LiteralKind::Bool(true),
             T![false] => LiteralKind::Bool(false),
-            BYTE_STRING | RAW_BYTE_STRING => LiteralKind::ByteString,
+            BYTE_STRING => LiteralKind::ByteString,
             CHAR => LiteralKind::Char,
             BYTE => LiteralKind::Byte,
             _ => unreachable!(),
diff --git a/crates/syntax/src/ast/generated/tokens.rs b/crates/syntax/src/ast/generated/tokens.rs
index 1b8449221..728b72cd7 100644
--- a/crates/syntax/src/ast/generated/tokens.rs
+++ b/crates/syntax/src/ast/generated/tokens.rs
@@ -70,16 +70,16 @@ impl AstToken for String {
 }
 
 #[derive(Debug, Clone, PartialEq, Eq, Hash)]
-pub struct RawString {
+pub struct ByteString {
     pub(crate) syntax: SyntaxToken,
 }
-impl std::fmt::Display for RawString {
+impl std::fmt::Display for ByteString {
     fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
         std::fmt::Display::fmt(&self.syntax, f)
     }
 }
-impl AstToken for RawString {
-    fn can_cast(kind: SyntaxKind) -> bool { kind == RAW_STRING }
+impl AstToken for ByteString {
+    fn can_cast(kind: SyntaxKind) -> bool { kind == BYTE_STRING }
     fn cast(syntax: SyntaxToken) -> Option<Self> {
         if Self::can_cast(syntax.kind()) {
             Some(Self { syntax })
diff --git a/crates/syntax/src/ast/node_ext.rs b/crates/syntax/src/ast/node_ext.rs
index c5cd1c504..ce35ac01a 100644
--- a/crates/syntax/src/ast/node_ext.rs
+++ b/crates/syntax/src/ast/node_ext.rs
@@ -7,7 +7,7 @@ use itertools::Itertools;
 use parser::SyntaxKind;
 
 use crate::{
-    ast::{self, support, token_ext::HasStringValue, AstNode, AstToken, NameOwner, SyntaxNode},
+    ast::{self, support, AstNode, AstToken, NameOwner, SyntaxNode},
     SmolStr, SyntaxElement, SyntaxToken, T,
 };
 
@@ -55,13 +55,7 @@ impl ast::Attr {
         let key = self.simple_name()?;
         let value_token = lit.syntax().first_token()?;
 
-        let value: SmolStr = if let Some(s) = ast::String::cast(value_token.clone()) {
-            s.value()?.into()
-        } else if let Some(s) = ast::RawString::cast(value_token) {
-            s.value()?.into()
-        } else {
-            return None;
-        };
+        let value: SmolStr = ast::String::cast(value_token.clone())?.value()?.into();
 
         Some((key, value))
     }
diff --git a/crates/syntax/src/ast/token_ext.rs b/crates/syntax/src/ast/token_ext.rs
index 8d3fad5a6..bf0035986 100644
--- a/crates/syntax/src/ast/token_ext.rs
+++ b/crates/syntax/src/ast/token_ext.rs
@@ -114,39 +114,24 @@ impl QuoteOffsets {
     }
 }
 
-pub trait HasQuotes: AstToken {
-    fn quote_offsets(&self) -> Option<QuoteOffsets> {
-        let text = self.text().as_str();
-        let offsets = QuoteOffsets::new(text)?;
-        let o = self.syntax().text_range().start();
-        let offsets = QuoteOffsets {
-            quotes: (offsets.quotes.0 + o, offsets.quotes.1 + o),
-            contents: offsets.contents + o,
-        };
-        Some(offsets)
-    }
-    fn open_quote_text_range(&self) -> Option<TextRange> {
-        self.quote_offsets().map(|it| it.quotes.0)
+impl ast::String {
+    pub fn is_raw(&self) -> bool {
+        self.text().starts_with('r')
     }
-
-    fn close_quote_text_range(&self) -> Option<TextRange> {
-        self.quote_offsets().map(|it| it.quotes.1)
-    }
-
-    fn text_range_between_quotes(&self) -> Option<TextRange> {
-        self.quote_offsets().map(|it| it.contents)
+    pub fn map_range_up(&self, range: TextRange) -> Option<TextRange> {
+        let contents_range = self.text_range_between_quotes()?;
+        assert!(TextRange::up_to(contents_range.len()).contains_range(range));
+        Some(range + contents_range.start())
     }
-}
-
-impl HasQuotes for ast::String {}
-impl HasQuotes for ast::RawString {}
 
