Introduce ReadableBuffer and WriteableBuffer Union aliases (#4232)

Since typing doesn't yet have a way to express buffer protocol objects
(python/typing#593), various interfaces have ended up with a mish-mash
of options: some list just bytes (or just bytearray, when writable),
some include mmap, some include memoryview, I think none of them include
array.array even though it's explicitly mentioned as bytes-like, etc. I
ran into problems because RawIOBase.readinto didn't allow for
memoryview.

To allow for some uniformity until the fundamental issue is resolved,
I've introduced _typeshed.ReadableBuffer and _typeshed.WriteableBuffer,
and applied them in stdlib/3/io.pyi as an example. If these get rolled
out in more places, it will mean that we have only one place where they
have to get tweaked in future, or swapped out for a public protocol.

This unfortunately does have the potential to break code that inherits
from RawIOBase/BufferedIOBase and overrides these methods, because the
base method is now more general and so the override now needs to accept
these types as well (which is why I've also updated gzip and lzma).
However, it should be a reasonably easy fix, and will make the
downstream annotations more correct.

stdlib/3/gzip.pyi

@@ -2,7 +2,7 @@ import sys
 import zlib
 from typing import IO, Optional, TextIO, Union, overload
 import _compression
-from _typeshed import AnyPath
+from _typeshed import AnyPath, ReadableBuffer
 from typing_extensions import Literal
 
 _OpenBinaryMode = Literal["r", "rb", "a", "ab", "w", "wb", "x", "xb"]
@@ -63,7 +63,7 @@ class GzipFile(_compression.BaseStream):
     @property
     def mtime(self) -> Optional[int]: ...
     crc: int
-    def write(self, data: bytes) -> int: ...
+    def write(self, data: ReadableBuffer) -> int: ...
     def read(self, size: Optional[int] = ...) -> bytes: ...
     def read1(self, size: int = ...) -> bytes: ...
     def peek(self, n: int) -> bytes: ...