Added variant to fineoffset_wh1050.c, which contains time data every 2h

[blackholefinder]

The "fine offset wh1050" decoder works in principle also with station TFA STRATOS 35.1077.

See the product page here:
https://www.tfa-dostmann.de/en/product/wireless-weather-station-with-wind-and-rain-gauge-stratos-35-1077/

The only issue is, that every two hours some packages with wrong data is reported.

The test data which belongs to this pull request can be found here:
blackholefinder/rtl_433_tests@976e273

[zuckschwerdt]

Great code and well documented, thanks!
This should be exactly the case for a "subtype": different sensors in the same sender. I need to check if adding a subtype is a breaking change though. Not really sure.

[zuckschwerdt]

Sadly adding "subtype" is a breaking change, e.g. to the MQTT topics. We need to consider not using "subtype" here.

[blackholefinder]

OK, understand. Braking change is always bad idea.

At the moment I see several options:

• keep old code almost as is, but omit messages which contains the time information (e.g. field B != 0x5). This would lead to consistent weather data, but time information would not be decoded.
• use subtype only for the time information frames (e.g. field B == 0x6)
• only add msg_type field to be able to detect possible wrong weather data (which is in reality time data) in consuming application.
• create new model for this weather station device. But this maybe lead to double decoding, as telegrams would match to both models.
• not support this weather station at all and keep it in my fork only.

Do you have further ideas?
Please tell whats your preference to proceed, than I can adapt the code accordingly.

[zuckschwerdt]

I was thinking to just remove the subtype and keep everything else in this PR the same. Consumers will then usually get the weather fields and sometimes a message with just a time information.