-pub trait HasStringValue: HasQuotes {
-    fn value(&self) -> Option<Cow<'_, str>>;
-}
+    pub fn value(&self) -> Option<Cow<'_, str>> {
+        if self.is_raw() {
+            let text = self.text().as_str();
+            let text =
+                &text[self.text_range_between_quotes()? - self.syntax().text_range().start()];
+            return Some(Cow::Borrowed(text));
+        }
 
-impl HasStringValue for ast::String {
-    fn value(&self) -> Option<Cow<'_, str>> {
         let text = self.text().as_str();
         let text = &text[self.text_range_between_quotes()? - self.syntax().text_range().start()];
 
@@ -164,22 +149,31 @@ impl HasStringValue for ast::String {
         let res = if buf == text { Cow::Borrowed(text) } else { Cow::Owned(buf) };
         Some(res)
     }
-}
 
-// FIXME: merge `ast::RawString` and `ast::String`.
-impl HasStringValue for ast::RawString {
-    fn value(&self) -> Option<Cow<'_, str>> {
+    pub fn quote_offsets(&self) -> Option<QuoteOffsets> {
         let text = self.text().as_str();
-        let text = &text[self.text_range_between_quotes()? - self.syntax().text_range().start()];
-        Some(Cow::Borrowed(text))
+        let offsets = QuoteOffsets::new(text)?;
+        let o = self.syntax().text_range().start();
+        let offsets = QuoteOffsets {
+            quotes: (offsets.quotes.0 + o, offsets.quotes.1 + o),
+            contents: offsets.contents + o,
+        };
+        Some(offsets)
+    }
+    pub fn text_range_between_quotes(&self) -> Option<TextRange> {
+        self.quote_offsets().map(|it| it.contents)
+    }
+    pub fn open_quote_text_range(&self) -> Option<TextRange> {
+        self.quote_offsets().map(|it| it.quotes.0)
+    }
+    pub fn close_quote_text_range(&self) -> Option<TextRange> {
+        self.quote_offsets().map(|it| it.quotes.1)
     }
 }
 
-impl ast::RawString {
-    pub fn map_range_up(&self, range: TextRange) -> Option<TextRange> {
-        let contents_range = self.text_range_between_quotes()?;
-        assert!(TextRange::up_to(contents_range.len()).contains_range(range));
-        Some(range + contents_range.start())
+impl ast::ByteString {
+    pub fn is_raw(&self) -> bool {
+        self.text().starts_with("br")
     }
 }
 
@@ -522,22 +516,6 @@ impl HasFormatSpecifier for ast::String {
     }
 }
 
-impl HasFormatSpecifier for ast::RawString {
-    fn char_ranges(
-        &self,
-    ) -> Option<Vec<(TextRange, Result<char, rustc_lexer::unescape::EscapeError>)>> {
-        let text = self.text().as_str();
-        let text = &text[self.text_range_between_quotes()? - self.syntax().text_range().start()];
-        let offset = self.text_range_between_quotes()?.start() - self.syntax().text_range().start();
-
-        let mut res = Vec::with_capacity(text.len());
-        for (idx, c) in text.char_indices() {
-            res.push((TextRange::at(idx.try_into().unwrap(), TextSize::of(c)) + offset, Ok(c)));
-        }
-        Some(res)
-    }
-}
-
 impl ast::IntNumber {
     #[rustfmt::skip]
     pub(crate) const SUFFIXES: &'static [&'static str] = &[
diff --git a/crates/syntax/src/parsing/lexer.rs b/crates/syntax/src/parsing/lexer.rs
index 5674ecb84..8afd7e53b 100644
--- a/crates/syntax/src/parsing/lexer.rs
+++ b/crates/syntax/src/parsing/lexer.rs
@@ -235,7 +235,7 @@ fn rustc_token_kind_to_syntax_kind(
                         RawStrError::TooManyDelimiters { .. } => "Too many `#` symbols: raw strings may be delimited by up to 65535 `#` symbols",
                     };
                 };
-                RAW_STRING
+                STRING
             }
             rustc_lexer::LiteralKind::RawByteStr { err: raw_str_err, .. } => {
                 if let Some(raw_str_err) = raw_str_err {
@@ -250,7 +250,7 @@ fn rustc_token_kind_to_syntax_kind(
                     };
                 };
 
