Fix start_tls() loses buffer data and hangs when upgrading the connection

[fizzi01]

This PR fixes a critical issue where buffered data in StreamReader is lost when upgrading a connection to TLS using StreamWriter.start_tls() or loop.start_tls(), causing the call to hang indefinitely. The issue exists in both asyncio (see python/cpython#142352) and uvloop.

The main focus is to verify and guarantee that any data buffered before upgrading a connection to TLS is not lost and is correctly transferred. These changes are effectively a port of python/cpython#142354 with absolutely no difference, except for the test that creates a TCP server, establishes a connection, buffers data, performs a TLS upgrade, and verifies that all messages (both pre-upgrade and post-upgrade) are received correctly.

Fixes#707

[fizzi01]

The current fix breaks the Test:test_start_tls because it uses protocol classes without the _stream_reader attribute.
A workaround could be to fallback to the old behavior when that attribute is missing:

asyncdefstart_tls(self, transport, protocol, ...):
# Transfer buffered data from the old protocol to the new one.stream_buff=Noneifhasattr(protocol, '_stream_reader'):
stream_reader=protocol._stream_readerifstream_readerisnotNoneandhasattr(stream_reader, '_buffer'):
stream_buff=stream_reader._bufferifstream_buffisnotNone:
ssl_protocol._incoming.write(stream_buff)
stream_buff.clear()

or to modify the base protocol class implementation in `test_context.py:

class_BaseProtocol(asyncio.BaseProtocol):
def__init__(self, cvar, *, loop=None):
...
self._stream_reader: asyncio.StreamReader=asyncio.StreamReader(loop=loop)

Do we always have a protocol with a _stream_reader attribute?

[fantix]

Thanks for the PR!

Do we always have a protocol with a _stream_reader attribute?

No I don't think so. Though, I don't understand the test failure - looks like you already have if not hasattr(protocol, '_stream_reader'): return?

[fizzi01]

Though, I don't understand the test failure - looks like you already have if not hasattr(protocol, '_stream_reader'): return?

The function start_tls() should either return a transport or raise an error. Since we don't consistently handle a buffer, we can just skip the new buffer logic and run as before in that cases.

transport=awaitself.loop.start_tls(
proto.transport, proto, client_sslctx,
server_hostname='127.0.0.1',
)

[fizzi01]

@fantix Everything should be fine now, can you review please?