J. Nick Koston

8.1.3 Requires-Python >=3.10

https://github.com/ijl/orjson/commit/f1cbc0a Likely regressed in that commit Maybe some type of buffer alignment issue.

> @bdraco would you mind checking this out. The library already supports passing in an existing event loop so if I understand what you're saying the better fix would be...

There isn't a software workaround for this as we need to be able to match the device somehow, and the only stable identifier we have is the MAC address. If...

Did you try the factory reset?

It looks like its not fixable as UI isn't sending the person detection until the end of the event now. Event starts: notice that `smart_detect_types` is empty `2024-09-02 19:59:02.808 DEBUG...

It looks like this is a fundamental change in protect 4.1.x which means we likely can't do anything about it. This is either a regression in protect 4.1.x or an...

> Hi @TheJulianJES, please do not rename this issue, your renaming is not reflecting my problem. Thank you. Please check your logbook. Mine is still turning off and on but...

Sadly it looks like there are multiple problems as I can definitely still see person events, but they are microseconds long now. Can you post the debug for the events...

The select distinct optimizes poorly with PostgreSQL because it doesn't support loose/skip index scans https://wiki.postgresql.org/wiki/Loose_indexscan