Executing a command after the preceding command finished processing

Locutus_ has asked for the wisdom of the Perl Monks concerning the following question:

Replies are listed 'Best First'.
Re: Executing a command after the preceding command finished processing by codiac (Beadle) on Mar 30, 2018 at 04:43 UTC
If you want the Perl script to wait for the shell command to complete then the easiest way is to remove the ampersand '&' from the command.	[reply]
Re: Executing a command after the preceding command finished processing by Laurent_R (Canon) on Mar 30, 2018 at 06:26 UTC
Just a naive question: why do you run your system commend in the background if you want to wait for it to complete before you proceed? In other words, it seems to me that you don't really need the ampersand and the `nohup` statement. If you remove them, your Perl script will wait for the system command to complete before proceeding to execute the next code statement. In a way, this is not a Perl question, since you would have the same problem if you did that in a shell script. Just as a side comment, please note that ≒erl is compiling the whole script before it starts executing anything (except for special phaser blocks such as `BEGIN { ...}`). Update: fixed a typo.	[reply] [d/l] [select]
Re: Executing a command after the preceding command finished processing by choroba (Cardinal) on Mar 29, 2018 at 23:46 UTC
Just print the message from the background process: `system '( nohup sleep 5 ; echo Shell done. ) &'; print "Perl done.\n";` [download] Beware: The message can appear in the middle of the command line. Precede it with a newline to prevent it, but it can still do weird things if you e.g. run `mc` in the meantime. ($q=q:Sq=~/;[c](.)(.)/;chr(-\|\|-\|5+lengthSq)`"S\|oS2"`map{chr \|+ord }map{substrSq`S_+\|`\|}3E\|-\|`7**2-3:)=~y+S\|`+$1,++print+eval$q,q,a, [download]	[reply] [d/l] [select]
Re^2: Executing a command after the preceding command finished processing by Locutus_ (Initiate) on Mar 29, 2018 at 23:56 UTC
Thanks for the prompt reply. I'm just confused as to where the code you provided would be placed in relation to the lines I provided. The two lines I provided both have a Perl script that is executed. Would your code follow my code or does it need to be placed elsewhere?	[reply]
Re: Executing a command after the preceding command finished processing by mr_ron (Chaplain) on Mar 30, 2018 at 12:51 UTC
If you want to run the two programs at the same time and wait for both on a system with UNIX shell (wfm on bash and csh) you might try something like: `perl -e 'system q/\ (sleep 5; echo after 5) &\ (sleep 2; echo after 2) &\ wait; echo all done\ /'` [download] Ron	[reply] [d/l]
Re: Executing a command after the preceding command finished processing by bliako (Monsignor) on Mar 31, 2018 at 09:10 UTC
The first point to consider will be codiac's comment: why do you need to run your processes in the background (the `&` at the end of your `system()` calls). If you do need these ampersands it means you have to `wait()/waitpid()` on them to finish. Otherwise you seriously run the risk of your perl program reaching its end and terminating while your sub-processes spawned with `system()` have not yet finished. So, the pattern would be: obtain the PID of each spawned command and `waitpid()` for it. When waitpid() has returned, it means process with that PID has finished and you then print your messages. A good question would be how to get the PID of the command spawned by `system()`. A suggestion can be found in marufcetin's comment and also in this node http://www.perlmonks.org/?node_id=680938 which has this: `my $pid = open my $cmd, "-\|", 'ls -al' or die $!;` [download] Note: "If the "open" involved a pipe, the return value happens to be the pid of the subprocess.", from the perlfunc manual. Alternatively, the "purest" way would be to `fork()` first and `system()` within, without the ampersands. Then, all you have to do is to wait on the forked pid. This is demonstrated in this node: http://www.perlmonks.org/?node_id=671047 And if all these forks look scary, you can use a high level module such as `Parallel::ForkManager` or similar (e.g. `Parallel::Simple`) to spawn your `system()` commands without the ampersands and wait for them to finish. However, from the code you supplied it looks to me that you can improve on your design a lot especially those ampersands...	[reply] [d/l] [select]
Re: Executing a command after the preceding command finished processing by haukex (Archbishop) on Apr 01, 2018 at 10:30 UTC
Seems like you posted a very similar question on StackOverflow as How to use Perl to check when a Unix command has finished processing?	[reply]
Re: Executing a command after the preceding command finished processing by mr_ron (Chaplain) on Mar 31, 2018 at 19:43 UTC
I wonder if a basic point here still needs clarification: I am creating a basic Perl script ... to automate a few different Unix commands that need to be executed in a specific order. Using system() to call these commands: system("nohup scripts_pl/RunAll.pl &"); Do the commands invoked by system need to be executed in a specific order with respect to each other or does the calling script want to run them in parallel before it moves on to other processing? If the invoked scripts need to run in order or don't need to run in parallel then codiac's post may be all you need. If it is the latter case then we need to look at somewhat a more involved approach like some of the other posts suggesting `open`, shell process control, or `fork`. Ron	[reply] [d/l] [select]
Re: Executing a command after the preceding command finished processing by marufcetin (Initiate) on Mar 30, 2018 at 09:56 UTC
try as open PIPE, "\| nohup run_decode.pl";	[reply]
Re: Executing a command after the preceding command finished processing by karthiknix (Sexton) on Apr 03, 2018 at 16:10 UTC
In your case I would suggest to find out the process id for the previous command or dependent command you are running. Wait until that process is complete then go for the next command to execute.	[reply]
Re: Executing a command after the preceding command finished processing by Anonymous Monk on Mar 30, 2018 at 14:00 UTC
You cannot reliably suspend Perl compiling until another command runs 枲I do not think that `BEGIN` would serve you here. Instead, run the predecessor command, then run the Perl script.	[reply]
Re^2: Executing a command after the preceding command finished processing by Your Mother (Archbishop) on Mar 30, 2018 at 16:38 UTC
You really should hush. # File: Mod.pm package Mod; sub import { print `date`; sleep 10 }; 1; # File: code.pl BEGIN { print `date` } use Mod; I'm a little tea pot. # Command line- moo@cow~>perl code.pl \|\| date Fri Mar 30 09:37:09 PDT 2018 Fri Mar 30 09:37:09 PDT 2018 syntax error at code.pl line 3, at EOF Execution of code.pl aborted due to compilation errors. Fri Mar 30 09:37:19 PDT 2018 [download]	[reply] [d/l]
Re^2: Executing a command after the preceding command finished processing by Laurent_R (Canon) on Mar 30, 2018 at 17:53 UTC
I think that the OP really meant "execute," not "compile": ... needs time to process before the next (and final) line should ~~compile~~ execute...	[reply]
Re^3: Executing a command after the preceding command finished processing by Anonymous Monk on Mar 30, 2018 at 22:45 UTC
I did indeed mean execute, sorry about that. Corrected.	[reply]


Keep It Simple, Stupid
	PerlMonks