Currently, in the launcher script the whoami and updater loop are masking the exit code of Jackett. So even if Jackett errors out, the systemd service sees an exit code of "0", which e.g. breaks the `Restart=on-failure` in the [non-mono service template](https://github.com/Jackett/Jackett/wiki/Systemd-service#not-using-mono).
This commit stores the exit code of Jackett in a variable and exits the script explicitly with this exit code, after the updater has been waited for. This way `Restart=on-failure` again can apply and `systemctl status jackett` reports the correct status if Jackett failed.
Signed-off-by: MichaIng <micha@dietpi.com>
+ Execute script with system default bourne shell. On Debian this is "dash" by default, which has a much lower memory foot print compared to bash. As well systems without bash won't fail. "${BASH_SOURCE[0]}" cannot be used then but "readlink -f" is another reliable method to get a full file path, using another coreutils binary like "dirname".
+ Execute Jackett with "--NoRestart" option. This has currently not effect, but it assures that the updater never attempts to restart the Jackett process, which is done by the systemd unit already. This avoids possible restart collisions if anything about the updater behaviour changes, e.g. to fix the originating issue.
+ Remove the "Jackett update complete" print. This is simply wrong if Jackett was stopped ordinarily and the updater itself logs to the same destination, including a "finished to copy files" after completed.
Signed-off-by: MichaIng <micha@dietpi.com>
MichaIng