-                RAW_BYTE_STRING
+                BYTE_STRING
             }
         };
 
diff --git a/crates/syntax/src/parsing/reparsing.rs b/crates/syntax/src/parsing/reparsing.rs
index 4149f856a..190f5f67a 100644
--- a/crates/syntax/src/parsing/reparsing.rs
+++ b/crates/syntax/src/parsing/reparsing.rs
@@ -44,7 +44,7 @@ fn reparse_token<'node>(
     let prev_token = algo::find_covering_element(root, edit.delete).as_token()?.clone();
     let prev_token_kind = prev_token.kind();
     match prev_token_kind {
-        WHITESPACE | COMMENT | IDENT | STRING | RAW_STRING => {
+        WHITESPACE | COMMENT | IDENT | STRING => {
             if prev_token_kind == WHITESPACE || prev_token_kind == COMMENT {
                 // removing a new line may extends previous token
                 let deleted_range = edit.delete - prev_token.text_range().start();
diff --git a/crates/syntax/src/validation.rs b/crates/syntax/src/validation.rs
index 0f9a5e8ae..62a37c50a 100644
--- a/crates/syntax/src/validation.rs
+++ b/crates/syntax/src/validation.rs
@@ -4,7 +4,7 @@ mod block;
 
 use crate::{
     algo, ast, match_ast, AstNode, SyntaxError,
-    SyntaxKind::{BYTE, BYTE_STRING, CHAR, CONST, FN, INT_NUMBER, STRING, TYPE_ALIAS},
+    SyntaxKind::{BYTE, CHAR, CONST, FN, INT_NUMBER, TYPE_ALIAS},
     SyntaxNode, SyntaxToken, TextSize, T,
 };
 use rowan::Direction;
@@ -121,18 +121,19 @@ fn validate_literal(literal: ast::Literal, acc: &mut Vec<SyntaxError>) {
         acc.push(SyntaxError::new_at_offset(rustc_unescape_error_to_string(err), off));
     };
 
-    match token.kind() {
-        BYTE => {
-            if let Some(Err(e)) = unquote(text, 2, '\'').map(unescape_byte) {
-                push_err(2, e);
-            }
-        }
-        CHAR => {
-            if let Some(Err(e)) = unquote(text, 1, '\'').map(unescape_char) {
-                push_err(1, e);
+    if let Some(s) = literal.as_string() {
+        if !s.is_raw() {
+            if let Some(without_quotes) = unquote(text, 1, '"') {
+                unescape_literal(without_quotes, Mode::Str, &mut |range, char| {
+                    if let Err(err) = char {
+                        push_err(1, (range.start, err));
+                    }
+                })
             }
         }
-        BYTE_STRING => {
+    }
+    if let Some(s) = literal.as_byte_string() {
+        if !s.is_raw() {
             if let Some(without_quotes) = unquote(text, 2, '"') {
                 unescape_byte_literal(without_quotes, Mode::ByteStr, &mut |range, char| {
                     if let Err(err) = char {
@@ -141,13 +142,17 @@ fn validate_literal(literal: ast::Literal, acc: &mut Vec<SyntaxError>) {
                 })
             }
         }
-        STRING => {
-            if let Some(without_quotes) = unquote(text, 1, '"') {
-                unescape_literal(without_quotes, Mode::Str, &mut |range, char| {
-                    if let Err(err) = char {
-                        push_err(1, (range.start, err));
-                    }
-                })
+    }
+
+    match token.kind() {
+        BYTE => {
+            if let Some(Err(e)) = unquote(text, 2, '\'').map(unescape_byte) {
+                push_err(2, e);
+            }
+        }
+        CHAR => {
+            if let Some(Err(e)) = unquote(text, 1, '\'').map(unescape_char) {
+                push_err(1, e);
             }
         }
         _ => (),