January 12, 2005, 12:30 PM — In one of last month's columns, we looked at a method for determining whether a process was being run by cron. In that column, we used the ptree command, which shows a process' lineage in an indented display like that shown below, to determine whether the "ancester" of the process is the cron process. In this week's column, we'll examine an even simpler way to determine whether a process is being run interactively -- by using the tty command.
% ptree $$
335 /usr/local/sbin/sshd
2675 /usr/local/sbin/sshd
2677 /usr/local/sbin/sshd
2679 -bash
2682 ptree 2679
The tty command displays the name of the terminal that is open as standard input. In other words, it identifies the terminal device that is associated with your login session. Type the tty command at your system prompt and you will see something like this:
% tty /dev/pts/2
The who command displays this same information along with some additional data, but displays your terminal in a different and less useful format. In particular, if you are logged in more than once on the same system, the tty command will quickly tell you which tty is associated with each login session; the who command will not.
% who shs pts/2 Jan 13 08:22 (10.11.10.11) shs pts/3 Jan 13 09:11 (10.11.10.11)
You can use the tty command in a script to easily determine whether the script has a controlling terminal and is, thus, being run interactively. When the tty command is run non-interactively, the output of the command (at least on a Solaris system) is "not a tty". To test this, you can insert a line like this in your crontab file:
10 10 * * * /usr/bin/tty > /export/home/shs/tty-log
Once every day, at 10:10 AM, a line "not a tty" will be added to your log.
A skeletal script for executing different logic depending on the output of the tty command might look like this:
---------------------------- cut here ----------------------------
#!/bin/bash
if `tty >/dev/null 2>&1` ; then
echo "interactive"
else
echo "non-interactive"
fi
---------------------------- cut here ----------------------------
The "if `tty >/dev/null 2>&1`" statement is one of those cute little Unix tricks that requires a little explanation. Knowing that the tty command will either yield a string such as "/dev/pts/2" or will issue the phrase "not a tty", we need to determine which response we get to determine whether or not we have a controlling terminal.
=One way we could do this would be to construct an (excuse the pun!) awkward command such as this:
if [ `tty | awk -F/ '{print $2}'` == "dev" ]; then
This command would strip out the string "dev" from "/dev/pts/2" or, in the case of a non-interactive execution, yield a null string. This would work, but the command used in our skeletal script is considerably more efficient. Let's look at it again:
if `tty >/dev/null 2>&1` ; then
By tossing out the output from the tty command, the "tty >/dev/null 2>&1" command leaves us with only the return code from the tty command to evaluate. If the tty command is run interactively (i.e., with a controlling terminal), the return code is 0 (success). When run non-interactively (i.e., no controlling terminal), the return code is 1. Our simple logic boils down to "if true, do this, otherwise". This is, of course, the essence of any if statement, but we have arrived at our determination of true/false with very little processing and very little coding.
Thanks to Brian Hatch for recommending use of the tty command for determining whether a script is being run interactively.
