wait normally returns the exit status of the last job which terminated. It may also return 127 in the event that n specifies a non-existent job or zero if there were no jobs to wait for.
wait needs to be aware of the job table of the current shell execution environment, it is usually implemented as a shell builtin.
This command can be useful where part of a script can execute in parallel to implement a barrier where an upcoming section depends on the successful completion of the preceding sections.
The following example will fetch the src/ directory from a machine named iona using rsync and simultaneously update the libraries on which this program depends, before building the combination.
#!/usr/bin/env bash # Parallel update script which makes use of the wait command # Update local copy rsync iona:src/ . & # Upgrade required libraries, or exit indicating failure if make failed for some reason make -C lib || exit 1 # Wait for rsync to terminate (may have already happened) and finish the job, unless rsync failed wait && make
Wait for specified job control id number:
$ ls -R / > /dev/null 2>&1 & # start any long running background process  1986 $ wait %2 # waits for background job number 2 to terminate, then returns