Skip to content

gh-129726: Break gzip.GzipFile reference loop #130055

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

Merged
vstinner merged 3 commits into from
Feb 28, 2025
Merged

gh-129726: Break gzip.GzipFile reference loop #130055

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
2d67d94
gh-129726: Break `gzip.GzipFile` reference loop
cmaloney Feb 12, 2025
194cf3f
PEP-8 import order, move to weakref.ref
cmaloney Feb 28, 2025
32f3f87
reduce diff
cmaloney Feb 28, 2025
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
Next Next commit
gh-129726: Break gzip.GzipFile reference loop 
A reference loop was resulting in the `fileobj` held by the `GzipFile`
being closed before the `GzipFile`.

The issue started with gh-89550 in 3.12, but was hidden in most cases
until 3.13 when gh-62948 made it more visible.
  • Loading branch information
@cmaloney
cmaloney committed Feb 12, 2025
commit 2d67d946a5842360b85992893f0214ceaeca4a07
4 changes: 3 additions & 1 deletion Lib/gzip.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -10,6 +10,7 @@
import builtins
import io
import _compression
import weakref

@cmaloney cmaloney Feb 12, 2025

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

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

the ordering of imports here is painful to me, but resisted bringing to PEP8 because this fix needs backporting.

@vstinner vstinner Feb 27, 2025

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

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

Please reformat these imports for PEP 8 :-) I will handle merge conflicts on backports if needed.


__all__ = ["BadGzipFile", "GzipFile", "open", "compress", "decompress"]

Expand Down Expand Up @@ -226,7 +227,8 @@ def __init__(self, filename=None, mode=None,
0)
self._write_mtime = mtime
self._buffer_size = _WRITE_BUFFER_SIZE
self._buffer = io.BufferedWriter(_WriteBufferStream(self),
write_wrap = _WriteBufferStream(weakref.proxy(self))

@vstinner vstinner Feb 27, 2025

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

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

Would it be possible to modify _WriteBufferStream to use a simpler weakref.ref() to store gzip_file instead? This object is private so we can change its API.

Something like:

class _WriteBufferStream(io.RawIOBase):
    """Minimal object to pass WriteBuffer flushes into GzipFile"""
    def __init__(self, gzip_file):
        self.gzip_file = weakref.ref(gzip_file)

    def write(self, data):
        gzip_file = self.gzip_file()
        if gzip_file is None:
            raise RuntimeError("lost gzip_file")
        return gzip_file._write_raw(data)

@cmaloney cmaloney Feb 28, 2025

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

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

Updated to your suggestion here. I Spent a bit of time trying to remove _WriteBufferStream and just use weakref.proxy, but getting so GzipFile.write has a different write function from the underlying self._buffer: BufferedWriter (which needs to use self._write_raw) led me in a lot of circles. More understandable for things I tried to just keep _WriteBufferStream.

self._buffer = io.BufferedWriter(write_wrap,
buffer_size=self._buffer_size)
else:
raise ValueError("Invalid mode: {!r}".format(mode))
Expand Down
13 changes: 13 additions & 0 deletions Lib/test/test_gzip.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3,12 +3,14 @@

import array
import functools
import gc
import io
import os
import struct
import sys
import unittest
from subprocess import PIPE, Popen
from test.support import catch_unraisable_exception
from test.support import import_helper
from test.support import os_helper
from test.support import _4G, bigmemtest, requires_subprocess
Expand Down Expand Up @@ -859,6 +861,17 @@ def test_write_seek_write(self):
self.assertEqual(gzip.decompress(data), message * 2)


def test_refloop_unraisable(self):
# Ensure a GzipFile referring to a temporary fileobj deletes cleanly.
# Previously an unraisable exception would occur on close because the
# fileobj would be closed before the GzipFile as the result of a
# reference loop. See issue gh-129726
with catch_unraisable_exception() as cm:
gzip.GzipFile(fileobj=io.BytesIO(), mode="w")
gc.collect()
self.assertIsNone(cm.unraisable)


class TestOpen(BaseTest):
def test_binary_modes(self):
uncompressed = data1 * 50
Expand Down
3 changes: 3 additions & 0 deletions Misc/NEWS.d/next/Library/2025-02-12-12-38-24.gh-issue-129726.jB0sxu.rst
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,3 @@
Fix :class:`gzip.GzipFile` raising an unraisable exception during garbage
collection when referring to a temporary object by breaking the reference
loop with :mod:`weakref`.