Afraid of Makefiles? Don't be!
In the last few years, I've had the pleasure to work with a lot of talented Software Engineers. One thing that struck me is that many of them did not have any working knowledge of
Makefiles and why they are useful.
When faced with the task to automate a build process, they often roll their own shell scripts. Common culprits are called
doall.sh in a project folder.
They implement the same basic functionality over and over again:
- Parsing input parameters and environment variables.
- Manually managing dependencies between build steps.
- Error handling (...maybe).
Along the way, they keep making the same basic mistakes:
- Incorrectly handling input parameters and environment variables.
- Missing dependencies between build steps.
- Forgetting to handle errors and — even worse — carrying on with the program execution.
These are issues Makefiles were invented to solve.
If you think that
make is scary, you probably think of complicated build machinery for big software projects. It doesn't need to be that way. Let's hear from the author of
make, Stuart Feldman himself:
It began with an elaborate idea of a dependency analyzer, boiled down to something much simpler, and turned into Make that weekend. Use of tools that were still wet was part of the culture. Makefiles were text files, not magically encoded binaries because that was the Unix ethos: printable, debuggable, understandable stuff.
Make was built in one weekend to solve a reoccuring problem in a simple way.
Before I leave the house, I need to get dressed. I use the same simple routine every time: Underpants, trousers, shirt, pullover, socks, shoes, jacket. Most likely you also have a routine, even though yours might be different.
Some of these steps depend on each other.
Make is useful for handling dependencies.
Let's try to express my routine as a
dress: trousers shoes jacket @echo "All done. Let's go outside!" jacket: pullover @echo "Putting on jacket." pullover: shirt @echo "Putting on pullover." shirt: @echo "Putting on shirt." trousers: underpants @echo "Putting on trousers." underpants: @echo "Putting on underpants." shoes: socks @echo "Putting on shoes." socks: pullover @echo "Putting on socks."
If we execute the
Makefile, we get the following output:
$ make dress Putting on underpants. Putting on trousers. Putting on shirt. Putting on pullover. Putting on socks. Putting on shoes. Putting on jacket. All done. Let's go outside!
Noticed how the steps are in the correct order? By plainly writing down the dependencies between the steps,
make helps us to execute them correctly.
Each build step has the following structure:
target: [dependencies] <shell command to execute> <shell command to execute> ...
The first target in a
Makefilewill be executed by default when we call
The order of the targets does not matter.
Shell commands must be indented with a tab.
@sign to suppress output of the command that is executed.
targetisn't a file you want to build, please add
.PHONY <target>at the end of the build step. Common phony targets are: clean, install, run,... Otherwise, if somebody creates an
makewill silently fail, because the build target already exists.
.PHONY: install install: npm install
Congratulations! You've learned 90% of what you need to know about
Makefiles can do much more! They will only build the files that have changed instead of doing a full rebuild. And they will do as much as possible in parallel. Just try to keep them simple please.
Thanks for reading! I mostly write about Rust and my (open-source) projects. If you would like to receive future posts automatically, you can subscribe via RSS or email: