What is shell's maximum background jobs PID stored exit-codes?

Question

When running a shell job in the background, the shell keeps track of the exit-code that can be retrieved with wait <PID>.

When a new background job is run, it gets its own new PID, even if previous background jobs are ended, because the exit-code of a background-job is stored and kept as a PID indexed entry, until the parent shell ends.

How many background jobs can be repeatedly run, before the allocated PID counter overflows and overwrites previously stored exit-codes?

Since the stored exit-codes are not released even by wait <PID>, it appears to cause a memory leak of an ever growing exit-codes store.

Is there a command or option to release a stored background job's exit-code if it is no longer needed?

Here is a test code to show that background-jobs per PID exit-codes are kept stored.

#!/usr/bin/env bash

typeset -a background_jobs

report() {
  # wait get exit-code of background job 0
  wait "${background_jobs[0]}"
  printf 'Exit-code of background job 0 is: 0x%02X\n' $?
  # wait get exit-code of background job 1
  wait "${background_jobs[1]}"
  printf 'Exit-code of background job 1 is: 0x%02X\n' $?
  # exit-code of job 0 still not released
  wait "${background_jobs[0]}"
  printf 'Still stored exit-code of background job 0 is: 0x%02X\n' $?
  # exit-code of job 1 still not released
  wait "${background_jobs[1]}"
  printf 'Still stored exit-code of background job 1 is: 0x%02X\n' $?
  local -i job_pid
  printf '\nPID       Ret\n-------------\n'
  for job_pid in "${background_jobs[@]}"; do
    wait "$job_pid"
    printf '% 8d 0x%02X\n' "$job_pid" "$?"
  done
  exit 0
}

trap report EXIT

for ((i = 0; i < 25; i++)); do
  {
    exit $((RANDOM & 16#7F))
  } &
  background_jobs+=($!)
done 2>/dev/null

What is shell's maximum background jobs PID stored exit-codes?

0 Answers0