gh-103082: Fix shifted field initialization in instrumentation.c

[arhadthedev]

If Py_TRACE_REFS is defined, PyObject initializers without field names like:

cpython/Python/instrumentation.c

Lines 17 to 21 in 2b6f5c3

	static PyObject DISABLE =
	{
	_PyObject_IMMORTAL_REFCNT,
	&PyBaseObject_Type
	};

are compiled in a totally unusable way:

_ob_next = _PyObject_IMMORTAL_REFCNT (999999999)
_ob_prev = &PyBaseObject_Type
ob_refcnt = 0
ob_type = NULL

The reason is implicit insertion of two extra fields under #ifdef Py_TRACE_REFS:

cpython/Include/object.h

Lines 102 to 106 in 2b6f5c3

	struct _object {
	_PyObject_HEAD_EXTRA
	Py_ssize_t ob_refcnt;
	PyTypeObject *ob_type;
	};

cpython/Include/object.h

Lines 65 to 76 in 2b6f5c3

	#ifdef Py_TRACE_REFS
	/* Define pointers to support a doubly-linked list of all live heap objects. */
	#define _PyObject_HEAD_EXTRA \
	PyObject *_ob_next; \
	PyObject *_ob_prev;
	#define _PyObject_EXTRA_INIT _Py_NULL, _Py_NULL,
	#else
	# define _PyObject_HEAD_EXTRA
	# define _PyObject_EXTRA_INIT
	#endif

This commit fixes declarations that fail the AMD64 Arch Linux TraceRefs PR buildbot. Other declarations will be fixed in another, pending PR. (edit: all other places use _PyObject_HEAD_EXTRA)

I desided to not use _PyObject_HEAD_EXTRA because it can be forgotten easily (unlike explicit field names) so we should phase it out from CPython codebase if possible.

• Issue: Implement and document PEP 669. #103082

[bedevere-bot]

🤖 New build scheduled with the buildbot fleet by @arhadthedev for commit 286fdc9 🤖

If you want to schedule another build, you need to add the 🔨 test-with-buildbots label again.

[arhadthedev]

@markshannon as an author of gh-103083 this PR fixes.

[orsenthil]

Explicit setting of these fields seems a bit odd to me.
@markshannon should review and merge it since this is recently introduced file.

[markshannon]

I think the use of named initializers is fine.

Since we can always change it to use macros later, and this fixes compilation, I'm merging this.