Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

bpo-23689: re module, fix memory leak when a match is terminated by a signal or memory allocation failure #32283

Merged
merged 11 commits into from Apr 3, 2022
Merged

bpo-23689: re module, fix memory leak when a match is terminated by a signal or memory allocation failure #32283

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
11 commits
Select commit Hold shift + click to select a range
e5317fe
1. add variables
wjssz Apr 3, 2022
28e4d2d
2. add _CompileData class to _compile.py
wjssz Apr 3, 2022
7493fad
3. add repeat_count parameter to _sre.compile() function
wjssz Apr 3, 2022
b928335
4. emit repeat_count in sre_compile.py
wjssz Apr 3, 2022
dfec05d
5. change _validate_outer() parameter
wjssz Apr 3, 2022
db7e88b
6. validate in _validate_inner() / _validate_outer()
wjssz Apr 3, 2022
a43cb6e
7. allocate repeats_array for SRE_STATE
wjssz Apr 3, 2022
1560a0c
8. support code in sre_lib.h
wjssz Apr 3, 2022
96be025
9. add unit-tests
wjssz Apr 3, 2022
7ee3e66
a. skip -= field_number
wjssz Apr 3, 2022
e95b19f
b. improve unit-test
wjssz Apr 3, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
61 changes: 39 additions & 22 deletions Lib/re/_compiler.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -67,14 +67,21 @@
_ignorecase_fixes = {i: tuple(j for j in t if i != j)
for t in _equivalences for i in t}

class _CompileData:
__slots__ = ('code', 'repeat_count')
def __init__(self):
self.code = []
self.repeat_count = 0

def _combine_flags(flags, add_flags, del_flags,
TYPE_FLAGS=_parser.TYPE_FLAGS):
if add_flags & TYPE_FLAGS:
flags &= ~TYPE_FLAGS
return (flags | add_flags) & ~del_flags

def _compile(code, pattern, flags):
def _compile(data, pattern, flags):
# internal: compile a (sub)pattern
code = data.code
emit = code.append
_len = len
LITERAL_CODES = _LITERAL_CODES
Expand Down Expand Up @@ -147,15 +154,19 @@ def _compile(code, pattern, flags):
skip = _len(code); emit(0)
emit(av[0])
emit(av[1])
_compile(code, av[2], flags)
_compile(data, av[2], flags)
emit(SUCCESS)
code[skip] = _len(code) - skip
else:
emit(REPEATING_CODES[op][0])
skip = _len(code); emit(0)
emit(av[0])
emit(av[1])
_compile(code, av[2], flags)
# now op is in (MIN_REPEAT, MAX_REPEAT, POSSESSIVE_REPEAT)
if op != POSSESSIVE_REPEAT:
emit(data.repeat_count)
data.repeat_count += 1
_compile(data, av[2], flags)
code[skip] = _len(code) - skip
emit(REPEATING_CODES[op][1])
elif op is SUBPATTERN:
Expand All @@ -164,7 +175,7 @@ def _compile(code, pattern, flags):
emit(MARK)
emit((group-1)*2)
# _compile_info(code, p, _combine_flags(flags, add_flags, del_flags))
_compile(code, p, _combine_flags(flags, add_flags, del_flags))
_compile(data, p, _combine_flags(flags, add_flags, del_flags))
if group:
emit(MARK)
emit((group-1)*2+1)
Expand All @@ -176,7 +187,7 @@ def _compile(code, pattern, flags):
# pop their stack if they reach it
emit(ATOMIC_GROUP)
skip = _len(code); emit(0)
_compile(code, av, flags)
_compile(data, av, flags)
emit(SUCCESS)
code[skip] = _len(code) - skip
elif op in SUCCESS_CODES:
Expand All @@ -191,13 +202,13 @@ def _compile(code, pattern, flags):
if lo != hi:
raise error("look-behind requires fixed-width pattern")
emit(lo) # look behind
_compile(code, av[1], flags)
_compile(data, av[1], flags)
emit(SUCCESS)
code[skip] = _len(code) - skip
elif op is CALL:
emit(op)
skip = _len(code); emit(0)
_compile(code, av, flags)
_compile(data, av, flags)
emit(SUCCESS)
code[skip] = _len(code) - skip
elif op is AT:
Expand All @@ -216,7 +227,7 @@ def _compile(code, pattern, flags):
for av in av[1]:
skip = _len(code); emit(0)
# _compile_info(code, av, flags)
_compile(code, av, flags)
_compile(data, av, flags)
emit(JUMP)
tailappend(_len(code)); emit(0)
code[skip] = _len(code) - skip
Expand Down Expand Up @@ -244,12 +255,12 @@ def _compile(code, pattern, flags):
emit(op)
emit(av[0]-1)
skipyes = _len(code); emit(0)
_compile(code, av[1], flags)
_compile(data, av[1], flags)
if av[2]:
emit(JUMP)
skipno = _len(code); emit(0)
code[skipyes] = _len(code) - skipyes + 1
_compile(code, av[2], flags)
_compile(data, av[2], flags)
code[skipno] = _len(code) - skipno
else:
code[skipyes] = _len(code) - skipyes + 1
Expand Down Expand Up @@ -608,17 +619,17 @@ def isstring(obj):
def _code(p, flags):

flags = p.state.flags | flags
code = []
data = _CompileData()

