-
-
Notifications
You must be signed in to change notification settings - Fork 29.8k
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.
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
CMultiplyNested closure crash: Python 3.13 regression #121863
Comments
Just checking - the code here crashes when run with Python and you don't need to compile with Cython to get the bug? |
The reproducer doesn't need Cython at all. Only these few lines of Python code without any import. |
The bug can be worked around by reverting codeobject.c changes: diff --git a/Objects/codeobject.c b/Objects/codeobject.c
index 7b1244a8d5f..55b512b3f1a 100644
--- a/Objects/codeobject.c
+++ b/Objects/codeobject.c
@@ -137,7 +137,6 @@ static PyObject *intern_one_constant(PyObject *op);
static int
intern_strings(PyObject *tuple)
{
- PyInterpreterState *interp = _PyInterpreterState_GET();
Py_ssize_t i;
for (i = PyTuple_GET_SIZE(tuple); --i >= 0; ) {
@@ -147,7 +146,7 @@ intern_strings(PyObject *tuple)
"non-string found in code slot");
return -1;
}
- _PyUnicode_InternMortal(interp, &_PyTuple_ITEMS(tuple)[i]);
+ PyUnicode_InternInPlace(&_PyTuple_ITEMS(tuple)[i]);
}
return 0;
}
@@ -158,13 +157,12 @@ intern_strings(PyObject *tuple)
static int
intern_constants(PyObject *tuple, int *modified)
{
- PyInterpreterState *interp = _PyInterpreterState_GET();
for (Py_ssize_t i = PyTuple_GET_SIZE(tuple); --i >= 0; ) {
PyObject *v = PyTuple_GET_ITEM(tuple, i);
if (PyUnicode_CheckExact(v)) {
if (should_intern_string(v)) {
PyObject *w = v;
- _PyUnicode_InternMortal(interp, &v);
+ PyUnicode_InternInPlace(&v);
if (w != v) {
PyTuple_SET_ITEM(tuple, i, v);
if (modified) {
@@ -460,13 +458,12 @@ init_code(PyCodeObject *co, struct _PyCodeConstructor *con)
con->stacksize = 1;
}
- PyInterpreterState *interp = _PyInterpreterState_GET();
co->co_filename = Py_NewRef(con->filename);
co->co_name = Py_NewRef(con->name);
co->co_qualname = Py_NewRef(con->qualname);
- _PyUnicode_InternMortal(interp, &co->co_filename);
- _PyUnicode_InternMortal(interp, &co->co_name);
- _PyUnicode_InternMortal(interp, &co->co_qualname);
+ PyUnicode_InternInPlace(&co->co_filename);
+ PyUnicode_InternInPlace(&co->co_name);
+ PyUnicode_InternInPlace(&co->co_qualname);
co->co_flags = con->flags;
co->co_firstlineno = con->firstlineno;
@@ -492,6 +489,7 @@ init_code(PyCodeObject *co, struct _PyCodeConstructor *con)
co->co_framesize = nlocalsplus + con->stacksize + FRAME_SPECIALS_SIZE;
co->co_ncellvars = ncellvars;
co->co_nfreevars = nfreevars;
+ PyInterpreterState *interp = _PyInterpreterState_GET();
#ifdef Py_GIL_DISABLED
PyMutex_Lock(&interp->func_state.mutex);
#endif |
Thanks for the bisect! |
|
I'll investigate that. I'd still like to merge the quick partial revert for 3.13.0b4 today. |
…thonGH-121903) (cherry picked from commit cffad5c) Co-authored-by: Petr Viktorin <encukou@gmail.com>
The regression is fixed, to the best of my knowledge. |
Test extracted from Cython
tests/run/methodmangling_T5.py
:Output:
Linked PRs
The text was updated successfully, but these errors were encountered: