René's URL Explorer Experiment

{"@context":"https://schema.org","@type":"DiscussionForumPosting","headline":"Single-threaded re-entrancy via a recursive generator causes an access violation (segfault) in `pairwise_next`","articleBody":"# Crash report\n\n### What happened?\n\nThe existing partial fix in `main` (the `it = po-\u003eit; if (it == NULL)` re-read guard after the first `tp_iternext` call) does not cover the second `tp_iternext` call, where `it` is still a borrowed reference. In the single-threaded case a re-entrant call can trigger `Py_CLEAR(po-\u003eit)` which drops the generator's refcount to zero and frees it, leaving the outer call's local `it` as a dangling pointer.\n\n**Minimal reproducer**:\n\n```python\nfrom itertools import pairwise\n\ndef g():\n    yield next(it)\n\nfor _ in range(10000):\n    try:\n        it = pairwise(g())\n        next(it)\n    except ValueError:\n        pass\n```\n\n```\nException ignored in: \u003cgenerator object g at 0x...\u003e\nValueError: generator already executing\nWindows fatal exception: access violation\n```\n\nThis is also the root cause of https://github.com/spyder-ide/qtconsole/issues/635.\n\nSome options to address:\n- `Py_INCREF(it)` before the second `tp_iternext(it)` call + `Py_DECREF` after, or\n- a re-entrancy guard flag (like `teedataobject.running`) as suggested by @rhettinger.\n\nNote: `Py_BEGIN_CRITICAL_SECTION` (added in PR #144489) does **not** protect against same-thread re-entrancy.\n\n\n\n### CPython versions tested on:\n\nCPython main branch\n\n### Operating systems tested on:\n\n_No response_\n\n### Output from running 'python -VV' on the command line:\n\n_No response_\n\n\u003c!-- gh-linked-prs --\u003e\n### Linked PRs\n* gh-150589\n\u003c!-- /gh-linked-prs --\u003e\n","author":{"url":"https://github.com/eendebakpt","@type":"Person","name":"eendebakpt"},"datePublished":"2026-05-08T12:25:25.000Z","interactionStatistic":{"@type":"InteractionCounter","interactionType":"https://schema.org/CommentAction","userInteractionCount":3},"url":"https://github.com/149557/cpython/issues/149557"}