this "splits" the output into sections so it's easier to attribute what
is curretly run. useful for debugging failures. in the end, the script
reports total CPU time it took to run the update.
the apply.sh script should assume current directory is arbitrary after
executing each step: base, bitcoin, lnd and so on.
this commit simply changes current dir to the repo root before executing
every step.
the first run of update.sh sets up a cron job to run itself
periodically. the update script takes one position argument, the git
branch to pull changes from, which is also known as "sysupdates channel"
- master for stable, dev for testing.
the sysupdates channel was lost in the cron job setup, always defaulting
to master (stable). this commit preserves the original arg value given
to update.sh during the initial run.