@cathoderay to be honest I've not look at it in a while. Anything you suggest I look for?

@batalto - if you can, see if you can see any record of a recent update? I can't, but my 'smart' phone isn't very smart.

@cathoderay this is what I have

@batalto - that is interesting, I don't appear to have such a screen. But that does make it appear that the latest update was over a couple of weeks ago, so it seems unlikely that a covert upgrade broke my midea_ac_lan connection.

This morning, even after rebooting the mini PC, I can't connect to it, and it doesn't appear on the LAN. I'm going to have to get out the monitor again and see what is going on. All very tedious and boring...  

@cathoderay

The 30 second refresh time delay comes from the 2010 software, a link to which you posted on page 2 of the present thread. This is the version that I have been using to investigate the software. It is now obvious that this may have differences from the version that you are using.

@derek-m - I don't think they should be too different, if at all. Only a very small number of the individual files in the full library get changed, I think only wasilukm versions of core/device.py c3/device.py and c3/message.py have had changes, and I sent you my working copies not so long ago (the zipped ones you couldn't open initially). I'm pretty sure we are working on the same code.

I think the 2010 is just part of the developer's name on github, all the files are much more recent, months not years old.

I've managed to get HA and midea_ac_lan back again, albeit with missing data for the time it was FUBAR. I think it is probably a dynamic IP address problem, I had to redo all the connections because IP addresses had changed. I seem to recall this is yet another HAPITA.  

@cathoderay

Working on the theory that the files are virtually the same, I think that rather than adding further code at the moment, it would be best to confirm what action the present code is performing. I suggest that you try commenting certain lines of code in turn, to see  what happens to the message produced.

Start by commenting the following line in C3/device.py

class DeviceAttributes(StrEnum):

tank_actual_temperature = "tank_actual_temperature"

@cathoderay

Once the previous test has been completed, then remove the comment and comment the following line as shown below.

class MideaC3Device(MiedaDevice):

# DeviceAttributes.tank_actual_temperature: None,

@cathoderay

Continue the testing procedure in C3/message.py, as shown below.

class C3MessageBody(MessageBody):

# self.tank_actual_temperature = body[data_offset + 21]

Posted by: @derek-m

↑

Start by commenting the following line in C3/device.py

class DeviceAttributes(StrEnum):

tank_actual_temperature = "tank_actual_temperature"

Unfortunately, it breaks the code, and midea_ac_lan fails to start:

AttributeError: tank_actual_temperature
2023-03-07 08:39:58.571 ERROR (MainThread) [homeassistant.setup] Setup failed for custom integration midea_ac_lan: Unable to import component: Exception importing custom_components.midea_ac_lan

I've noticed this before, python is very fastidious, even the smallest exception breaks it.

Posted by: @derek-m

↑

@cathoderay

Once the previous test has been completed, then remove the comment and comment the following line as shown below.

class MideaC3Device(MiedaDevice):

# DeviceAttributes.tank_actual_temperature: None,

 

Okay, try changing the value to something like 100, to see if that value appears in the message or whether it is overwritten by the true value.

Posted by: @derek-m

↑

try changing the value to something like 100, to see if that value appears in the message or whether it is overwritten by the true value

It gets overwritten by the true value, 'tank_actual_temperature': 38 in the log 'Status update' and 38 degrees C shows on HA dashboard. That seems plausible, the None is perhaps a default value, in the event nothing is returned.