diff options
| author | shadchin <shadchin@yandex-team.com> | 2024-12-23 19:39:02 +0300 |
|---|---|---|
| committer | shadchin <shadchin@yandex-team.com> | 2024-12-23 19:54:20 +0300 |
| commit | 65a5bf9d37a3b29eb394f560b9a09318196c40e8 (patch) | |
| tree | e5cd68fb0682b2388e52d9806bb87adc348e21a8 /contrib/tools/python3/Lib/re/_compiler.py | |
| parent | a1dd87a52878ab3e46e5fd2dba5ecbba6113d7e0 (diff) | |
| download | ydb-65a5bf9d37a3b29eb394f560b9a09318196c40e8.tar.gz | |
Update Python 3 to 3.12.8
commit_hash:c20045b8a987d8720e1f3328270357491d5530f3
Diffstat (limited to 'contrib/tools/python3/Lib/re/_compiler.py')
| -rw-r--r-- | contrib/tools/python3/Lib/re/_compiler.py | 23 |
1 files changed, 14 insertions, 9 deletions
diff --git a/contrib/tools/python3/Lib/re/_compiler.py b/contrib/tools/python3/Lib/re/_compiler.py index 285c21936f..bb97f9fdd1 100644 --- a/contrib/tools/python3/Lib/re/_compiler.py +++ b/contrib/tools/python3/Lib/re/_compiler.py @@ -250,11 +250,11 @@ def _optimize_charset(charset, iscased=None, fixup=None, fixes=None): while True: try: if op is LITERAL: - if fixup: - lo = fixup(av) - charmap[lo] = 1 - if fixes and lo in fixes: - for k in fixes[lo]: + if fixup: # IGNORECASE and not LOCALE + av = fixup(av) + charmap[av] = 1 + if fixes and av in fixes: + for k in fixes[av]: charmap[k] = 1 if not hascased and iscased(av): hascased = True @@ -262,7 +262,7 @@ def _optimize_charset(charset, iscased=None, fixup=None, fixes=None): charmap[av] = 1 elif op is RANGE: r = range(av[0], av[1]+1) - if fixup: + if fixup: # IGNORECASE and not LOCALE if fixes: for i in map(fixup, r): charmap[i] = 1 @@ -289,8 +289,7 @@ def _optimize_charset(charset, iscased=None, fixup=None, fixes=None): # Character set contains non-BMP character codes. # For range, all BMP characters in the range are already # proceeded. - if fixup: - hascased = True + if fixup: # IGNORECASE and not LOCALE # For now, IN_UNI_IGNORE+LITERAL and # IN_UNI_IGNORE+RANGE_UNI_IGNORE work for all non-BMP # characters, because two characters (at least one of @@ -301,7 +300,13 @@ def _optimize_charset(charset, iscased=None, fixup=None, fixes=None): # Also, both c.lower() and c.lower().upper() are single # characters for every non-BMP character. if op is RANGE: - op = RANGE_UNI_IGNORE + if fixes: # not ASCII + op = RANGE_UNI_IGNORE + hascased = True + else: + assert op is LITERAL + if not hascased and iscased(av): + hascased = True tail.append((op, av)) break |