This is part twelve of a series on how to approach bash programming in a way that’s safer and more structured than your basic script.
See part 1 if you want to catch the series from the start.
Last time, we showed the elements of a strict mode for bash. This time, let’s discuss some of the implications of strict mode for how you approach programming with bash.
Bash is very forgiving to errors and bash commands don’t tend to have the same success rate as functions typically do, at least in other languages.
For these reasons, failure is something that has to be dealt with much more frequently than you might expect. Scripts that might work fine most of the time will exit unexpectedly when strict mode is enabled.
This is probably not a bad thing, since most of the time you’d want to know that a command failed if you didn’t know.
However, you really may not care that a command fails because it doesn’t matter. There are plenty of examples of commands which fail because what they are attempting to do has already been done.
There are three methods for dealing with “acceptable” errors:
find a method that doesn’t error
toggle strict mode off and on
negate the command with an exclamation point
Then Don’t Do That
The first method is to find a way to run the same function or command without causing an error in the first place.
This is useful when the error that the command is reporting doesn’t matter to the rest of your program. The archetypical example is making a directory. If the directory already exists, then that’s all you were trying to accomplish in the first place.
However, mkdir will error if you tell it to make a directory which already exists. Fortunately, there’s an option to mkdir which will suppress the error when the directory already exists: –parents, or -p.
Similarly for the rm command, there’s the –force option, or -f. A number of commands have such a force option.
There are more examples, but this is a method you’ll have to learn for yourself, since it’s command-specific. The documentation is your friend.
Toggle Strict Mode
If necessary, turn strict mode off and on again. This is usually required when you are sourcing code which wasn’t written for strict mode.
We’ll make a function later to turn strict mode off and on so that we can toggle it easily.
Errexit does not apply to conditional expressions. There are several ways that commands can be part of conditional expressions:
as the condition of an if statement
as the condition of a while loop
as the left side of an || (or) expression where the right-hand-side is true
in a negation
The simplest of these is negation:
Even when the command succeeds and is then negated (resulting in false), the false doesn’t trigger errexit because it’s considered a conditional expression.
Continue with part 13 - implementing strict mode