[clang] [Clang][Lexer][Performance] Optimize Lexer whitespace skipping logic (PR #180819)
Thibault Monnier via cfe-commits
cfe-commits at lists.llvm.org
Sat Feb 14 07:04:14 PST 2026
https://github.com/Thibault-Monnier updated https://github.com/llvm/llvm-project/pull/180819
>From faa899a6ce518c1176f2bf59f199eb42e59d840e Mon Sep 17 00:00:00 2001
From: Thibault-Monnier <thibaultmonni at gmail.com>
Date: Tue, 10 Feb 2026 19:41:47 +0100
Subject: [PATCH] Try prioritizing skipping space
---
clang/lib/Lex/Lexer.cpp | 8 ++++----
1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/clang/lib/Lex/Lexer.cpp b/clang/lib/Lex/Lexer.cpp
index 1498657047bd6..483cca32e08a2 100644
--- a/clang/lib/Lex/Lexer.cpp
+++ b/clang/lib/Lex/Lexer.cpp
@@ -2533,8 +2533,8 @@ bool Lexer::SkipWhitespace(Token &Result, const char *CurPtr,
// Skip consecutive spaces efficiently.
while (true) {
- // Skip horizontal whitespace very aggressively.
- while (isHorizontalWhitespace(Char))
+ // Skip horizontal whitespace, especially space, very aggressively.
+ while (LLVM_LIKELY(Char == ' ') || isHorizontalWhitespace(Char))
Char = *++CurPtr;
// Otherwise if we have something other than whitespace, we're done.
@@ -3756,10 +3756,10 @@ bool Lexer::LexTokenInternal(Token &Result, bool TokAtPhysicalStartOfLine) {
const char *CurPtr = BufferPtr;
// Small amounts of horizontal whitespace is very common between tokens.
- if (isHorizontalWhitespace(*CurPtr)) {
+ if (LLVM_LIKELY(*CurPtr == ' ') || isHorizontalWhitespace(*CurPtr)) {
do {
++CurPtr;
- } while (isHorizontalWhitespace(*CurPtr));
+ } while (LLVM_LIKELY(*CurPtr == ' ') || isHorizontalWhitespace(*CurPtr));
// If we are keeping whitespace and other tokens, just return what we just
// skipped. The next lexer invocation will return the token after the
More information about the cfe-commits
mailing list