Fixes #22055

As discussed in the issue. This is not super-principled, but it works,
and also gives nice standard colorized blocker if the fallback doesn't
work either.

(cherry picked from commit 1648291)

Initial changelog for the latest mypy release

(cherry picked from commit c34b73f)

(cherry picked from commit 051f2c9)

Primarily add overviews to beginnings of various sections.

(cherry picked from commit 5c45130)

Add a cherry picked commit to the changelog and remove Unreleased along
with another styling nit

(cherry picked from commit 3ad769c)

Minor fix

(cherry picked from commit 05e50c0)

(cherry picked from commit 7382985)

Fixes #22090

(cherry picked from commit d176870)

The recursive setup could cause `AttributeError` at runtime when a
function contains both a frame-private as well as a captured nested
function. The setup would add the frame-private function object as an
attribute to the env class and access it even though it was never
initialized.

Fix by omitting the recursive setup since it's not needed - a
frame-private nested function can't call itself.

Also fix possibly inconsistent storage for a redefined nested function
that is captured in one branch and not captured in another by using
`original_def` in capture/free variable checks.

(cherry picked from commit a9928e3)

Fixes #22229

This fixes a regression that can (often) cause false negatives in
settable properties.

The problem is that we currently may have some part of an overloaded
function to be scheduled for processing as part of interface, and other
parts to be processed as parts of implementation. This is especially
problematic for properties (which are internally stored as overloads). I
propose to always keep overloads as a whole. It may be marginally slower
in parallel mode, but much simpler to reason about.

(cherry picked from commit 45992fc)

(cherry picked from commit 033f958)