Instead to implement own logic how the database scheme should be updated
or migrated to a newer or older version flucky use now instead the
go-migrate package.
Add additional functions to the repository to add or update devices,
sensors or measured values. Furthermore the test has been adapt to the
new functions.
Add timezone for the columns creation_date and update_date, otherwise
it's difficult to compare the times correctly in the unit test.
Furthermore the default value of the creation_date will now be defined
by the postgres implementation of the database interface.
changes:
- fix: read temperature values without daemon
Add subcommand to read temperature values without starting the daemon
- fix: implement measured value types
Replace measured value types with constants
- fix: add sensor pipelines
Add functions which returns a channel with measured values
- fix: filter measured values from a channel
Add functions to filter measured values by sensor id or measured
value types.
The default directory of the sqlite db has now been changed depending of
the user who executes flucky.
The place where the sqlite db will be stored is unchanged for the root
user, but for all other users the sqlite db will now be created to
~/.cache/flucky
changes:
- Remove obsolete architectures and environment variables from the
Makefile. Remove rpm-build steps. This will be exported in an own
repository.
- Adapt the travis file. Remove obsolete architecture steps.
Changes:
- Renamed storage endpoint into dsn (data source name).
- Add additional dsn fallback property. This dsn will be used in futes
to store informations, if the main dsn backend does not work
correctly. For example, if no connection can be established over the
network to a database.
changes:
- Implement repository test for the sqlite backend
- Add testutils package to start container images
- Remove deprecated till_date in measured values
- Renamed columns of the table humidities, pressures and temperatures
