pwait waits for process supplied on the command line to finish. It's
different from wait because it can wait for any process to finish not just
children of the current shell. When all the processes supplied can no longer be
found pwait exits with 0.
The above replaces idioms such as tail --pid=<PID> --follow /dev/null && do_something
which now becomes pwait <PID> && do_something.
The program can be invoked as follows pwait [options..] <PID..> <NAME..>
Process can be given by PID or by name, pwait 1735 gcc. Name resolution is
not exact, meaning all processes that <NAME> is a substring of will be waited.
Several options can be given that control the behavior of pwait.
-t, --timeout <NUMBER> tells pwait to exit with 1 after <NUMBER> of
seconds have passed and not all the processes have finished.
pwait --timeout 30 7204 waits for the process with PID 7204 for a
maximum of 30 seconds, if the process hasn't finished by then it exits
with 1. <NUMBER> can also be a decimal number, such as 0.4. By default,
pwait waits indefinitely for the processes to finish.
-n, --interval <NUMBER> tells pwait to sleep for <NUMBER> of seconds
after each check for the existence of the processes.
pwait --interval 3 7204 periodically sleeps for 3 seconds after checking
for the existence of the process with PID 7204. Default is 1.
The program is written to conform to the C++ 98 standard.
It can be compiled the following way:
git clone https://github.com/begone-prop/pwait.git
cd pwait
g++ -Wall -Wextra -O2 -std=c++98 -pedantic ./pwait.cpp -o ./pwait
Note: There is no guarantee that the process you are waiting for is the one
you actually want because it's only defined to wait for processes that are
your children. It's possible that in the time interval between checks for the
existence of the process the original one terminated and a new, different
process was created and was assigned the same PID. Using smaller values for the
--interval option make the odds of this happening smaller.