Shell Commands

The simplest way to run a command is to simply type its name into the terminal and hit enter. Commands are small programs, often dubbed binaries or executables (akin to a .exe in Windows), that the shell "calls" (finds and runs) - this is commonly known as command execution.

Commands, for the most part, have a similar syntax that it generally goes like this:

[command name] [option] [argument]

Instead of chucking a massive list of commands at you now, I'll weave them in as we go along, explaining each as they pop up. For now, we're going to hone in on the art of command management.

PATH

By now, you've played with a few commands like cat, grep, and chsh. You might be scratching your head thinking, "Where do these commands come from?" or "How on earth does the shell execute the commands I type?". The magic behind this is called the PATH.

In essence, the PATH is a list of directories nestled within the Linux File System (LFS). Each time you type a command, the shell looks through this list and if the command is found, it executes. Pretty simple, right?

Remember a moment ago when I said that the shell "calls" the command. Well, if you consider the PATH as an old school telephone book that lists where applications "live", then you are looking up their number and literally calling them...

Im not sure that's the intended analogy, but it works for me!

So, it begs the question, "What's actually in the PATH?". The directories that make up the PATH are stored in a shell environment variable in much the same way as the current running shell we saw previously. Therefore, it can be viewed in much the same way:

echo $PATH

An example of the contents of the $PATH variable

As you can see, a list of directories is displayed, each separated by a colon (:).

The path directory order is important

When you enter the name of a command, and the shell looks to the PATH, the listed directories are checked from left to right. Consider the following scenario:

We enter the command test. The shell attempts to call the test command and so consults the PATH. In doing so, it discovers that multiple binaries named test exists in multiple directories:

/usr/local/sbin /bin So, which version of test is executed?

In this scenario, the binary present in /usr/local/sbin is executed as it is present within a directory earlier (leftmost) in the list.

"But what if I want to execute test from /bin rather than from /usr/local/sbin?"...

To override the natural behaviour of the shell, you can specify the exact location of the desired binary you wish to execute by typing out its full location within the LFS - this is properly known as the Absolute Path:

/bin/test

Alternatively, you can modify the order of the PATH variable to ensure the correct binary is called by the shell.

In contrast to certain other operating systems, Linux doesn't initially scan the present directory for an executable before referencing the PATH. Instead, it sets off scouring the PATH right away.

To run executables present in the current directory, they either have to be mentioned in the PATH, or you must provide their absolute (like /home/user/scriptx.sh) or relative path (e.g., ./scriptx.sh). We'll circle back to the notions of Absolute and Relative PATHs shortly...

Locate Commands in PATH with which

Suppose you're certain about the command you intend to use, but you want to know precisely where within the LFS it resides. For that, we turn to the which command:

which COMMAND

Example of the which command

When you execute the which command, it will look in each directory specified in PATH for any binary file that has a corresponding name. However, if there is a "Symbolic Link" to a command (the rough equivalent in Linux to a shortcut in Windows) , then which will display the location of the link only; it will not follow the link to display the directory containing the actual binary in question.