Move the unref to below the multi-interpreter check.

[b-pass]

Description

This is a fix for #6174.

AI's assessment seems plausible

Thread A observes has_seen_non_main_interpreter() == false in get_pp() and prepares to return internals_singleton_pp_.
A non-main-interpreter thread B enters ensure_internals(), still observes the flag as false in unref(), and clears internals_singleton_pp_.
Thread A returns the now-null shared member.

Publishing has_seen_non_main_interpreter() = true before calling unref() in a non-main interpreter would cause that unref() to clear only the thread-local cache instead of the shared singleton.

Moving the unref makes sense. We cannot outright remove the unref, there is a case (covered by the "Restart the interpreter" unit test) which crashes without it.

Suggested changelog entry:

• Fixed occasional test failure under heavy threading load (especially on windows)