bleak
bleak copied to clipboard
hbldh
OSError: [WinError -2147418113] when connecting to a same device for the second time
- bleak version:
0.13.0 - Python version:
3.9.7 - Operating System:
Windows 10 20H1 - BlueZ version (
bluetoothctl -v) in case of Linux:
Description
When connecting to a same device for the second time in a short time(10 min or more), error occured.
What I Did
- Connected to a device, transfered some data(or do nothing), then disconnected (python exited)
- Connected to the same device again, got error below:
Traceback (most recent call last):
File "E:\bmc1025\bmc\learnlua\bletest3.py", line 107, in <module>
sys.exit(asyncio.run(main()))
File "D:\Programmes\anaconda\envs\bleenv\lib\asyncio\runners.py", line 44, in run
return loop.run_until_complete(main)
File "D:\Programmes\anaconda\envs\bleenv\lib\asyncio\base_events.py", line 642, in run_until_complete
return future.result()
File "E:\bmc1025\bmc\learnlua\bletest3.py", line 67, in main
raise e
File "E:\bmc1025\bmc\learnlua\bletest3.py", line 64, in main
await client.connect()
File "D:\Programmes\anaconda\envs\bleenv\lib\site-packages\bleak\backends\winrt\client.py", line 257, in connect
await self.get_services()
File "D:\Programmes\anaconda\envs\bleenv\lib\site-packages\bleak\backends\winrt\client.py", line 426, in get_services
await self._requester.get_gatt_services_async(
OSError: [WinError -2147418113] 灾难性故障
灾难性故障: maybe 'Fatal error' in English
More
Restarting OS, or waiting for a long time solves. Maybe windows does some clean up.
After debugging for hours, I found if I click "Step over" slowly, the code is more likely to run as expected.
After trying sleep in many places, I found the key.
...\backends\winrt\client.py, about line 258
...
await asyncio.sleep(1) # +
# Obtain services, which also leads to connection being established.
await self.get_services()
Interesting. It would be nice if you could do a Bluetooth packet capture for both the working (sleep workaround) and non-working case so we can see what the difference is.
There is also #695 which adds the possibility to use cached services on the second connection which may help work around the issue.
I've had the same issue on one computer. And despite using this modification of the client.py script the OS error still remains :
...\backends\winrt\client.py, about line 258
... await asyncio.sleep(1) # + # Obtain services, which also leads to connection being established. await self.get_services()
Do you have any ideas on how to fix this ?
As mentioned by OP when I restart the computer it works fine sometimes. The issue is probably related to the winrt backend (I have the same logs as OP except in French for Fatal Error / Catastrophic Failure : Défaillance Irrémédiable), is https://github.com/hbldh/bleak/pull/695 this feature going to be integrated in bleak package as the issues seem connected.
