[gustaf]

There are two exit codes and therefore, two possible points of failure: doas and the command.

Wrapping an exit code test in a loop does what I expect -- until doas(1) succeeds (exit 0), but the command fails (exit >0). The command failure is reported correctly, but the program is stuck in an infinite loop because the exit code for the command is not zero.

To test this, I made a simple script:

# Write a trace for each command.
set -x

doas ls ~/NoDirectory
until [ $? -eq 0 ]
do
  doas ls ~/NoDirectory
done

echo 'Script successfully completed!'

OUTPUT:
#   + doas ls /home/gustaf/NoDirectory
#   doas (gustaf...) password: 
#   ls: /home/gustaf/NoDirectory: No such file or directory
#   + [ 1 -eq 0 ]
#   + doas ls /home/gustaf/NoDirectory
#   doas (gustaf...) password:
#   ls: /home/gustaf/NoDirectory: No such file or directory
#   + [ 1 -eq 0 ]
#   + doas ls /home/gustaf/NoDirectory
#   doas (gustaf...) password

The loop is supposed to be for doas(1) only. I need some way to report an error and exit the script if the command fails.