[webkit-changes] [WebKit/WebKit] 8e5b3f: [JSC] Improve RegExp Lookbehind Character Class Ba...

[Michael Saboff]

  Branch: refs/heads/main
  Home:   https://github.com/WebKit/WebKit
  Commit: 8e5b3f5761f011abfa644ddab3f0f1f6561ec5fd
      https://github.com/WebKit/WebKit/commit/8e5b3f5761f011abfa644ddab3f0f1f6561ec5fd
  Author: Michael Saboff <msaboff at apple.com>
  Date:   2023-06-28 (Wed, 28 Jun 2023)

  Changed paths:
    M JSTests/stress/regexp-lookaround-captures.js
    M Source/JavaScriptCore/yarr/YarrInterpreter.cpp

  Log Message:
  -----------
  [JSC] Improve RegExp Lookbehind Character Class Backtracking
https://bugs.webkit.org/show_bug.cgi?id=258531
rdar://111051833

Reviewed by Yusuke Suzuki.

Changed character class backtracking in the Yarr interpreter to unread one codepoint instead of rematching
one less codepoint.  This can dramatically improve matching performance for large strings where we need to
backtrack a large number of codepoints.

Added test cases that specifically test backtracking character classes that need to backtrack both BMP and
non-BMP codepoints when processing the input string.

* JSTests/stress/regexp-lookaround-captures.js:
* Source/JavaScriptCore/yarr/YarrInterpreter.cpp:
(JSC::Yarr::Interpreter::matchCharacterClass):
(JSC::Yarr::Interpreter::backtrackCharacterClass):

Canonical link: https://commits.webkit.org/265593@main