eUgEntOptIc44/mrab-regex
1
1
Fork 0
mirror of https://github.com/mrabarnett/mrab-regex.git synced 2025-10-05 20:02:39 +02:00
Code Issues Releases Activity

The escape function no longer escapes \x00. It's not necessary.

Browse Source 
Inline flags can now be turned off and apply to what follows.

Added \R to match line endings.
This commit is contained in:
Matthew Barnett
2023-12-24 00:13:19 +00:00
parent 34333d5009
commit cdcbf36aeb
7 changed files with 54 additions and 58 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
.github/workflows/main.yml vendored
View File

@@ -45,7 +45,7 @@ jobs:
env:
# macOS archs
CIBW_ARCHS_MACOS: "x86_64 arm64"
CIBW_ARCHS_MACOS: "x86_64 arm64 universal2"
steps:
- uses: actions/checkout@v3
@@ -64,6 +64,14 @@ jobs:
name: regex-files
path: wheelhouse/*.whl
- name: Create GitHub release
uses: actions/create-release@v1
env:
GITHUB_TOKEN: ${{ secrets.RELEASE_TOKEN }}
with:
tag_name: ${{ github.ref }}
title: Release ${{ github.ref }}
# Build source distribution & manylinux1_x86_64 wheels
# These two jobs build:
# 1, build_wheels (above): manylinux1_i686 / manylinux2014_x86_64

8
changelog.txt
View File

@@ -1,3 +1,11 @@
Version: 2023.12.23
The escape function no longer escapes \x00. It's not necessary.
Inline flags can now be turned off and apply to what follows.
Added \R to match line endings.
Version: 2023.10.3
Updated to Unicode 15.1.0.

25
regex_3/_regex_core.py
View File

@@ -1150,22 +1150,7 @@ def parse_flags_subpattern(source, info):
def parse_positional_flags(source, info, flags_on, flags_off):
"Parses positional flags."
version = (info.flags & _ALL_VERSIONS) or DEFAULT_VERSION
if version == VERSION0:
# Positional flags are global and can only be turned on.
if flags_off:
raise error("bad inline flags: cannot turn flags off",
source.string, source.pos)
new_global_flags = flags_on & ~info.global_flags
if new_global_flags:
info.global_flags |= new_global_flags
# A global has been turned on, so reparse the pattern.
raise _UnscopedFlagSet(info.global_flags)
else:
info.flags = (info.flags | flags_on) & ~flags_off
info.flags = (info.flags | flags_on) & ~flags_off
source.ignore_space = bool(info.flags & VERBOSE)
def parse_name(source, allow_numeric=False, allow_group_0=False):
@@ -1233,6 +1218,14 @@ def parse_escape(source, info, in_set):
elif ch in "pP":
# A Unicode property, positive or negative.
return parse_property(source, info, ch == "p", in_set)
elif ch == "R" and not in_set:
# A line ending.
charset = [0x0A, 0x0B, 0x0C, 0x0D]
if info.guess_encoding == UNICODE:
charset.extend([0x85, 0x2028, 0x2029])
return Atomic(Branch([String([0x0D, 0x0A]), SetUnion(info, [Character(c)
for c in charset])]))
elif ch == "X" and not in_set:
# A grapheme cluster.
return Grapheme()

6
regex_3/regex.py
View File

@@ -241,7 +241,7 @@ __all__ = ["cache_all", "compile", "DEFAULT_VERSION", "escape", "findall",
"VERSION1", "X", "VERBOSE", "W", "WORD", "error", "Regex", "__version__",
"__doc__", "RegexFlag"]
__version__ = "2.5.135"
__version__ = "2.5.136"
# --------------------------------------------------------------------
# Public interface.
@@ -392,8 +392,6 @@ def escape(pattern, special_only=True, literal_spaces=False):
elif c in _METACHARS or c.isspace():
s.append("\\")
s.append(c)
elif c == "\x00":
s.append("\\000")
else:
s.append(c)
else:
@@ -402,8 +400,6 @@ def escape(pattern, special_only=True, literal_spaces=False):
s.append(c)
elif c in _ALNUM:
s.append(c)
elif c == "\x00":
s.append("\\000")
else:
s.append("\\")
s.append(c)

