perf(es/parser): Optimize skip_space

.changeset/fluffy-cobras-agree.md

@@ -0,0 +1,6 @@
+---
+swc_core: patch
+swc_ecma_parser: patch
+---
+
+perf(es/parser): optimize `skip_space`

crates/swc_ecma_parser/src/lexer/mod.rs

@@ -260,7 +260,7 @@ impl<'a> Lexer<'a> {
             if self.state.had_line_break && C == b'-' && self.eat(b'>') {
                 self.emit_module_mode_error(start, SyntaxError::LegacyCommentInModule);
                 self.skip_line_comment(0);
-                self.skip_space::<true>();
+                self.skip_space();
                 return self.read_token();
             }
 
@@ -303,7 +303,7 @@ impl<'a> Lexer<'a> {
                     if had_line_break_before_last && self.is_str("====") {
                         self.emit_error_span(fixed_len_span(start, 7), SyntaxError::TS1185);
                         self.skip_line_comment(4);
-                        self.skip_space::<true>();
+                        self.skip_space();
                         return self.read_token();
                     }
 
@@ -347,7 +347,7 @@ impl Lexer<'_> {
         if C == b'<' && self.is(b'!') && self.peek() == Some('-') && self.peek_ahead() == Some('-')
         {
             self.skip_line_comment(3);
-            self.skip_space::<true>();
+            self.skip_space();
             self.emit_module_mode_error(start, SyntaxError::LegacyCommentInModule);
 
             return self.read_token();
@@ -399,7 +399,7 @@ impl Lexer<'_> {
         {
             self.emit_error_span(fixed_len_span(start, 7), SyntaxError::TS1185);
             self.skip_line_comment(5);
-            self.skip_space::<true>();
+            self.skip_space();
             return self.read_token();
         }
 
@@ -871,45 +871,6 @@ impl<'a> Lexer<'a> {
         }
     }
 
-    /// Skip comments or whitespaces.
-    ///
-    /// See https://tc39.github.io/ecma262/#sec-white-space
-    #[inline(never)]
-    fn skip_space<const LEX_COMMENTS: bool>(&mut self) {
-        loop {
-            let (offset, newline) = {
-                let mut skip = self::whitespace::SkipWhitespace {
-                    input: self.input().as_str(),
-                    newline: false,
-                    offset: 0,
-                };
-
-                skip.scan();
-
-                (skip.offset, skip.newline)
-            };
-
-            self.input_mut().bump_bytes(offset as usize);
-            if newline {
-                self.state_mut().mark_had_line_break();
-            }
-
-            if LEX_COMMENTS && self.input().is_byte(b'/') {
-                if let Some(c) = self.peek() {
-                    if c == '/' {
-                        self.skip_line_comment(2);
-                        continue;
-                    } else if c == '*' {
-                        self.skip_block_comment();
-                        continue;
-                    }
-                }
-            }
-
-            break;
-        }
-    }
-
     /// Ensure that ident cannot directly follow numbers.
     fn ensure_not_ident(&mut self) -> LexResult<()> {
         match self.cur() {
@@ -2178,7 +2139,7 @@ impl<'a> Lexer<'a> {
                 let span = fixed_len_span(start, 7);
                 self.emit_error_span(span, SyntaxError::TS1185);
                 self.skip_line_comment(5);
-                self.skip_space::<true>();
+                self.skip_space();
                 return self.error_span(span, SyntaxError::TS1185);
             }
 

crates/swc_ecma_parser/src/lexer/state.rs

@@ -211,7 +211,7 @@ impl crate::input::Tokens for Lexer<'_> {
     }
 
     fn scan_jsx_open_el_terminal_token(&mut self) -> TokenAndSpan {
-        self.skip_space::<true>();
+        self.skip_space();
         let start = self.input.cur_pos();
         let res = match self.scan_jsx_attrs_terminal_token() {
             Ok(res) => Ok(res),
@@ -382,7 +382,7 @@ impl Lexer<'_> {
         self.state.had_line_break = self.state.is_first;
         self.state.is_first = false;
 
-        self.skip_space::<true>();
+        self.skip_space();
         *start = self.input.cur_pos();
 
         if self.input.last_pos() == self.input.end_pos() {

crates/swc_ecma_parser/src/lexer/table.rs

@@ -19,8 +19,9 @@ use crate::{
 pub(super) type ByteHandler = fn(&mut Lexer<'_>) -> LexResult<Token>;
 
 /// Lookup table mapping any incoming byte to a handler function defined below.
+#[rustfmt::skip]
 pub(super) static BYTE_HANDLERS: [ByteHandler; 256] = [
-    //   0    1    2    3    4    5    6    7    8    9    A    B    C    D    E    F   //
+//   0    1    2    3    4    5    6    7    8    9    A    B    C    D    E    F   //
     ERR, ERR, ERR, ERR, ERR, ERR, ERR, ERR, ERR, ERR, ERR, ERR, ERR, ERR, ERR, ERR, // 0
     ERR, ERR, ERR, ERR, ERR, ERR, ERR, ERR, ERR, ERR, ERR, ERR, ERR, ERR, ERR, ERR, // 1
     ERR, EXL, QOT, HSH, IDN, PRC, AMP, QOT, PNO, PNC, ATR, PLS, COM, MIN, PRD, SLH, // 2