Recent udev versions do not want long-running tasks (like co2monitor) to be invoked with RUN+="...", they will be killed after some time. This is very bad and breaks functionality of the data logger completely.
Thank you for the working software.
Is it possible to set an interval time. If the software writes every 10sec the data in the csv file, it will very big after some time.
Does it split the file after some time (day/month files)?
On systems without systemd, there is no automaic data logging on plugging in. A solution could be to use a udev RUN+="..." again but use a script like this and check if there's systemd support to conditionally start logging.
When hotplugging like hell, co2monitor gets confused and crashes, because the hotplugging handling is not yet backed by enough exception tests.
The following is thus to implemented:
service: on hotplug, check if the device node is already monitored (i.e., does the dbus object exist already)
For some reason, co2monitor in version 0.0.1 crashes everytime at bootup. Actually, there is no reason for the service to be started at all. It does not matter because it's started again when the co2 device is plugged in. But it's annoying.
The service does not recognize if the device is unplugged during warmup.
implement an service method stop_device_logging(devicefile) that takes a device file/path and is also exported via dbus. The invoker can then call that method on the REMOVE action.
For some reason, dh_installdocs installs the manpages all over the place under /usr/share/doc/co2monitor/ so they end up twice there plus gzipped in the man system. This is useless and needs to be fixed.