From 70ca0210e8958d2665541ddd38fce2965075674e Mon Sep 17 00:00:00 2001 From: Serhiy Storchaka Date: Sat, 16 Feb 2013 16:47:47 +0200 Subject: [PATCH] Issue #13169: The maximal repetition number in a regular expression has been increased from 65534 to 2147483647 (on 32-bit platform) or 4294967294 (on 64-bit). --- Lib/sre_compile.py | 1 + Lib/sre_constants.py | 4 ---- Lib/sre_parse.py | 9 +++++++-- Lib/test/test_re.py | 34 +++++++++++++++++++++++++++++++++- Misc/NEWS | 4 ++++ Modules/_sre.c | 18 ++++++++++++------ Modules/sre.h | 5 +++++ 7 files changed, 62 insertions(+), 13 deletions(-) diff --git a/Lib/sre_compile.py b/Lib/sre_compile.py index f52ea014597..46eac9c070a 100644 --- a/Lib/sre_compile.py +++ b/Lib/sre_compile.py @@ -13,6 +13,7 @@ import _sre, sys import sre_parse from sre_constants import * +from _sre import MAXREPEAT assert _sre.MAGIC == MAGIC, "SRE module mismatch" diff --git a/Lib/sre_constants.py b/Lib/sre_constants.py index 417670bf8db..71ccb23f621 100644 --- a/Lib/sre_constants.py +++ b/Lib/sre_constants.py @@ -15,10 +15,6 @@ MAGIC = 20031017 -# max code word in this release - -MAXREPEAT = 65535 - # SRE standard exception (access as sre.error) # should this really be here? diff --git a/Lib/sre_parse.py b/Lib/sre_parse.py index 19dd4fc4bce..045a5ebfef5 100644 --- a/Lib/sre_parse.py +++ b/Lib/sre_parse.py @@ -15,6 +15,7 @@ import sys from sre_constants import * +from _sre import MAXREPEAT SPECIAL_CHARS = ".\\[{()*+?^$|" REPEAT_CHARS = "*+?{" @@ -505,10 +506,14 @@ def _parse(source, state): continue if lo: min = int(lo) + if min >= MAXREPEAT: + raise OverflowError("the repetition number is too large") if hi: max = int(hi) - if max < min: - raise error("bad repeat interval") + if max >= MAXREPEAT: + raise OverflowError("the repetition number is too large") + if max < min: + raise error("bad repeat interval") else: raise error("not supported") # figure out which item to repeat diff --git a/Lib/test/test_re.py b/Lib/test/test_re.py index 6b047e48dbb..f7e76dcfbad 100644 --- a/Lib/test/test_re.py +++ b/Lib/test/test_re.py @@ -1,4 +1,5 @@ -from test.support import verbose, run_unittest, gc_collect, bigmemtest, _2G +from test.support import verbose, run_unittest, gc_collect, bigmemtest, _2G, \ + cpython_only import io import re from re import Scanner @@ -883,6 +884,37 @@ class ReTests(unittest.TestCase): self.assertEqual(n, size + 1) + def test_repeat_minmax_overflow(self): + # Issue #13169 + string = "x" * 100000 + self.assertEqual(re.match(r".{65535}", string).span(), (0, 65535)) + self.assertEqual(re.match(r".{,65535}", string).span(), (0, 65535)) + self.assertEqual(re.match(r".{65535,}?", string).span(), (0, 65535)) + self.assertEqual(re.match(r".{65536}", string).span(), (0, 65536)) + self.assertEqual(re.match(r".{,65536}", string).span(), (0, 65536)) + self.assertEqual(re.match(r".{65536,}?", string).span(), (0, 65536)) + # 2**128 should be big enough to overflow both SRE_CODE and Py_ssize_t. + self.assertRaises(OverflowError, re.compile, r".{%d}" % 2**128) + self.assertRaises(OverflowError, re.compile, r".{,%d}" % 2**128) + self.assertRaises(OverflowError, re.compile, r".{%d,}?" % 2**128) + self.assertRaises(OverflowError, re.compile, r".{%d,%d}" % (2**129, 2**128)) + + @cpython_only + def test_repeat_minmax_overflow_maxrepeat(self): + try: + from _sre import MAXREPEAT + except ImportError: + self.skipTest('requires _sre.MAXREPEAT constant') + string = "x" * 100000 + self.assertIsNone(re.match(r".