Skip to content

[3.12] gh-124785: Revert "gh-116510: Fix crash due to shared immortal interned strings (gh-124646)" (gh-124807) #124818

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

Closed
wants to merge 1 commit into from
Closed

[3.12] gh-124785: Revert "gh-116510: Fix crash due to shared immortal interned strings (gh-124646)" (gh-124807) #124818

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
5 changes: 0 additions & 5 deletions Misc/NEWS.d/next/Core_and_Builtins/2024-09-25-12-37-58.gh-issue-116510.WeBAx1.rst
View file
Open in desktop

This file was deleted.

48 changes: 6 additions & 42 deletions Objects/unicodeobject.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -286,37 +286,13 @@ hashtable_unicode_compare(const void *key1, const void *key2)
}
}

/* Return true if this interpreter should share the main interpreter's
intern_dict. That's important for interpreters which load basic
single-phase init extension modules (m_size == -1). There could be interned
immortal strings that are shared between interpreters, due to the
PyDict_Update(mdict, m_copy) call in import_find_extension().

It's not safe to deallocate those strings until all interpreters that
potentially use them are freed. By storing them in the main interpreter, we
ensure they get freed after all other interpreters are freed.
*/
static bool
has_shared_intern_dict(PyInterpreterState *interp)
{
PyInterpreterState *main_interp = _PyInterpreterState_Main();
return interp != main_interp && interp->feature_flags & Py_RTFLAGS_USE_MAIN_OBMALLOC;
}

static int
init_interned_dict(PyInterpreterState *interp)
{
assert(get_interned_dict(interp) == NULL);
PyObject *interned;
if (has_shared_intern_dict(interp)) {
interned = get_interned_dict(_PyInterpreterState_Main());
Py_INCREF(interned);
}
else {
interned = PyDict_New();
if (interned == NULL) {
return -1;
}
PyObject *interned = interned = PyDict_New();
Copy link
Contributor

Choose a reason for hiding this comment

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

Why do you want to do this? Rather than

PyObject *interned = PyDict_New();

Copy link
Member

Choose a reason for hiding this comment

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

It's just a straight reversion of my change so that matches the previous code. I agree with you though, looks odd to write it that way.

Copy link
Member Author

Choose a reason for hiding this comment

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

It's a revert. Even worse, it's a backport of a revert. That's what the original code was. I imagine it was a mistake in the original PR (#102339) but that, too, may have been moved from somewhere else.

If you want to send a PR to fix it (in main) I'm sure Eric wouldn't mind reviewing :)

if (interned == NULL) {
return -1;
}
_Py_INTERP_CACHED_OBJECT(interp, interned_strings) = interned;
return 0;
Expand All @@ -327,10 +303,7 @@ clear_interned_dict(PyInterpreterState *interp)
{
PyObject *interned = get_interned_dict(interp);
if (interned != NULL) {
if (!has_shared_intern_dict(interp)) {
// only clear if the dict belongs to this interpreter
PyDict_Clear(interned);
}
PyDict_Clear(interned);
Py_DECREF(interned);
_Py_INTERP_CACHED_OBJECT(interp, interned_strings) = NULL;
}
Expand Down Expand Up @@ -15163,13 +15136,6 @@ _PyUnicode_ClearInterned(PyInterpreterState *interp)
}
assert(PyDict_CheckExact(interned));

if (has_shared_intern_dict(interp)) {
// the dict doesn't belong to this interpreter, skip the debug
// checks on it and just clear the pointer to it
clear_interned_dict(interp);
return;
}

#ifdef INTERNED_STATS
fprintf(stderr, "releasing %zd interned strings\n",
PyDict_GET_SIZE(interned));
Expand Down Expand Up @@ -15688,10 +15654,8 @@ _PyUnicode_Fini(PyInterpreterState *interp)
{
struct _Py_unicode_state *state = &interp->unicode;

if (!has_shared_intern_dict(interp)) {
// _PyUnicode_ClearInterned() must be called before _PyUnicode_Fini()
assert(get_interned_dict(interp) == NULL);
}
// _PyUnicode_ClearInterned() must be called before _PyUnicode_Fini()
assert(get_interned_dict(interp) == NULL);

_PyUnicode_FiniEncodings(&state->fs_codec);

Expand Down
Loading