Popen3 Problem
 

Hello everyone. I have created a sales system for my company. Basically
you telnet to the system and enter your login name, password ... From
there you can access the sales system.

I am working on a feature to write reports for a certain area of the
system. For this, I want it to just shell out to VIM (Vi iMproved). For
this I will be using Popen3, so the telnet user can access the sytem.

This is some example code I have ... ( I know its Ruby code, but I know that Popen3 is popular with Perl. I'm sure a Perl user can answer this question as it more than likely isn't a problem with the language itself. )

def runVimForReport( node_number, file_name )

  stdin, stdout, stderr = Open3.popen3('vim #{file_name}' )



  read_thread = Thread.new { start_read(stdout,stderr) }

  write_thread = Thread.new { start_write(stdin,stderr) }



  write_thread.join

  read_thread.join

end



def start_read( stdout , stderr )

  loop do

    the_input = stdout.getc

    nodeSendData( @node_number, the_input.chr )

    STDOUT.flush

  end



  puts("EXIT READ THREAD")

end



def start_write( stdin , stderr )

  loop do

    STDOUT.flush

    input = nodeGetChar( @node_number )

    stdin.putc(input)

  end

end

This code works well, the user can access VIM just fine. There is only
one problem ...

When the user exits the program it continues to read/write to the
process. It doesn't just simply exit the functions. How do I tell when a
program has exited? I tried other programs and all are the same ... It
just hangs. The user has to logout and log back in. Even when
disconnecting the telnet session, the program is still running in the
background.

Is there a variable I can check in my loop to see if the program has
exited?

Thanks for all your help!

- Matt

Re: Popen3 Problem
 

I figured it out!

It wasn't that the program wasn't exiting. Ruby didn't know how to handle a nil variable in my code.

I changed my start_read function to:

def start_read( stdout , stderr )

  puts("Read thread is running ...")

  loop do

    # errorline = stderr.readline

    #puts("STDOUT #{errorline}")

    the_input = stdout.getc

    if the_input == nil                        # <=== This is what I added.

      Thread.kill(@write_thread)          #

      break                                          #

    end                                               #

    nodeSendData( @node_number, the_input.chr )

    STDOUT.flush

  end

end

So, when the program exits the input becomes nil. After discovering that it is nil, I killed the other thread. Everything works great now.

Thanks for nobody's help :)

- Matt