{%d}" % (MAXREPEAT - 1), string)) + self.assertEqual(re.match(r".{,%d}" % (MAXREPEAT - 1), string).span(), + (0, 100000)) + self.assertIsNone(re.match(r".{%d,}?" % (MAXREPEAT - 1), string)) + self.assertRaises(OverflowError, re.compile, r".{%d}" % MAXREPEAT) + self.assertRaises(OverflowError, re.compile, r".{,%d}" % MAXREPEAT) + self.assertRaises(OverflowError, re.compile, r".{%d,}?" % MAXREPEAT) + + def run_re_tests(): from test.re_tests import tests, SUCCEED, FAIL, SYNTAX_ERROR if verbose: diff --git a/Misc/NEWS b/Misc/NEWS index 4bc6584c043..83405c424b7 100644 --- a/Misc/NEWS +++ b/Misc/NEWS @@ -224,6 +224,10 @@ Core and Builtins Library ------- +- Issue #13169: The maximal repetition number in a regular expression has been + increased from 65534 to 2147483647 (on 32-bit platform) or 4294967294 (on + 64-bit). + - Issue #16743: Fix mmap overflow check on 32 bit Windows. - Issue #16800: tempfile.gettempdir() no longer left temporary files when diff --git a/Modules/_sre.c b/Modules/_sre.c index 88bbf6a941e..4421eae83ac 100644 --- a/Modules/_sre.c +++ b/Modules/_sre.c @@ -517,7 +517,7 @@ SRE_COUNT(SRE_STATE* state, SRE_CODE* pattern, Py_ssize_t maxcount) Py_ssize_t i; /* adjust end */ - if (maxcount < end - ptr && maxcount != 65535) + if (maxcount < end - ptr && maxcount != SRE_MAXREPEAT) end = ptr + maxcount; switch (pattern[0]) { @@ -1132,7 +1132,7 @@ entrance: } else { /* general case */ LASTMARK_SAVE(); - while ((Py_ssize_t)ctx->pattern[2] == 65535 + while ((Py_ssize_t)ctx->pattern[2] == SRE_MAXREPEAT || ctx->count <= (Py_ssize_t)ctx->pattern[2]) { state->ptr = ctx->ptr; DO_JUMP(JUMP_MIN_REPEAT_ONE,jump_min_repeat_one, @@ -1218,7 +1218,7 @@ entrance: } if ((ctx->count < ctx->u.rep->pattern[2] || - ctx->u.rep->pattern[2] == 65535) && + ctx->u.rep->pattern[2] == SRE_MAXREPEAT) && state->ptr != ctx->u.rep->last_ptr) { /* we may have enough matches, but if we can match another item, do so */ @@ -1296,7 +1296,7 @@ entrance: LASTMARK_RESTORE(); if (ctx->count >= ctx->u.rep->pattern[2] - && ctx->u.rep->pattern[2] != 65535) + && ctx->u.rep->pattern[2] != SRE_MAXREPEAT) RETURN_FAILURE; ctx->u.rep->count = ctx->count; @@ -3072,7 +3072,7 @@ _validate_inner(SRE_CODE *code, SRE_CODE *end, Py_ssize_t groups) GET_ARG; max = arg; if (min > max) FAIL; - if (max > 65535) + if (max > SRE_MAXREPEAT) FAIL; if (!_validate_inner(code, code+skip-4, groups)) FAIL; @@ -3091,7 +3091,7 @@ _validate_inner(SRE_CODE *code, SRE_CODE *end, Py_ssize_t groups) GET_ARG; max = arg; if (min > max) FAIL; - if (max > 65535) + if (max > SRE_MAXREPEAT) FAIL; if (!_validate_inner(code, code+skip-3, groups)) FAIL; @@ -3979,6 +3979,12 @@ PyMODINIT_FUNC PyInit__sre(void) Py_DECREF(x); } + x = PyLong_FromUnsignedLong(SRE_MAXREPEAT); + if (x) { + PyDict_SetItemString(d, "MAXREPEAT", x); + Py_DECREF(x); + } + x = PyUnicode_FromString(copyright); if (x) { PyDict_SetItemString(d, "copyright", x); diff --git a/Modules/sre.h b/Modules/sre.h index d389b46dd13..68196985584 100644 --- a/Modules/sre.h +++ b/Modules/sre.h @@ -16,6 +16,11 @@ /* size of a code word (must be unsigned short or larger, and large enough to hold a UCS4 character) */ #define SRE_CODE Py_UCS4 +#if SIZEOF_SIZE_T > 4 +# define SRE_MAXREPEAT (~(SRE_CODE)0) +#else +# define SRE_MAXREPEAT ((SRE_CODE)PY_SSIZE_T_MAX + 1u) +#endif typedef struct { PyObject_VAR_HEAD