59
regex_3/test_regex.py
View File

@@ -911,10 +911,9 @@ class RegexTests(unittest.TestCase):
p = regex.compile('(?iu)' + lower_char)
self.assertEqual(bool(p.match(upper_char)), True)
# Changed to positional flags in regex 2023.12.23.
self.assertEqual(bool(regex.match(r"(?i)a", "A")), True)
self.assertEqual(bool(regex.match(r"a(?i)", "A")), True)
self.assertEqual(bool(regex.match(r"(?iV1)a", "A")), True)
self.assertEqual(regex.match(r"a(?iV1)", "A"), None)
self.assertEqual(regex.match(r"a(?i)", "A"), None)
def test_dollar_matches_twice(self):
# $ matches the end of string, and just before the terminating \n.
@@ -1396,18 +1395,15 @@ class RegexTests(unittest.TestCase):
# Issues 433028, 433024, 433027.
self.assertEqual(regex.search(r"(?i)Ab", "ab").span(), (0, 2))
self.assertEqual(regex.search(r"(?i:A)b", "ab").span(), (0, 2))
self.assertEqual(regex.search(r"A(?i)b", "ab").span(), (0, 2))
self.assertEqual(regex.search(r"A(?iV1)b", "ab"), None)
self.assertRaisesRegex(regex.error, self.CANT_TURN_OFF, lambda:
regex.search(r"(?V0-i)Ab", "ab", flags=regex.I))
# Changed to positional flags in regex 2023.12.23.
self.assertEqual(regex.search(r"A(?i)b", "ab"), None)
self.assertEqual(regex.search(r"(?V0)Ab", "ab"), None)
self.assertEqual(regex.search(r"(?V1)Ab", "ab"), None)
self.assertEqual(regex.search(r"(?V1-i)Ab", "ab", flags=regex.I), None)
self.assertEqual(regex.search(r"(?-i)Ab", "ab", flags=regex.I), None)
self.assertEqual(regex.search(r"(?-i:A)b", "ab", flags=regex.I), None)
self.assertEqual(regex.search(r"A(?V1-i)b", "ab",
flags=regex.I).span(), (0, 2))
self.assertEqual(regex.search(r"A(?-i)b", "ab", flags=regex.I).span(),
(0, 2))
def test_repeated_repeats(self):
# Issue 2537.
@@ -1820,12 +1816,10 @@ class RegexTests(unittest.TestCase):
('a.*b', 'acc\nccb', '', ascii(None)),
('a.{4,5}b', 'acc\nccb', '', ascii(None)),
('a.b', 'a\rb', '0', ascii('a\rb')),
# The new behaviour is that the inline flag affects only what follows.
('a.b(?s)', 'a\nb', '0', ascii('a\nb')),
('a.b(?sV1)', 'a\nb', '', ascii(None)),
# Changed to positional flags in regex 2023.12.23.
('a.b(?s)', 'a\nb', '', ascii(None)),
('(?s)a.b', 'a\nb', '0', ascii('a\nb')),
('a.*(?s)b', 'acc\nccb', '0', ascii('acc\nccb')),
('a.*(?sV1)b', 'acc\nccb', '', ascii(None)),
('a.*(?s)b', 'acc\nccb', '', ascii(None)),
('(?s)a.*b', 'acc\nccb', '0', ascii('acc\nccb')),
('(?s)a.{4,5}b', 'acc\nccb', '0', ascii('acc\nccb')),
@@ -2345,12 +2339,9 @@ class RegexTests(unittest.TestCase):
# Not an error under PCRE/PRE:
# When the new behaviour is turned on positional inline flags affect
# only what follows.
('w(?i)', 'W', '0', ascii('W')),
('w(?iV1)', 'W', '0', ascii(None)),
('w(?i)', 'W', '0', ascii(None)),
('w(?i)', 'w', '0', ascii('w')),
('w(?iV1)', 'w', '0', ascii('w')),
('(?i)w', 'W', '0', ascii('W')),
('(?iV1)w', 'W', '0', ascii('W')),
# Comments using the x embedded pattern modifier.
("""(?x)w# comment 1
@@ -2403,14 +2394,10 @@ xyzabc
# Bug 114033: nothing to repeat.
(r'(x?)?', 'x', '0', ascii('x')),
