refactor(regular_expression): Align diagnostics (#5543)

Manage all diagnostics for LiteralParser, FlagsParser, PatternParser in one place, same message format.

crates/oxc_regular_expression/src/body_parser/mod.rs

@@ -1,4 +1,3 @@
-mod diagnostics;
 mod parser;
 mod reader;
 mod state;

crates/oxc_regular_expression/src/body_parser/parser.rs

@@ -4,7 +4,8 @@ use oxc_span::Atom as SpanAtom;
 
 use crate::{
     ast,
-    body_parser::{diagnostics, reader::Reader, state::State, unicode, unicode_property},
+    body_parser::{reader::Reader, state::State, unicode, unicode_property},
+    diagnostics,
     options::ParserOptions,
     span::SpanFactory,
     surrogate_pair,

crates/oxc_regular_expression/src/diagnostics.rs

@@ -3,6 +3,43 @@ use oxc_span::Span;
 
 const PREFIX: &str = "Invalid regular expression:";
 
+// For (Literal)Parser ---
+
+#[cold]
+pub fn unexpected_literal_char(span0: Span) -> OxcDiagnostic {
+    OxcDiagnostic::error(format!("{PREFIX} Unexpected literal character")).with_label(span0)
+}
+
+#[cold]
+pub fn unterminated_literal(span0: Span, kind: &str) -> OxcDiagnostic {
+    OxcDiagnostic::error(format!("{PREFIX} Unterminated {kind}")).with_label(span0)
+}
+
+#[cold]
+pub fn empty_literal(span0: Span) -> OxcDiagnostic {
+    OxcDiagnostic::error(format!("{PREFIX} Empty literal")).with_label(span0)
+}
+
+// For FlagsParser ---
+
+#[cold]
+pub fn duplicated_flag(span0: Span) -> OxcDiagnostic {
+    OxcDiagnostic::error(format!("{PREFIX} Duplicated flag")).with_label(span0)
+}
+
+#[cold]
+pub fn unknown_flag(span0: Span) -> OxcDiagnostic {
+    OxcDiagnostic::error(format!("{PREFIX} Unknown flag")).with_label(span0)
+}
+
+#[cold]
+pub fn invalid_unicode_flags(span0: Span) -> OxcDiagnostic {
+    OxcDiagnostic::error(format!("{PREFIX} Invalid flags, `u` and `v` should be used alone"))
+        .with_label(span0)
+}
+
+// For PatternParser ---
+
 #[cold]
 pub fn duplicated_capturing_group_names(spans: Vec<Span>) -> OxcDiagnostic {
     OxcDiagnostic::error(format!("{PREFIX} Duplicated capturing group names")).with_labels(spans)

crates/oxc_regular_expression/src/flag_parser.rs

@@ -1,8 +1,8 @@
 use oxc_allocator::Allocator;
-use oxc_diagnostics::{OxcDiagnostic, Result};
+use oxc_diagnostics::Result;
 use rustc_hash::FxHashSet;
 
-use crate::{ast, options::ParserOptions, span::SpanFactory};
+use crate::{ast, diagnostics, options::ParserOptions, span::SpanFactory};
 
 pub struct FlagsParser<'a> {
     source_text: &'a str,
@@ -20,8 +20,7 @@ impl<'a> FlagsParser<'a> {
     }
 
     pub fn parse(&mut self) -> Result<ast::Flags> {
-        let mut existing_flags = FxHashSet::default();
-
+        let span = self.span_factory.create(0, self.source_text.len());
         let mut global = false;
         let mut ignore_case = false;
         let mut multiline = false;
@@ -31,9 +30,10 @@ impl<'a> FlagsParser<'a> {
         let mut has_indices = false;
         let mut unicode_sets = false;
 
-        for c in self.source_text.chars() {
+        let mut existing_flags = FxHashSet::default();
+        for (idx, c) in self.source_text.char_indices() {
             if !existing_flags.insert(c) {
-                return Err(OxcDiagnostic::error(format!("Duplicated flag `{c}`")));
+                return Err(diagnostics::duplicated_flag(self.span_factory.create(idx, idx)));
             }
 
             match c {
@@ -45,17 +45,16 @@ impl<'a> FlagsParser<'a> {
                 's' => dot_all = true,
                 'd' => has_indices = true,
                 'v' => unicode_sets = true,
-                _ => return Err(OxcDiagnostic::error(format!("Invalid flag `{c}`"))),
+                _ => return Err(diagnostics::unknown_flag(self.span_factory.create(idx, idx))),
             }
         }
 
