[3.13] gh-105936: Properly update closure cells for __setattr__ and __delattr__ in frozen dataclasses with slots (GH-144021)

[gpshead]

gh-105936: Properly update closure cells for __setattr__ and __delattr__ in frozen dataclasses with slots (GH-144021) (cherry picked from commit 8a398bf)

The cherry-pick required additional changes beyond the original commit because 3.13 lacks the __class__ closure cell fixup machinery that was added in 3.14 by GH-124455 (gh-90562). Specifically:

Co-authored-by: Sviataslau 35541026+Prometheus3375@users.noreply.github.com

• Backported _update_func_cell_for__class__() helper function and the closure fixup loop in _add_slots() from gh-90562: Support zero argument super with dataclasses when slots=True #124455. Without these, renaming the closure variable from cls to __class__ has no effect because nothing updates the cell when the class is recreated with slots.
• Changed _add_slots() to use newcls instead of reusing cls for the recreated class, so both old and new class references are available for the fixup loop.
• Replaced assertNotHasAttr with assertFalse(hasattr(...)) in tests (assertNotHasAttr was added in 3.14).
• Dropped test_original_class_is_gced additions (that test does not exist on 3.13; it was added by gh-135228: When @dataclass(slots=True) replaces a dataclass, make the original class collectible (take 2) #137047 for 3.14 regression: slot dataclasses classes leak original class #135228 which was not backported to 3.13).

[gpshead]

The is left as a Draft because it required more changes and backporting pieces of some other stuff that is already in 3.14 in order for it to be possible on 3.13. Closer review required.

[JelleZijlstra]

This makes sense but I think we need a small fixup, I'll send a PR for that on master first.