gh-153852: Fix data race in list.sort()

[weixlu]

As reported in #153852(TSAN0014), list.sort() writes each slot of the array with a plain, non-atomic store. When another thread concurrently accesses the same list , TSAN will report a data race between the in-place sort and the reader.

The fix is straightforward: publish each slot write with FT_ATOMIC_STORE_PTR_RELEASE, just like the other list mutators already do. For example, ins1() (used by list.insert) also uses FT_ATOMIC_STORE_PTR_RELEASE(items[i+1], items[i])

Testing

• The reproducer now runs clean.
• PASS ./python -m test test_sort test_list

Notes for reviewer:

• On the default build (GIL-enabled), FT_ATOMIC_STORE_PTR_RELEASE expands to a plain store, so there is no change there.
• This PR only adds a atomic store, so basic functional behavior is unchanged.

• Issue: Free-threading data races in CPython: 15 findings (9 new + 6 residual of existing FT work) #153852

[brijkapadia]

1. Could you create a new issue instead of referencing the parent?
2. Are you sure using a lot of release atomics is the best approach?