I'm trying to do the following:
1) Run a bunch of jobs in the background
2) Determine if any one of them returns with a non-zero exit status

Here's what I've come up with so far:
##########################################
#!/bin/ksh

while [ ${#pid_array[*]} -lt 1024 ]
do
SLEEP_TIME=`expr 1024 - ${#pid_array[*]}`
nohup exit_script ${SLEEP_TIME} ${#pid_array[*]} > /dev/null &
SAVE_PID=$!
echo "PID:${SAVE_PID} SLOT:${#pid_array[*]} SLEEP:${SLEEP_TIME}"
pid_array[${#pid_array[*]}]=${SAVE_PID}
done

jobs -l

echo "Waiting..."
CNT=0
while [ $CNT -lt ${#pid_array[*]} ]
do
wait ${pid_array[${CNT}]}
echo "PID:${pid_array[${CNT}]} CODE:$?"
CNT=`expr $CNT + 1`
done

exit 0
##########################################

exit_script is just a test program that sleeps for $1 and returns with a code of $2 (mod 256, of course).

The problem I'm having is if I perform the wait on a given pid, and that pid is already done, it doesn't always return with the proper code. Specifically, the first 25 will return with the correct code while the rest return with 127, wait's way of saying it doesn't know the code.

This leaves me with two questions:
1) Is there a way to make it "remember" more than 25 return codes?
2) Is there a better way to do this?

If it matters, this is being ran using ksh on Solaris 8. Thanks for any help you can provide.

Joel