| Age | Commit message (Collapse) | Author |
|---|
|
infrastructure to serve this.
I believe this is a useful improvement. The main two advantages are that a) it is very cheap to check whether a record exists for a given day - just check for file presence and b) there is a reasonable limit to the size of these files (never more than a few hundred records per day) which means that they can just be loaded into ram for processing without any complex splitting operations
Currently, the necessary `~/asteroid-healthloggerd/stepCounter` and `...loggerd/heartrateMonitor` directories are not automatically created and the code just vomits errors into log if it can't write to them.
|
|
initialising.
- This is quite crude - Firstly, we assume that the last record in the file is chronologically the last. This needs more consideration.
- we do check if we're on a fresh boot, but we just give up completely if it's not a fresh boot, which may work, but needs testing to figure out how it actually works in practice
|
|
- this should make things a lot easier to maintain. instead of having everything in the same file, each sensor provides a constructor that sets it up correctly etc.
- this should also get the step counter running when the service starts, which is necessary, since the sensor won't count on most devices unless the service is running.
- this also temporarily removes the gps module from the service. it's not implemented right now anyway.
|
