python / python/cpython

FrameLocalsProxy `|=` and `update()` mishandle errors raised while merging mappings

Open
#153,418 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

interpreter-core type-bug
Dominant language
Python
Stars
77.2k
Forks
35.9k
PR merge metrics
PR metrics pending

Description

Bug report

Bug description:

Bug description

FrameLocalsProxy does not correctly propagate exceptions raised while merging a dict subclass or another supported mapping-like object.

There are two related problems in Objects/frameobject.c:

  1. framelocalsproxy_inplace_or() returns Py_NotImplemented when framelocalsproxy_merge() fails. If the merge failed because Python code raised an exception, the slot returns a normal object while an exception is still set. In a pydebug build this triggers a fatal _Py_CheckSlotResult error.
  2. framelocalsproxy_update() replaces any failure from framelocalsproxy_merge() with a generic TypeError, masking the original exception.

This is a sub-issue of https://github.com/python/cpython/issues/146102 with gist details
A minimal reproducer is below.

import operator
import sys

class BadDict(dict):
    def keys(self):
        raise RuntimeError("keys() failed!")

def repro_update():
    locs = sys._getframe().f_locals
    print(type(locs))

    try:
        locs.update(BadDict())
    except TypeError as exc:
        print("BUG: update() masked RuntimeError as TypeError:", exc)
    except RuntimeError as exc:
        print("OK: update() propagated RuntimeError:", exc)

def repro_inplace_or():
    locs = sys._getframe().f_locals
    print(type(locs))

    try:
        operator.ior(locs, BadDict())
    except RuntimeError as exc:
        print("OK: |= propagated RuntimeError:", exc)
    except BaseException as exc:
        print("BUG: |= raised wrong exception:", type(exc).__name__, exc)
    else:
        print("BUG: |= swallowed RuntimeError")

repro_update()
repro_inplace_or()

operator.ior(locs, BadDict()) is used to exercise the same nb_inplace_or slot as locs |= BadDict().

Actual behavior

For .update(), the original RuntimeError from BadDict.keys() is replaced with a generic TypeError:

<class 'FrameLocalsProxy'>
BUG: update() masked RuntimeError as TypeError: update() argument must be dict or another FrameLocalsProxy

For |= on a pydebug build, the interpreter aborts because the slot reports success while an exception remains set:

<class 'FrameLocalsProxy'>
Fatal Python error: _Py_CheckSlotResult: Slot |= of type FrameLocalsProxy succeeded with an exception set
Python runtime state: initialized
Traceback (most recent call last):
  File "/tmp/repro_ior.py", line 6, in keys
    raise RuntimeError("keys() failed!")
RuntimeError: keys() failed!
Aborted (core dumped)

Expected behavior

Both FrameLocalsProxy.update() and FrameLocalsProxy.__ior__() should propagate the original exception raised while merging:

<class 'FrameLocalsProxy'>
OK: update() propagated RuntimeError: keys() failed!
<class 'FrameLocalsProxy'>
OK: |= propagated RuntimeError: keys() failed!

I will submit a PR to fix this

CPython versions tested on:

CPython main branch

Operating systems tested on:

Linux

Linked PRs
  • gh-153421

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start in Objects/frameobject.c at framelocalsproxy_inplace_or() and framelocalsproxy_update(), then run the minimal reproducer with BadDict.keys() raising RuntimeError. Done means both operations propagate the original exception without a debug-build slot failure, with regression coverage for the reported behavior.

Written by the indexing model from the issue text.

Assessment

Tech stack
c, python
Domain
backend
Issue type
Bug
Difficulty
3/5
Estimated time
1-2 days
Activity status
Stale
Clarity
Clearly specified
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.