Performance regression in `shlex.quote` from 3.13 to 3.14

Bug report

Bug description:

#132036 included an algorithmic change to shlex.quote that made it slower when the input has to be quoted. This is because the regular expression search was able to short-circuit at the first unsafe character.

However, the isascii check is worthwhile

Cc @picnixz

import re
import shlex
import timeit

# From 3.13
_find_unsafe = re.compile(r'[^\w@%+=:,./-]', re.ASCII).search
def old_quote(s):
    """Return a shell-escaped version of the string *s*."""
    if not s:
        return "''"
    # BEST: if s.isascii() and _find_unsafe(s) is None:
    if _find_unsafe(s) is None:
        return s

    # use single quotes, and put single quotes into double quotes
    # the string $'b is then quoted as '$'"'"'b'
    return "'" + s.replace("'", "'\"'\"'") + "'"


g = {'old_quote': old_quote, 'new_quote': shlex.quote}
print('with spaces')
print('  old', timeit.timeit("old_quote('the quick brown fox jumps over the lazy dog')", globals=g, number=1000000))
print('  new', timeit.timeit("new_quote('the quick brown fox jumps over the lazy dog')", globals=g, number=1000000))
print('without spaces')
print('  old', timeit.timeit("old_quote('thequickbrownfoxjumpsoverthelazydog')", globals=g, number=1000000))
print('  new', timeit.timeit("new_quote('thequickbrownfoxjumpsoverthelazydog')", globals=g, number=1000000))
print('non-ASCII')
print('  old', timeit.timeit("old_quote('mötley')", globals=g, number=1000000))
print('  new', timeit.timeit("new_quote('mötley')", globals=g, number=1000000))
print('short')
print('  old', timeit.timeit("old_quote('a')", globals=g, number=1000000))
print('  new', timeit.timeit("new_quote('a')", globals=g, number=1000000))

sample output:

with spaces
  old 0.4148377259989502
  new 0.5036935329990229
without spaces
  old 0.3872929839999415
  new 0.3540855330065824
ascii
  old 0.4636239370011026
  new 0.20726546400692314
short
  old 0.1217202929983614
  new 0.2977778149943333

CPython versions tested on:

3.14

Operating systems tested on:

Linux

Linked PRs

• gh-146408

Bắt đầu từ entry point shlex.quote và so sánh hành vi hiện tại của nó với implementation 3.13 được nêu trong báo cáo. Xem lại PR được liên kết gh-146408, sau đó sử dụng các trường hợp timeit được cung cấp để xác minh rằng hồi quy đã được khắc phục.