# compile info block
_compile_info(code, p, flags)
_compile_info(data.code, p, flags)

# compile the pattern
_compile(code, p.data, flags)
_compile(data, p.data, flags)

code.append(SUCCESS)
data.code.append(SUCCESS)

return code
return data

def _hex_code(code):
return '[%s]' % ', '.join('%#0*x' % (_sre.CODESIZE*2+2, x) for x in code)
Expand Down Expand Up @@ -719,14 +730,21 @@ def print_2(*args):
else:
print_(FAILURE)
i += 1
elif op in (REPEAT, REPEAT_ONE, MIN_REPEAT_ONE,
elif op in (REPEAT_ONE, MIN_REPEAT_ONE,
POSSESSIVE_REPEAT, POSSESSIVE_REPEAT_ONE):
skip, min, max = code[i: i+3]
if max == MAXREPEAT:
max = 'MAXREPEAT'
print_(op, skip, min, max, to=i+skip)
dis_(i+3, i+skip)
i += skip
elif op is REPEAT:
skip, min, max, repeat_index = code[i: i+4]
if max == MAXREPEAT:
max = 'MAXREPEAT'
print_(op, skip, min, max, repeat_index, to=i+skip)
dis_(i+4, i+skip)
i += skip
elif op is GROUPREF_EXISTS:
arg, skip = code[i: i+2]
print_(op, arg, skip, to=i+skip)
Expand Down Expand Up @@ -781,11 +799,11 @@ def compile(p, flags=0):
else:
pattern = None

code = _code(p, flags)
data = _code(p, flags)

if flags & SRE_FLAG_DEBUG:
print()
dis(code)
dis(data.code)

# map in either direction
groupindex = p.state.groupdict
Expand All @@ -794,7 +812,6 @@ def compile(p, flags=0):
indexgroup[i] = k

return _sre.compile(
pattern, flags | p.state.flags, code,
p.state.groups-1,
groupindex, tuple(indexgroup)
)
pattern, flags | p.state.flags, data.code,
p.state.groups-1, groupindex, tuple(indexgroup),
data.repeat_count)
2 changes: 1 addition & 1 deletion Lib/re/_constants.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -13,7 +13,7 @@

# update when constants are added or removed

MAGIC = 20220318
MAGIC = 20220402

from _sre import MAXREPEAT, MAXGROUPS

Expand Down
28 changes: 26 additions & 2 deletions Lib/test/test_re.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1643,9 +1643,12 @@ def test_dealloc(self):
long_overflow = 2**128
self.assertRaises(TypeError, re.finditer, "a", {})
with self.assertRaises(OverflowError):
_sre.compile("abc", 0, [long_overflow], 0, {}, ())
_sre.compile("abc", 0, [long_overflow], 0, {}, (), 0)
with self.assertRaises(TypeError):
_sre.compile({}, 0, [], 0, [], [])
_sre.compile({}, 0, [], 0, [], [], 0)
with self.assertRaises(RuntimeError):
# invalid repeat_count -1
_sre.compile("abc", 0, [1], 0, {}, (), -1)

def test_search_dot_unicode(self):
self.assertTrue(re.search("123.*-", '123abc-'))
Expand Down Expand Up @@ -2334,6 +2337,27 @@ def test_possesive_repeat(self):
14. SUCCESS
''')

def test_repeat_index(self):
self.assertEqual(get_debug_out(r'(?:ab)*?(?:cd)*'), '''\
MIN_REPEAT 0 MAXREPEAT
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

You just read my mind! I was going to propose such a change, but I thought that I was already bothering you too much.

Copy link
Author

@ghost ghost Apr 3, 2022 •
edited by ghost
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I just thought this after posting this PR.

I thought that I was already bothering you too much.

As an inactive contributor, this is not a matter.
I'm not practised, so need continuously improve the patch to get to a good state.
When I think it's good, I can always find its shortcomings afterwards.

I have only one question: how to prove that we need only one SRE_REPEAT structure per the REPEAT code?

I have to think about how to answer your question.

LITERAL 97
LITERAL 98
MAX_REPEAT 0 MAXREPEAT
LITERAL 99
LITERAL 100

0. INFO 4 0b0 0 MAXREPEAT (to 5)
5: REPEAT 8 0 MAXREPEAT 0 (to 14)
10. LITERAL 0x61 ('a')
12. LITERAL 0x62 ('b')
14: MIN_UNTIL
15. REPEAT 8 0 MAXREPEAT 1 (to 24)
20. LITERAL 0x63 ('c')
22. LITERAL 0x64 ('d')
24: MAX_UNTIL
25. SUCCESS
''')


class PatternReprTests(unittest.TestCase):
def check(self, pattern, expected):
Expand Down
2 changes: 2 additions & 0 deletions Misc/NEWS.d/next/Library/2022-04-03-13-19-08.bpo-23689.TFSc3E.rst
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,2 @@
:mod:`re` module: fix memory leak when a match is terminated by a signal or
memory allocation failure. Patch by Ma Lin.
Loading