Bash：在Shell脚本中使用进程ID

Question

I am a bit confused on the uses of PIDs and parent/children processes. I have been reading up on them and I understand as far as the fact that when a program starts it makes an exact copy of itself (the child) and each of them have unique PIDs. But I am not sure if I can use it in a shell to tell me when certain aspects of that shell program have finished.

For a better example (pseudo code):

 for ((i = 0; i < 10; i++))
  for a_name in "${anArray[@]}";do
     a series of math equations that allow the values associated with a_name to run in the background simultaneously using '&' earlier in the code         
  done
wait

  for a_name in "${anArray[@]}";do
     same as above but diff equations
  done
wait
done

I am hoping to be able to see when any given value is finished in the command, so that that particular value can move on to the next for loop and command.

I have seen that wait can take a job identifier as an argument (wait%1 or wait $PPID) but I am not really sure how those would be implemented.

Does anyone have advice on how to use PIDs and/or have a link to a super good tutorial? (and I mean super good, I need some laymen's terms thrown in there)

Thanks!

Answer 1

You can wait on a single process:

command ${array[0]} &
waiton=$!
# Do some more stuff which might finish before process $waiton
wait $waiton

You can wait on all children processes:

someLongRunningCommand &
(
    for value in "${array[@]}"; do
        command "$value" &
    done
    wait
)
# wait with no arguments waits on all children processes of the current
# process. That doesn't include `someLongRunningCommand`, as it is not
# a child of the process running the subshell.

Other situations are trickier and may be better handled by xargs , parallel , or some other method.