Block incorrect raw timestamp

[Navid200]

After a hard reset, the raw timestamp is incorrect until after sensor start.
With the incorrect timestamp, the corresponding activation time is incorrect resulting in the transmitter days parameter to be reported as -1.
As long as transmitter days is -1, xDrip refuses to start sensor.
This creates a catch-22 situation as reported here: #3078.

This PR does not allow an incorrect raw timestamp to be used to update the transmitter age.
Tests show that the catch-22 situation is avoided after this PR.

[Navid200]

This has been tested.

[jamorham]

Can you elaborate on what sort of raw timestamp is being seen during this condition? I'm not sure if this is the best way to deal with this without knowing what the condition is.

[Navid200]

Currently, xDrip sees a raw timestamp of 17 years. When it sees that, it refuses to enquire for a native timestamp. Without one, it returns -1 for transmitter days. With -1, it refuses to start a sensor.

After this, when xDrip sees a raw timestamp greater than 10 months (300 days), it ignores it and replaces it with 0.
I am happy to change that to 600. But, I don't think any battery currently lasts 10 months.

With a 0 raw timestamp, xDrip goes ahead and asks for the native timestamp. Unless the transmitter is really broken, it should get 0. With 0 as transmitter days, xDrip allows a sensor start.

The following is cosmetic.
When the raw timestamp becomes 0, xDrip stops showing it on the status page. So, the user will only see 0.
When the raw timestamp grows to something like 0.1. The user will see 0/0.1 for transmitter days on the status page.

The only new added log is a detail log stating that we are ignoring an incoming timestamp, which we believe is invalid.

I hope I did not misunderstand the question.

[Navid200]

The raw timestamp sent after a hard reset is 6214 days (7 years).

[Navid200]

Let me be the devil's advocate:

Let's say someone solders wires to the battery contacts and wires them to a power supply such that in 7 years the transmitter is still running.
At that point, after this PR, the raw timestamp will be switched to 0. But, the native timestamp will show 6214.
But, as soon as transmitter days goes over 180 days, my understanding is that the transmitter will refuse to start a session.
I don't see the point of such an experiment.
Regardless, in 7 years, I don't think G6 will still be as popular as it is today.