feat: Implement string escape sequences (noir-lang#2803)

Sakapoi · Oct 19, 2023 · 617aec0 · 617aec0
1 parent a100103
commit 617aec0
Show file tree

Hide file tree

Showing 3 changed files with 46 additions and 12 deletions.
diff --git a/compiler/noirc_frontend/src/lexer/errors.rs b/compiler/noirc_frontend/src/lexer/errors.rs
@@ -21,6 +21,12 @@ pub enum LexerErrorKind {
     LogicalAnd { span: Span },
     #[error("Unterminated block comment")]
     UnterminatedBlockComment { span: Span },
+    #[error("Unterminated string literal")]
+    UnterminatedStringLiteral { span: Span },
+    #[error(
+        "'\\{escaped}' is not a valid escape sequence. Use '\\' for a literal backslash character."
+    )]
+    InvalidEscape { escaped: char, span: Span },
 }
 
 impl LexerErrorKind {
@@ -33,6 +39,8 @@ impl LexerErrorKind {
             LexerErrorKind::TooManyBits { span, .. } => *span,
             LexerErrorKind::LogicalAnd { span } => *span,
             LexerErrorKind::UnterminatedBlockComment { span } => *span,
+            LexerErrorKind::UnterminatedStringLiteral { span } => *span,
+            LexerErrorKind::InvalidEscape { span, .. } => *span,
         }
     }
 
@@ -46,30 +54,30 @@ impl LexerErrorKind {
                 let found: String = found.map(Into::into).unwrap_or_else(|| "<eof>".into());
 
                 (
-                    "an unexpected character was found".to_string(),
-                    format!(" expected {expected} , but got {found}"),
+                    "An unexpected character was found".to_string(),
+                    format!("Expected {expected}, but found {found}"),
                     *span,
                 )
             },
             LexerErrorKind::NotADoubleChar { span, found } => (
-                format!("tried to parse {found} as double char"),
+                format!("Tried to parse {found} as double char"),
                 format!(
                     " {found:?} is not a double char, this is an internal error"
                 ),
                 *span,
             ),
             LexerErrorKind::InvalidIntegerLiteral { span, found } => (
-                "invalid integer literal".to_string(),
+                "Invalid integer literal".to_string(),
                 format!(" {found} is not an integer"),
                 *span,
             ),
             LexerErrorKind::MalformedFuncAttribute { span, found } => (
-                "malformed function attribute".to_string(),
+                "Malformed function attribute".to_string(),
                 format!(" {found} is not a valid attribute"),
                 *span,
             ),
             LexerErrorKind::TooManyBits { span, max, got } => (
-                "integer literal too large".to_string(),
+                "Integer literal too large".to_string(),
                 format!(
                     "The maximum number of bits needed to represent a field is {max}, This integer type needs {got} bits"
                 ),
@@ -80,7 +88,11 @@ impl LexerErrorKind {
                 "Try `&` instead, or use `if` only if you require short-circuiting".to_string(),
                 *span,
             ),
-            LexerErrorKind::UnterminatedBlockComment { span } => ("unterminated block comment".to_string(), "Unterminated block comment".to_string(), *span),
+            LexerErrorKind::UnterminatedBlockComment { span } => ("Unterminated block comment".to_string(), "Unterminated block comment".to_string(), *span),
+            LexerErrorKind::UnterminatedStringLiteral { span } =>
+                ("Unterminated string literal".to_string(), "Unterminated string literal".to_string(), *span),
+            LexerErrorKind::InvalidEscape { escaped, span } =>
+                (format!("'\\{escaped}' is not a valid escape sequence. Use '\\' for a literal backslash character."), "Invalid escape sequence".to_string(), *span),
         }
     }
 }

diff --git a/compiler/noirc_frontend/src/lexer/lexer.rs b/compiler/noirc_frontend/src/lexer/lexer.rs
@@ -332,12 +332,34 @@ impl<'a> Lexer<'a> {
 
     fn eat_string_literal(&mut self) -> SpannedTokenResult {
         let start = self.position;
+        let mut string = String::new();
+
+        while let Some(next) = self.next_char() {
+            let char = match next {
+                '"' => break,
+                '\\' => match self.next_char() {
+                    Some('r') => '\r',
+                    Some('n') => '\n',
+                    Some('t') => '\t',
+                    Some('0') => '\0',
+                    Some('"') => '"',
+                    Some('\\') => '\\',
+                    Some(escaped) => {
+                        let span = Span::inclusive(start, self.position);
+                        return Err(LexerErrorKind::InvalidEscape { escaped, span });
+                    }
+                    None => {
+                        let span = Span::inclusive(start, self.position);
+                        return Err(LexerErrorKind::UnterminatedStringLiteral { span });
+                    }
+                },
+                other => other,
+            };
 
-        let str_literal = self.eat_while(None, |ch| ch != '"');
-
-        let str_literal_token = Token::Str(str_literal);
+            string.push(char);
+        }
 
-        self.next_char(); // Advance past the closing quote
+        let str_literal_token = Token::Str(string);
 
         let end = self.position;
         Ok(str_literal_token.into_span(start, end))

diff --git a/tooling/nargo_cli/tests/execution_success/strings/src/main.nr b/tooling/nargo_cli/tests/execution_success/strings/src/main.nr
@@ -19,7 +19,7 @@ fn main(message : pub str<11>, y : Field, hex_as_string : str<4>, hex_as_field :
     assert(y == 5); // Change to y != 5 to see how the later print statements are not called
     std::println(array);
 
-    bad_message = "helld world";
+    bad_message = "hell\0\"world";
     std::println(bad_message);
     assert(message != bad_message);