178

I have three scripts I need to run when I start up my Ubuntu machine, they start services I use in my development environment.

To do that, I manually open three terminals and type in the commands.

Is there any way to create a script that will open three terminals and execute one command in each of these? (Each command should be in a separate terminal window so I can see their output).

169
gnome-terminal -- command

or

xterm -e command

or

konsole -e command

Pretty much

terminal -e command

To make the terminal stay when the command exits:

In konsole there is a --noclose flag.

In xterm, there is a -hold flag.

In gnome-terminal, go to Edit -> Profile Preferences -> Title. Click the Command tab. Select Hold the terminal from the drop-down menu labelled When command exits. You should create a new profile for that and execute with

gnome-terminal --window-with-profile=NAMEOFTHEPROFILE -e command
| improve this answer | |
  • 2
    If I try to hold the terminal, I get "child process exited normally with status code 127" – Darshan Chaudhary Aug 12 '16 at 7:49
  • 2
    gnome-terminal does not have the title option any more :( – törzsmókus Feb 22 '17 at 13:04
  • 1
    @törzsmókus it is 2017 indeed! LTS releases have a 5 year support life. 14.04 does not end until April 2019. wiki.ubuntu.com/Releases – bhass1 Mar 5 '17 at 22:33
  • 1
    gnome-terminal -e command only works if command is quoted. So this does not work: gnome-terminal -e "echo hello world; sleep 3" but this does: gnome-terminal -e "bash -c 'echo hello world; sleep 3'". Sigh. – bgoodr Apr 16 '17 at 19:47
  • 5
    consider using now: gnome-terminal -- command – dallonsi Jul 1 '19 at 7:35
69

Instead of hard-coding gnome-terminal, konsole, et cetera, use the Alternatives system. The program that executes the default terminal emulator is:

x-terminal-emulator

On my system, it opens a new instance of Konsole every time I execute this command.

Luckily, the terminals seems to support the -e option for executing a command (I verified it for konsole and gnome-terminal). Arguments after the command are passed to the invoked command. Bash refuses to stay open in my terminal, an additional script is needed to get a terminal:

#!/bin/sh
"$@"
exec "$SHELL"

If you've saved the previous script as /home/user/hacky and made it executable, you would run your scripts with:

x-terminal-emulator -e /home/user/hacky your-script optional arguments here

The full path is required and /home/user/hacky has to be executable.

My previous attempt to run a script in a new terminal window can be found in revision #2, it was before I realised arguments can be passed to x-terminal-emulator.

| improve this answer | |
  • In this case, it won't help much as the asker wants to do something that isn't the same for all terminals. – nickguletskii Jun 2 '11 at 20:36
  • Attempt #3: this one should keep the terminal open and run the program with optional arguments. – Lekensteyn Jun 2 '11 at 21:06
  • 1
    I used gnome option, however, once I run my script, the main terminal closes !! .. any idea why ? – McLan May 20 '15 at 14:33
  • 3
    @Suda.nese That is by design, when the "terminal" is done executing the script it will quit because there is nothing more to do. You could "fix" this by invoking a shell where you can execute commands (bash) or have a line such as read -p "Press Return to continue". – Lekensteyn May 20 '15 at 15:44
  • 1
    How can you run more than one command in the terminal? For example cd xxx && start.sh. The interpretor sees the && as the second part of the command (which is logical), but if I quote it, then it tries to exec the whole thing as one big argument – Richard May 25 '17 at 7:02
10

Quite simply-

#!/bin/bash

/etc/init.d/ccpd status

This is enough for other commands that do not need to display anything in terminal. But here one has to see the status displayed.
So, it needs to run in a terminal window

#!/bin/bash

gnome-terminal -e "/etc/init.d/ccpd status"  --window-with-profile=NAMEOFTHEPROFILE

The other post intended [] to be a placeholder

Here "NAMEOFTHEPROFILE" is to be replaced with the name of the profile that "Holds the terminal when the command exits".

enter image description here

enter image description here

| improve this answer | |
  • 1
    @cipricus i believe [] was just placeholder – Karthik T Jan 23 '13 at 9:37
  • got it. but i have to make the terminal not to close so fast. i guess that is also in the linked question – user47206 Jan 23 '13 at 9:43
  • @cipricus have you tried the profile one? Just need to add --window-with-profile=NAMEOFTHEPROFILE to what I have given – Karthik T Jan 23 '13 at 9:44
  • 1
    @cipricus I would have to get back home to give better instructions, but the idea is to create a special profile with that option set, and use the name of the special profile in the place above. – Karthik T Jan 23 '13 at 9:51
  • 2
    @cipricus if that is enough for you, then that is ok. Profile is nothing more than a group of settings. You can setup the settings ONLY for use in your script, and not have to use it in all terminals. You can see Edit -> Profiles to see all the profiles you have, and you would add one there which was setup as explained in the post you linked – Karthik T Jan 23 '13 at 15:39
8

UPDATE 17 FEB 2020: this answer is now obsolete.

Click this link and use this other answer instead: Open Terminal with multiple tabs and execute application.


Aided by @nickguletskii's answer, and my own comment under his answer, and inspired by @grabantot's upvote of my comment, here's my preferred way to do it, especially when I want the terminal to stay open so I can then manually use it.

Ex. usage: this is really useful to add to your startup programs so this script will run, open a terminal, create and name a tab in the terminal, and run a command for you. Or, you can just add a symlink to this script to your desktop. I use this type of approach so I can double-click a single icon on my desktop and have it open up a bunch of terminals (with various tabs named according to what work I'm going to do in them) and programs to set up my programming environment, for instance, for daily work.

Here's a contrived example, which opens up a single tab, titles it "test", then runs the simple command cd /etc; ls inside it. The $SHELL part at the end forces the shell to stay open so you can then see its output and continue using it (I learned this somewhere else on Stack Overflow):

gnome-terminal --tab --title="test" --command="bash -c 'cd /etc; ls; $SHELL'"

Here's a more complicated example which opens up 3 separate tabs in the same gnome-terminal. This is exactly the type of thing my desktop shortcut does so I can open up a bunch of programming windows at once:

gnome-terminal --tab --title="tab 1" --command="bash -c 'cd /etc; ls; $SHELL'" --tab --title="tab 2" --command="bash -c 'cd ~; ls; $SHELL'" --tab --title="tab 3" --command="bash -c 'cd ~/temp3; ls; $SHELL'"

Here's a breakdown of that command above:

  • gnome-terminal = open up a gnome-terminal
  • --tab = open up a unique tab for what comes next
  • --title="tab 1" = title this tab "tab 1"
  • --command="bash -c 'cd /etc; ls; $SHELL'" = run the bash -c 'cd /etc; ls; $SHELL' command, which is a command I just made up as an example; here's what it does:
    • bash -c says it is a bash 'c'ommand
    • cd /etc = 'c'hange 'd'irectory into the "/etc" path
    • ls = 'l'i's't contents of this directory
    • $SHELL = this cryptic tidbit is required to keep the shell open so you can work with it. If you want the shell to open, run your command, then close, simply remove this part. I, however, want the tab to stay open so I can go make programming magic. :)
  • we then start back over at the --tab part to produce tab 2, then again for tab 3. Customize to your heart's content.

Screenshot:

enter image description here

| improve this answer | |
  • 1
    glad I was helpful) I also have scripts that I can just click on and start working on the project. There were two problems with them: lots of terminal windows (had a whole separate screen for them) and windows closing after server crashes for example. That answer solves both of my problems with --tab + $SHELL. Nice – grabantot Dec 31 '18 at 9:33
  • How can I open a tab and set it as active, at the moment I open a tab and then have to click on it as well to make it the active tab – Alfa Bravo Aug 21 at 17:47
  • @AlfaBravo, not sure. You might consider scripting actual keyboard presses as though a human had done them. Ex: xdotool key --clearmodifiers Super+d, as I use in my show-desktop.desktop file, presses Windows Key + D to toggle showing the desktop--same as if a human had done these key presses. I'm sure you could just script pressing Ctrl + PgUp or Ctrl + PgDn the right number of times to select the tab you want, as that's how you can switch tabs manually as a human. – Gabriel Staples Aug 21 at 18:01
1

commenting for the answer by Lekensteyn. I know this is a old post, but for anyone who finds this useful (as I just did) Instead of making another "hacky script" with just put a function inside the script you are calling

hacky_function()
{
"$@"
exec "$SHELL"
}

Call your script with "x-terminal-emulator -e /path/to/script hacky_function optional arguments here"

Don't forget to put "$@" at the end of the script

| improve this answer | |
0

Almost a decade late to the party but, here's my answer using Python.

In the .gif below I launched the program from an existing terminal with screen recorder running to show what it would look like at login:

dellstart.gif

The python program contains the commands to run after gnome-terminal is opened:

commands = [ 'xdotool type "cd ~"',             # Change to home directory
             'xdotool key Return',              # Enter Key
             'xdotool type "./ssh-activity"',   # Suspend after 15 minutes
             'xdotool key Return',              # Enter Key
             'xdotool key Control_L+Shift_L+T', # Open new terminal tab
             'xdotool type "cd ~/askubuntu"',   # Change to working directory
             'xdotool key Return',              # Enter Key
           ]

You would customize these commands to suit your needs.

