uasyncio --- Cleaning up after a task

[doceave]

Hi there everyone

It is very likely that owing to my poor implementation of the uasyncio library, on an ESP32 WROOM32, that after a few minutes of running memory intensive tasks, bugs begin revealing themselves. All functions test well initially, but after a while seemingly random error is raised in the REPL:

Code: Select all

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "ACT_v0b.py", line 1154, in <module>
  File "uasyncio/core.py", line 109, in run_forever
  File "ACT_v0b.py", line 1070, in ACT_Now_R
  File "uasyncio/core.py", line 180, in run_until_complete
  File "uasyncio/core.py", line 92, in run_forever
IndexError: empty
>>>

My gut tells me that I am not cleaning up well after running certain functions repeatedly with resultant memory issues that develop in a use-time dependent manner...

Put simply, is there a proper way to clean up after functions that are called using "create_task", which return immediately but may be called multiple times from different places? Example:

Code: Select all

loop = asyncio.get_event_loop()
loop.run_until_complete(json_get_settings())
loop.run_until_complete(wifi_connect())
loop.create_task(home_screen())

Or anything else which could explain odd behavior as a function of time/usage?

Thanks.

[kevinkk525]

I have a few questions:

1. Are you using the new uasyncio that got merged within the last weeks (and is present in daily builds) or the old one? (If you have asyncio.run then it is the new one)
2. Why are you calling loop.run_until_complete multiple times? Typically you're program should have one main loop that is always running and then create tasks.
3. You clean up the RAM by calling gc.collect(). Although that will happen automatically eventually, when the available RAM gets too low, but RAM fragmentation can cause problems and at some point there might not be enough contiguous RAM available. Try getting some output from gc.mem_free() and micropython.mem_info(1) when it crashes.

[pythoncoder]

You can test for the new version with

Code: Select all

import uasyncio as asyncio
asyncio.__version__

The old version will throw an exception, the new one will issue (3, 0, 0).

@doceave As Kevin says you are using uasyncio in a strange way. I suggest you look at the tutorial in this repo.

[doceave]

I am indeed using the older version of uasyncio
> Upgrade pending

Following numerous reviews of the linked tutorial I now finally get it!
> I now have a main loop which starts tasks, each of which await coros as needed

As always... many many thanks for the gentle assistance.