-        // This should be a `SyntaxError`
         if unicode && unicode_sets {
-            return Err(OxcDiagnostic::error("Invalid regular expression flags"));
+            return Err(diagnostics::invalid_unicode_flags(span));
         }
 
         Ok(ast::Flags {
-            span: self.span_factory.create(0, self.source_text.len()),
+            span,
             global,
             ignore_case,
             multiline,

crates/oxc_regular_expression/src/lib.rs

@@ -2,6 +2,7 @@
 
 pub mod ast;
 mod body_parser;
+mod diagnostics;
 mod display;
 mod flag_parser;
 mod literal_parser;

crates/oxc_regular_expression/src/literal_parser.rs

@@ -1,8 +1,8 @@
 use oxc_allocator::Allocator;
-use oxc_diagnostics::{OxcDiagnostic, Result};
+use oxc_diagnostics::Result;
 
 use crate::{
-    ast, body_parser::PatternParser, flag_parser::FlagsParser, options::ParserOptions,
+    ast, body_parser::PatternParser, diagnostics, flag_parser::FlagsParser, options::ParserOptions,
     span::SpanFactory,
 };
 
@@ -28,7 +28,7 @@ impl<'a> Parser<'a> {
         // Precheck if the source text is a valid regular expression literal
         // If valid, parse the pattern and flags with returned span offsets
         let (body_start_offset, body_end_offset, flag_start_offset) =
-            parse_reg_exp_literal(self.source_text)?;
+            parse_reg_exp_literal(self.source_text, &self.span_factory)?;
 
         // Parse flags first to know if unicode mode is enabled or not
         let flags = FlagsParser::new(
@@ -67,12 +67,15 @@ impl<'a> Parser<'a> {
 /// / RegularExpressionBody / RegularExpressionFlags
 /// ```
 /// Returns `(body_start_offset, body_end_offset, flag_start_offset)`.
-fn parse_reg_exp_literal(source_text: &str) -> Result<(usize, usize, usize)> {
+fn parse_reg_exp_literal(
+    source_text: &str,
+    span_factory: &SpanFactory,
+) -> Result<(usize, usize, usize)> {
     let mut offset = 0;
     let mut chars = source_text.chars().peekable();
 
     let Some('/') = chars.next() else {
-        return Err(OxcDiagnostic::error("Invalid regular expression: Unexpected character"));
+        return Err(diagnostics::unexpected_literal_char(span_factory.create(offset, offset)));
     };
     offset += 1; // '/'
 
@@ -84,9 +87,10 @@ fn parse_reg_exp_literal(source_text: &str) -> Result<(usize, usize, usize)> {
         match chars.peek() {
             // Line terminators are not allowed
             Some('\u{a}' | '\u{d}' | '\u{2028}' | '\u{2029}') | None => {
-                let kind =
-                    if in_character_class { "character class" } else { "regular expression" };
-                return Err(OxcDiagnostic::error(format!("Unterminated {kind}")));
+                return Err(diagnostics::unterminated_literal(
+                    span_factory.create(body_start, offset),
+                    if in_character_class { "character class" } else { "regular expression" },
+                ));
             }
             Some(&ch) => {
                 if in_escape {
@@ -112,12 +116,12 @@ fn parse_reg_exp_literal(source_text: &str) -> Result<(usize, usize, usize)> {
     }
 
     let Some('/') = chars.next() else {
-        return Err(OxcDiagnostic::error("Invalid regular expression: Unexpected character"));
+        return Err(diagnostics::unexpected_literal_char(span_factory.create(offset, offset)));
     };
     let body_end = offset;
 
     if body_end == body_start {
-        return Err(OxcDiagnostic::error("Invalid regular expression: Empty"));
+        return Err(diagnostics::empty_literal(span_factory.create(0, body_end + 1)));
     }
 
     Ok((body_start, body_end, body_end + 1))
@@ -141,7 +145,7 @@ mod test {
             "/👈🏻こっち/u",
         ] {
             let (body_start_offset, body_end_offset, flag_start_offset) =
-                parse_reg_exp_literal(literal_text)
+                parse_reg_exp_literal(literal_text, &SpanFactory::new(0))
                     .unwrap_or_else(|_| panic!("{literal_text} should be parsed"));
 
             let body_text = &literal_text[body_start_offset..body_end_offset];
@@ -155,7 +159,7 @@ mod test {
         for literal_text in
             ["", "foo", ":(", "a\nb", "/", "/x", "/y\nz/", "/1[\n]/", "//", "///", "/*abc/", "/\\/"]
         {
-            assert!(parse_reg_exp_literal(literal_text).is_err());
+            assert!(parse_reg_exp_literal(literal_text, &SpanFactory::new(0)).is_err());
         }
     }
 }