The entire python program:

#!/usr/bin/env python
# -*- coding: utf-8 -*-

#==============================================================================
#
#       dellstart - Autostart programs on Dell Fileserver
#
#==============================================================================

'''
CALL:

   dellstart

REQUIRES:

    sudo apt install xdotool
    
'''

from __future__ import print_function           # Must be first import
import external as ext
import time

commands = [ 'xdotool type "cd ~"',             # Change to home directory
             'xdotool key Return',              # Enter Key
             'xdotool type "./ssh-activity"',   # Suspend after 15 minutes
             'xdotool key Return',              # Enter Key
             'xdotool key Control_L+Shift_L+T', # Open new terminal tab
             'xdotool type "cd ~/askubuntu"',   # Change to working directory
             'xdotool key Return',              # Enter Key
           ]


def process_commands(command_list):

    active_pid = ext.launch_command('gnome-terminal')
    if active_pid == 0:
        print("ERROR launching 'gnome-terminal'. Aborting 'dellstart' script")
        exit()

    time.sleep(3)                                # Wait a few seconds

    for command in command_list:
        #print('starting command:',command)
        execute_one_command(command)


def execute_one_command(command):

    active_pid = ext.launch_command(command)
    if active_pid == 0:
        print('ERROR launching command:', command)
        exit()
        
    while ext.check_pid_running(active_pid):
        time.sleep(.1)                          # Sleep 50 milliseconds
        # TODO: Set timeout of 10 seconds

    #print('command completed:',command)


if __name__ == "__main__":

    process_commands(commands)
    exit()

Change #print to print to debug where you are at in the program as it runs.

The module that is included:

# -*- coding: utf-8 -*-
#==============================================================================
#
#       external.py - Used by mserve, encoding.py and dellstart
#
#==============================================================================

from __future__ import print_function       # Must be first import
import os
import time


def launch_command(ext_name):
    ''' Launch external command and return PID to parent
    '''

    all_pids = pid_list(ext_name)
    new_pids = all_pids

    os.popen(ext_name + ' &')           # Run command in background
    sleep_count = 0
    while new_pids == all_pids:
        new_pids = pid_list(ext_name)
        if sleep_count > 0:
            # self.lib_top.after(10)    # Fine tune for sleep count of 2
            time.sleep(.01)             # sleep 10 milliseconds
        sleep_count += 1
        if sleep_count == 1000:         # 10 second time-out
            print('launch_ext_command() ERROR: max sleep count reached')
            print('External command name:',ext_name)
            return 0

    #print('launch_ext_command() sleep_count:', sleep_count, all_pids)
    diff_list = list(set(new_pids) - set(all_pids))

    if len(diff_list) == 1:
        return int(diff_list[0])

    print('launch_ext_command() ERROR: A new PID could not be found')
    return 0


def pid_list(ext_name):
    ''' Return list of PIDs for program name and arguments
        Whitespace output is compressed to single space
    '''
    all_lines = []
    # Just grep up to first space in command line. It was failing on !
    prog_name = ext_name.split(' ',1)[0]
    all_lines = os.popen("ps aux | grep -v grep | grep " + \
                        "'" + prog_name + "'").read().strip().splitlines
    PID = []
    for l in all_lines():
        l = ' '.join(l.split())         # Compress whitespace
        PID.append(int(l.split(' ', 2)[1]))

    #print('pid_list:',PID)
    return PID


def check_pid_running(active_pid):
    ''' Parent will call us 10 times a second until process is finished. '''
    if active_pid == 0:
        print("check_pid_running() ERROR: argument is '0'")
        return 0                    # Programmer error

    try:
        os.kill(active_pid, 0)      # 0 is status check, 9 kills
        return active_pid           # pid is still running
    except OSError:
        return 0                    # pid has finished


# End of external.py

Note you may have to tinker with time.sleep(.01) depending on the speed of your machine. That's about the only drawback of this methodology I've discovered.

I plan to use this script to startup many application after login. Especially to move windows over three monitors for the perfect startup desktop.

| improve this answer | |
-3

Use the screen command and -d detach from an existing screen session, and reattach here -m force a new screen session -S create a named session instead of using the default name

| improve this answer | |
  • 1
    this answer is not clear at all, please care to make more understandable – azerafati Apr 20 '17 at 12:18
  • @azerafati indeed, screen doesn't open up any terminal window... is not even intended by that awesome package... – m3nda Jul 12 '18 at 19:55

Not the answer you're looking for? Browse other questions tagged or ask your own question.