|
 |
f1a5bc |
|
|
|
f1a5bc |
# HG changeset patch
|
|
|
f1a5bc |
# User Serhiy Storchaka <storchaka@gmail.com>
|
|
|
f1a5bc |
# Date 1369166013 -10800
|
|
|
f1a5bc |
# Node ID 8408eed151ebee1c546414f1f40be46c1ad76077
|
|
|
f1a5bc |
# Parent 7fce9186accb10122e45d975f4b380c2ed0fae35
|
|
|
f1a5bc |
Issue #17979: Fixed the re module in build with --disable-unicode.
|
|
|
f1a5bc |
|
|
|
f1a5bc |
diff --git a/Modules/sre.h b/Modules/sre.h
|
|
|
f1a5bc |
--- a/Modules/sre.h
|
|
|
f1a5bc |
+++ b/Modules/sre.h
|
|
|
f1a5bc |
@@ -23,8 +23,8 @@
|
|
|
f1a5bc |
# define SRE_MAXREPEAT ((SRE_CODE)PY_SSIZE_T_MAX + 1u)
|
|
|
f1a5bc |
# endif
|
|
|
f1a5bc |
#else
|
|
|
f1a5bc |
-# define SRE_CODE unsigned long
|
|
|
f1a5bc |
-# if SIZEOF_SIZE_T > SIZEOF_LONG
|
|
|
f1a5bc |
+# define SRE_CODE unsigned int
|
|
|
f1a5bc |
+# if SIZEOF_SIZE_T > SIZEOF_INT
|
|
|
f1a5bc |
# define SRE_MAXREPEAT (~(SRE_CODE)0)
|
|
|
f1a5bc |
# else
|
|
|
f1a5bc |
# define SRE_MAXREPEAT ((SRE_CODE)PY_SSIZE_T_MAX + 1u)
|
|
|
f1a5bc |
|
|
|
f1a5bc |
|
|
|
f1a5bc |
# HG changeset patch
|
|
|
f1a5bc |
# User Serhiy Storchaka <storchaka@gmail.com>
|
|
|
f1a5bc |
# Date 1375547193 -10800
|
|
|
f1a5bc |
# Node ID e5e425fd1e4f7e859abdced43621203cdfa87a16
|
|
|
f1a5bc |
# Parent 8205e72b5cfcdb7a3450c80f3368eff610bc650c
|
|
|
f1a5bc |
Issue #17998: Fix an internal error in regular expression engine.
|
|
|
f1a5bc |
|
|
|
f1a5bc |
diff --git a/Lib/test/test_re.py b/Lib/test/test_re.py
|
|
|
f1a5bc |
--- a/Lib/test/test_re.py
|
|
|
f1a5bc |
+++ b/Lib/test/test_re.py
|
|
|
f1a5bc |
@@ -907,6 +907,16 @@ class ReTests(unittest.TestCase):
|
|
|
f1a5bc |
self.assertEqual(m.group(1), "")
|
|
|
f1a5bc |
self.assertEqual(m.group(2), "y")
|
|
|
f1a5bc |
|
|
|
f1a5bc |
+ def test_issue17998(self):
|
|
|
f1a5bc |
+ for reps in '*', '+', '?', '{1}':
|
|
|
f1a5bc |
+ for mod in '', '?':
|
|
|
f1a5bc |
+ pattern = '.' + reps + mod + 'yz'
|
|
|
f1a5bc |
+ self.assertEqual(re.compile(pattern, re.S).findall('xyz'),
|
|
|
f1a5bc |
+ ['xyz'], msg=pattern)
|
|
|
f1a5bc |
+ pattern = pattern.encode()
|
|
|
f1a5bc |
+ self.assertEqual(re.compile(pattern, re.S).findall(b'xyz'),
|
|
|
f1a5bc |
+ [b'xyz'], msg=pattern)
|
|
|
f1a5bc |
+
|
|
|
f1a5bc |
|
|
|
f1a5bc |
|
|
|
f1a5bc |
def run_re_tests():
|
|
|
f1a5bc |
diff --git a/Modules/_sre.c b/Modules/_sre.c
|
|
|
f1a5bc |
--- a/Modules/_sre.c
|
|
|
f1a5bc |
+++ b/Modules/_sre.c
|
|
|
f1a5bc |
@@ -1028,7 +1028,7 @@ entrance:
|
|
|
f1a5bc |
TRACE(("|%p|%p|REPEAT_ONE %d %d\n", ctx->pattern, ctx->ptr,
|
|
|
f1a5bc |
ctx->pattern[1], ctx->pattern[2]));
|
|
|
f1a5bc |
|
|
|
f1a5bc |
- if (ctx->pattern[1] > end - ctx->ptr)
|
|
|
f1a5bc |
+ if ((Py_ssize_t) ctx->pattern[1] > end - ctx->ptr)
|
|
|
f1a5bc |
RETURN_FAILURE; /* cannot match */
|
|
|
f1a5bc |
|
|
|
f1a5bc |
state->ptr = ctx->ptr;
|
|
|
f1a5bc |
@@ -1111,7 +1111,7 @@ entrance:
|
|
|
f1a5bc |
TRACE(("|%p|%p|MIN_REPEAT_ONE %d %d\n", ctx->pattern, ctx->ptr,
|
|
|
f1a5bc |
ctx->pattern[1], ctx->pattern[2]));
|
|
|
f1a5bc |
|
|
|
f1a5bc |
- if (ctx->pattern[1] > end - ctx->ptr)
|
|
|
f1a5bc |
+ if ((Py_ssize_t) ctx->pattern[1] > end - ctx->ptr)
|
|
|
f1a5bc |
RETURN_FAILURE; /* cannot match */
|
|
|
f1a5bc |
|
|
|
f1a5bc |
state->ptr = ctx->ptr;
|
|
|
f1a5bc |
@@ -1210,7 +1210,7 @@ entrance:
|
|
|
f1a5bc |
TRACE(("|%p|%p|MAX_UNTIL %d\n", ctx->pattern,
|
|
|
f1a5bc |
ctx->ptr, ctx->count));
|
|
|
f1a5bc |
|
|
|
f1a5bc |
- if (ctx->count < ctx->u.rep->pattern[1]) {
|
|
|
f1a5bc |
+ if (ctx->count < (Py_ssize_t) ctx->u.rep->pattern[1]) {
|
|
|
f1a5bc |
/* not enough matches */
|
|
|
f1a5bc |
ctx->u.rep->count = ctx->count;
|
|
|
f1a5bc |
DO_JUMP(JUMP_MAX_UNTIL_1, jump_max_until_1,
|
|
|
f1a5bc |
@@ -1224,7 +1224,7 @@ entrance:
|
|
|
f1a5bc |
RETURN_FAILURE;
|
|
|
f1a5bc |
}
|
|
|
f1a5bc |
|
|
|
f1a5bc |
- if ((ctx->count < ctx->u.rep->pattern[2] ||
|
|
|
f1a5bc |
+ if ((ctx->count < (Py_ssize_t) ctx->u.rep->pattern[2] ||
|
|
|
f1a5bc |
ctx->u.rep->pattern[2] == SRE_MAXREPEAT) &&
|
|
|
f1a5bc |
state->ptr != ctx->u.rep->last_ptr) {
|
|
|
f1a5bc |
/* we may have enough matches, but if we can
|
|
|
f1a5bc |
@@ -1273,7 +1273,7 @@ entrance:
|
|
|
f1a5bc |
TRACE(("|%p|%p|MIN_UNTIL %d %p\n", ctx->pattern,
|
|
|
f1a5bc |
ctx->ptr, ctx->count, ctx->u.rep->pattern));
|
|
|
f1a5bc |
|
|
|
f1a5bc |
- if (ctx->count < ctx->u.rep->pattern[1]) {
|
|
|
f1a5bc |
+ if (ctx->count < (Py_ssize_t) ctx->u.rep->pattern[1]) {
|
|
|
f1a5bc |
/* not enough matches */
|
|
|
f1a5bc |
ctx->u.rep->count = ctx->count;
|
|
|
f1a5bc |
DO_JUMP(JUMP_MIN_UNTIL_1, jump_min_until_1,
|
|
|
f1a5bc |
@@ -1302,7 +1302,7 @@ entrance:
|
|
|
f1a5bc |
|
|
|
f1a5bc |
LASTMARK_RESTORE();
|
|
|
f1a5bc |
|
|
|
f1a5bc |
- if ((ctx->count >= ctx->u.rep->pattern[2]
|
|
|
f1a5bc |
+ if ((ctx->count >= (Py_ssize_t) ctx->u.rep->pattern[2]
|
|
|
f1a5bc |
&& ctx->u.rep->pattern[2] != SRE_MAXREPEAT) ||
|
|
|
f1a5bc |
state->ptr == ctx->u.rep->last_ptr)
|
|
|
f1a5bc |
RETURN_FAILURE;
|
|
|
f1a5bc |
diff --git a/Modules/sre.h b/Modules/sre.h
|
|
|
f1a5bc |
--- a/Modules/sre.h
|
|
|
f1a5bc |
+++ b/Modules/sre.h
|
|
|
f1a5bc |
@@ -20,14 +20,14 @@
|
|
|
f1a5bc |
# if SIZEOF_SIZE_T > 4
|
|
|
f1a5bc |
# define SRE_MAXREPEAT (~(SRE_CODE)0)
|
|
|
f1a5bc |
# else
|
|
|
f1a5bc |
-# define SRE_MAXREPEAT ((SRE_CODE)PY_SSIZE_T_MAX + 1u)
|
|
|
f1a5bc |
+# define SRE_MAXREPEAT ((SRE_CODE)PY_SSIZE_T_MAX)
|
|
|
f1a5bc |
# endif
|
|
|
f1a5bc |
#else
|
|
|
f1a5bc |
# define SRE_CODE unsigned int
|
|
|
f1a5bc |
# if SIZEOF_SIZE_T > SIZEOF_INT
|
|
|
f1a5bc |
# define SRE_MAXREPEAT (~(SRE_CODE)0)
|
|
|
f1a5bc |
# else
|
|
|
f1a5bc |
-# define SRE_MAXREPEAT ((SRE_CODE)PY_SSIZE_T_MAX + 1u)
|
|
|
f1a5bc |
+# define SRE_MAXREPEAT ((SRE_CODE)PY_SSIZE_T_MAX)
|
|
|
f1a5bc |
# endif
|
|
|
f1a5bc |
#endif
|
|
|
f1a5bc |
|
|
|
f1a5bc |
|