Switches only work until the first usb cable disconnect about acc HOT 3 CLOSED

Toggling multiple switches may trigger the issue.
Did you try using a single switch under such conditions? -- acc -s s

from acc.

I chose battery/charging_enabled 1 0 as my switch. And the switch test acc -t now reports that my device is supported. And acc now stops charging at my cutoff battery charge, which is 90% for me. So technically my issue is resolved.

But there are still some weird things:

• If i do the switch test with a battery charge over my cutoff battery charge, like 92%, the test fails again. And this sticks until a reboot. But charging still ends, if the charge is 90% or higher. So everything works, but the switch test says otherwise.
• What bugs me the most, is that if the charge is greater than 90%, charging still stops and the charge drops over time below 90%, but the little thunderbolt in the battery icon does not go away, the indicator led on my phone does not turn off (as it does if the charge hits 90% from below and cuts off charging and stays there). So something happens, but something else still goes wrong. And the discharge wont stop, even if the battery charge drops below the resume charge, which is 85% for me. Disconnecting and reconnecting the usb cable does not help, it wont start to charge the battery until a reboot is done. After the reboot everything works fine again.

So everything is fine, if the battery charge approaches the cutoff charge from below and does not go over 90%. But it glitches out between 90% and 100% (yes even at 90%, disconnecting and reconnecting the usb cable at 90% will glitch out acc). Maybe this is already known. But if you are interested in this glitch, here my new logs:
acc-logs-h1.tar.bz2.zip

And if you wish, i could commit some additions to the readme which recommends a reboot, if the switch tests fails. And suggests that choosing a switch might resolve the problem.

from acc.

Observations and Questions

1. battery/charging_enabled is working.

2. acc -t is working with that control file.

3. acc -t -- misbehaves, because it always tests all control files. Some of which were found to be problematic.

4. Charging indicators won't go off after disabling charging.

5. A reboot is required to re-enable charging (rebootOnPause=timeoutSeconds).

6. The log shows exit code 143, which suggests something is killing accd after it tries disabling charging a couple of times.

7. Assuming you have a specific charging switch set, is acc -t prone to the inconsistent behavior of acc -t --? Recall that the first toggles only the set switch, while the second always toggles all.

8. If you don't perform any test with acc -t nor acc -t --, and the control file is still battery/charging_enabled, does accd work as expected?

P.S., there's a new dev build available. It's ready for daily use. If you have a Telegram account, we can work through this issue much faster.

from acc.