# Bug 115040: rescan if flags are modified inside pattern.
# If the new behaviour is turned on then positional inline flags
# affect only what follows.
(r' (?x)foo ', 'foo', '0', ascii('foo')),
(r' (?V1x)foo ', 'foo', '0', ascii(None)),
# Changed to positional flags in regex 2023.12.23.
(r' (?x)foo ', 'foo', '0', ascii(None)),
(r'(?x) foo ', 'foo', '0', ascii('foo')),
(r'(?V1x) foo ', 'foo', '0', ascii('foo')),
(r'(?x)foo ', 'foo', '0', ascii('foo')),
(r'(?V1x)foo ', 'foo', '0', ascii('foo')),
# Bug 115618: negative lookahead.
(r'(?<!abc)(d.f)', 'abcdefdof', '0', ascii('dof')),
# Bug 116251: character class bug.
@@ -3154,10 +3141,8 @@ xyzabc
# Hg issue 39: regex.search("((?i)blah)\\s+\\1", "blah BLAH") doesn't
# return None
self.assertEqual(regex.search(r"(?V0)((?i)blah)\s+\1",
"blah BLAH").group(0, 1), ("blah BLAH", "blah"))
self.assertEqual(regex.search(r"(?V1)((?i)blah)\s+\1", "blah BLAH"),
None)
# Changed to positional flags in regex 2023.12.23.
self.assertEqual(regex.search(r"((?i)blah)\s+\1", "blah BLAH"), None)
# Hg issue 40: regex.search("(\()?[^()]+(?(1)\)|)", "(abcd").group(0)
# returns "bcd" instead of "abcd"
@@ -4336,10 +4321,10 @@ thing
self.assertEqual(regex.search(r"^a?(a?)b?c\1$", "abca").span(), (0, 4))
# Git issue 498: Conditional negative lookahead inside positive lookahead fails to match
self.assertEqual(regex.match(r"(?(?=a).|..)", "ab").span(), (0, 1))
self.assertEqual(regex.match(r"(?(?=b).|..)", "ab").span(), (0, 2))
self.assertEqual(regex.match(r"(?(?!a).|..)", "ab").span(), (0, 2))
self.assertEqual(regex.match(r"(?(?!b).|..)", "ab").span(), (0, 1))
self.assertEqual(regex.match(r'(?(?=a).|..)', 'ab').span(), (0, 1))
self.assertEqual(regex.match(r'(?(?=b).|..)', 'ab').span(), (0, 2))
self.assertEqual(regex.match(r'(?(?!a).|..)', 'ab').span(), (0, 2))
self.assertEqual(regex.match(r'(?(?!b).|..)', 'ab').span(), (0, 1))
def test_fuzzy_ext(self):
self.assertEqual(bool(regex.fullmatch(r'(?r)(?:a){e<=1:[a-z]}', 'e')),
@@ -4460,6 +4445,12 @@ thing
self.assertEqual([m.span() for m in regex.finditer(r'(?m)^\s*?$',
'foo\n\n\nbar')], [(4, 4), (4, 5), (5, 5)])
def test_line_ending(self):
self.assertEqual(regex.findall(r'\R', '\r\n\n\x0B\f\r\x85\u2028\u2029'),
['\r\n', '\n', '\x0B', '\f', '\r', '\x85', '\u2028', '\u2029'])
self.assertEqual(regex.findall(br'\R', b'\r\n\n\x0B\f\r\x85'), [b'\r\n',
b'\n', b'\x0B', b'\f', b'\r'])
def test_main():
unittest.main(verbosity=2)

2
setup.py
View File

@@ -7,7 +7,7 @@ with open('README.rst', encoding='utf-8') as file:
setup(
name='regex',
version='2023.10.3',
version='2023.12.23',
description='Alternative regular expression module, to replace re.',
long_description=long_description,
long_description_content_type='text/x-rst',

2
tools/build_regex_unicode.py
View File

@@ -1781,4 +1781,4 @@ binary_dict = make_binary_dict()
generate_code(unicode_data, UNICODE_VERSION, this_folder)
print('\nSuccessfully generated _reges_unicode.h and _reges_unicode.c in %s' % tools_folder)
print('\nSuccessfully generated _regex_unicode.h and _regex_unicode.c in %s' % tools_folder)