> Does this also deprecate "static_cast<char32_t>(u8' ')"?

It's not intended to, but looking at
https://eel.is/c++draft/expr.static.cast#5, the proposed wording
inadvertently does.

> In the change of [conv.integral] paragraph 1 - why is the second
> bullet needed? (btw, missing "is"?)

Yes, there is a typo indeed.
The second bullet is needed because in a char8_t to char32_t
conversion sequences, char8_t is first promoted to int (typically).
This would bypass the first bullet because we don't want to deprecate
int to char32_t conversion.