If discovery is triggered without credentials and discovers devices requiring authentication,
blank credentials are used to initialise the protocols and no connection is actually made.
In this instance we should not return the credentials_hash for blank credentials as it will be invalid.
* Add ChildDevice and ChildProtocolWrapper
* Initialize & update children
* Fix circular imports
* Add dummy_protocol fixture and tests for unwrapping responseData
* Use dummy_protocol for existing smartprotocol tests
* Move _ChildProtocolWrapper to smartprotocol.py
* Use dummy_protocol for test multiple requests
* Use device_id instead of position for selecting the child
* Fix wrapping for regular requests
* Remove unused imports
* tweak
* rename child_device to childdevice
* Fix import
* Fix connection indeterminate state on cancellation
If the task the query is running in it cancelled, we do
know the state of the connection so we must close. Previously
we would not close on BaseException which could result
in reading the previous response if the previous query was
cancelled after the request had been sent
* add test for cancellation
* Do login entirely within AesTransport
* Remove login and handshake attributes from BaseTransport
* Add AesTransport tests
* Synchronise transport and protocol __init__ signatures and rename internal variables
* Update after review
* Drop python 3.7 support
* CI: drop 3.7 and add 3.11
* Remove skipifs that were required for <3.8
* Use pypy-3.8 for CI, re-enable pypy for windows to see if it works now
* Bump readthedocs to use py3.8
* Remove py3.7 failure comment
* Avoid retrying open_connection on unrecoverable errors
- We can retry so hard that we block the event loop
Fixes
```
2022-04-16 22:18:51 WARNING (MainThread) [asyncio] Executing <Task finished name=Task-3576 coro=<open_connection() done, defined at /opt/homebrew/Cellar/python@3.9/3.9.12/Frameworks/Python.framework/Versions/3.9/lib/python3.9/asyncio/streams.py:25> exception=ConnectionRefusedError(61, "Connect call failed (192.168.107.200, 9999)") created at /opt/homebrew/Cellar/python@3.9/3.9.12/Frameworks/Python.framework/Versions/3.9/lib/python3.9/asyncio/tasks.py:460> took 1.001 seconds
```
* comment
* Keep connection open and lock to prevent duplicate requests
* option to not update children
* tweaks
* typing
* tweaks
* run tests in the same event loop
* memorize model
* Update kasa/protocol.py
Co-authored-by: Teemu R. <tpr@iki.fi>
* Update kasa/protocol.py
Co-authored-by: Teemu R. <tpr@iki.fi>
* Update kasa/protocol.py
Co-authored-by: Teemu R. <tpr@iki.fi>
* Update kasa/protocol.py
Co-authored-by: Teemu R. <tpr@iki.fi>
* dry
* tweaks
* warn when the event loop gets switched out from under us
* raise on unable to connect multiple times
* fix patch target
* tweaks
* isrot
* reconnect test
* prune
* fix mocking
* fix mocking
* fix test under python 3.7
* fix test under python 3.7
* less patching
* isort
* use mocker to patch
* disable on old python since mocking doesnt work
* avoid disconnect/reconnect cycles
* isort
* Fix hue validation
* Fix latitude_i/longitude_i units
Co-authored-by: Teemu R. <tpr@iki.fi>
* Add doctests to SmartBulb
* Add SmartDevice doctests, cleanup README.md
* add doctests for smartplug and smartstrip
* add discover doctests
* Fix bulb mock
* add smartdimmer doctests
* add sphinx-generated docs, cleanup readme a bit
* remove sphinx-click as it does not work with asyncclick
* in preparation for rtd hooking, move doc deps to be separate from dev deps
* pytestmark needs to be applied separately for each and every file, this fixes the tests
* use pathlib for resolving relative paths
* Skip discovery doctest on python3.7
The code is just fine, but some reason the mocking behaves differently between 3.7 and 3.8.
The latter seems to accept a discrete object for asyncio.run where the former expects a coroutine..
