Note to C++/Java programmers

(Other students can continue with What is a loop?)

Programmers at every level think they know it all. And yes, you know what a loop is - or at least you'd better. But loops in the shell are quite different: Lets compare them

If I've confused you, good. Start by forgetting how loops work in whatever language you use. Instead of trying to force the shell to do what you're used to, learn how the shell naturally controls loops. The resulting code will be shorter, easier to follow, and even faster.

What is a loop?

You've used loops before, whether you've ever programmed or not. Suppose you are moving, and your room is full of boxes. You can carry one box at a time, and your job is to carry the boxes to the moving truck. When all the boxes are moved, you get to eat lunch. You used a loop to accomplish the task. We could write the loop like this:

We're going to take a few minutes and discuss this simple loop. First, some terminology:

• the body of the loop consists of the two statements
• pick up box 
• move to truck
• the loop condition is there is a box in the room. 

In a while loop, the loop condition generates a positive or negative result, just like this one does. So that we can talk about shell commands, we will say the loop condition either succeeds or fails. In our case 'there is a box in the room' succeeds if there is a box in the room. If not, it fails.

The loop execution proceeds as follows:

1. the loop condition is evaluated. If it fails, the loop stops, and you do the next thing following the done statement (eat lunch) Otherwise
2. the loop body is executed:

• pick up box
• move box to truck

3. start again at 1.

• is there another box? succeeds
• pick up box
• move box to truck
• is there another box? succeeds
• pick up box
• move box to truck
• is there another box? succeeds
• pick up box
• move box to truck
• is there another box? fails
• eat lunch

Note that there is a dependency between the loop condition and the loop body. Specifically, something in the loop body must affect the loop condition! For us, moving a box to the truck decreases the number of boxes in the room so, eventually, there are no more boxes. If nothing in the loop body caused the loop condition to change, the loop would never stop, giving us an infinite loop. In our case, if we moved the box from one side of the room to the other, our loop would never stop, because we would never decrease the number of boxes in the room.

We have expressed this loop as a while loop. We accomplished this by finding a command that will give us a success or failure status that we can use to indicate the loop is finished. This particular task, however, can also be processed by the other type of loop - the for loop.

A for loop executes the loop body for each item in a list of items. The loop relies on an iteration variable, which, in our case, would be the current box we are carrying. To successfully write a for loop, we must be able to generate a list of the things we want to process. In our example, the list consists of the boxes we have to move. We will name them box1 box2 and box3 and our iteration variable currentbox. Our loop proceeds by assigning the next box in the list to currentbox each time through the list, and deleting that box from the list:

Remember, when the next box is assigned to currentbox, it is removed from the list-of-boxes. Before the loop starts, the list-of-boxes contains box1 box2 box3, and currentbox is empty. Then, our loop proceeds as follows:

The operation of removing a box from the list-of-boxes each iteration is reminiscent of a command we used with commandline arguments : shift! 

The loop continues until the list is empty. Note that the loop body here does not have anything to do with altering the loop exit condition! We could move the box around the room instead of to the truck and the loop would still stop!! The for loop is safer - it is guaranteed to stop because the list will eventually be exhausted!

Here's an example of a for loop in the shell. Let's see if you can figure out what it does:

You guessed it - it runs the cat program successively on each item in the current directory whose name ends in .txt

Copyright 2010 Greg Boyd - All Rights Reserved.