Beefy Boxes and Bandwidth Generously Provided by pair Networks
Problems? Is your data what you think it is?

Re^3: using online translation engines with perl

by Anonymous Monk
on Nov 07, 2018 at 07:25 UTC ( #1225348=note: print w/replies, xml ) Need Help??

in reply to Re^2: using online translation engines with perl
in thread using online translation engines with perl

I'm baffled that pid's are zero within a block, but not so when they finish.

fork causes the execution of a process to split into two different processes, sharing the initial memory layout but not the following changes to the variables. To help the program discern which of the two it has become, the newly produced ("child") process receives 0 as the return value of fork() and the old one ("parent") gets returned the PID of the new process. This is why you had to put an exit 0; at the end of the if ($pid == 0) block: otherwise both the parent and the child would continue executing the same parts of the program, causing a lot of confusion.

wait returns the real (former) PIDs of the freshly terminated child processes. If the child wanted for some reason to know its PID, it could have used the $$ variable.

This example might help:

if (fork() == 0) { system("sh", "-c", "sleep 3 # getting yandex translation"); exit(0); } if (fork() == 0) { system("sh", "-c", "sleep 2; # getting google translation"); exit(0); } system("pstree -Apal $$"); # pstree gets the PID of the parent process + as its argument while((my $pid = wait()) != -1) { say "$pid terminated" } __END__ perl,29173 |-perl,29174 | `-sh,29177 -c sleep 3 # getting yandex translation | `-sleep,29179 3 |-perl,29175 | `-sh,29178 -c sleep 2; # getting google translation `-pstree,29176 -Apal 29173 29175 terminated 29174 terminated
pstree command draws all child processes of the main Perl process (including itself) with their command line arguments. The Perl processes having the same command line arguments but different PIDs are the copies created by fork().
What alarm truly does here is unclear. Finally, there is no code to kill pid's when the timer reaches zero.
alarm arranges for a SIGALRM signal to be delivered to the calling process in the number of seconds specified as its argument. It behaves as if something else kills the process which had called alarm and forgot to disarm it. The good part is that one can use $SIG{ALRM} to trap the signal and do something meaningful instead of dying, but in your case the work is already being done in a child process, and we can let it be killed:
my $start = time; if (fork() == 0) { # arm the alarm clock alarm(10); # create a child process that sleeps system("sleep 365d"); exit(0); } system("pstree -Apal $$"); while((my $pid = wait()) != -1) { say "$pid terminated" } say time-$start, " seconds elapsed instead of a year" __END__ perl,29620 |-perl,29621 | `-sleep,29623 365d `-pstree,29622 -Apal 29620 29621 terminated 10 seconds elapsed instead of a year

Log In?

What's my password?
Create A New User
Node Status?
node history
Node Type: note [id://1225348]
and the web crawler heard nothing...

How do I use this? | Other CB clients
Other Users?
Others making s'mores by the fire in the courtyard of the Monastery: (3)
As of 2019-08-18 10:55 GMT
Find Nodes?
    Voting Booth?
    If you were the first to set foot on the Moon, what would be your epigram?

    Results (134 votes). Check out past polls.