Diagnose¶

• RTC drift

The drift of intern RTC is enormous. On CANRec its running way to fast (LSE not even enabled? and not used by RTC?).
The calibration is capable to adjust several seconds per month, not per day.
Not sure if an per-device config is even necessary or if it is an per type thing.

CANRec can request time values periodically from all devices. It can only compare it to the own clock, so this is not usefull.
Cordyceps can analyse this. It has a "real" working RTC form the computer.

Events¶

There is no logging available from devices that are only connected via CAN. The heartbeat "Warning" is nice. But useless as long as
the user does not know whats going on.
An event has to be send via CAN so at least Cordyceps can show the reason of warning.

Each warning needs to handle an Event-code additionally. Having it in the logging would be nice and clean for the using code.

Using something like errno would make it not thread-safe. Having extended logging functions/macros looks acceptable.

*Use-cases

• write unknown Object; Object is interesting; within eventloop
  • lCriticalE( UNKNOWN_OBJECT ); nicht mehr generisch
• undevoltage; Voltage interesting; within eventloop
• Buffer overflow, not within eventloop

sendEvent(Error, );
Logging can not handle units.