How to Speed Up Python Code Reviews With Linting

One fateful Thursday morning I sat down with a fresh cup of coffee, ready to dig in and give some feedback on a fix we wanted to ship before the end of the sprint.

When I loaded up the first set of changes into my trusty Sublime Text, my eyes nearly fell out—this was a serious “can’t see the forest for the trees” type of situation:

The formatting for this Python code was…All. Over. The. Place.

There was no consistency whatsoever in how the code was indented, how braces were positioned… even the spacing between operators inside expressions was seemingly randomized:

#the  worst code ever .
value +=10*  othervalue

Ugh.

It just seemed so sloppy! And the inconsistent formatting made it really hard to see what the code did, what the intention behind it was.

It felt like my brain was 90% occupied with parsing out the code, instead of being able to focus on the bigger picture and to hunt for actual bugs.

I must’ve spent at least an hour cleaning up the formatting, before I was able to give any substantial feedback on these changes. It was the most tedious code review of my dev career.

My busywork was of little value to the company, too:

They paid me a software engineer’s salary for nudging around braces and juggling whitespace…

That same day I pulled the whole team together to discuss the mandatory use of a code style checker before code reviews.

And guess what? It worked out great.

Most developers on the team were using Sublime Text so we all installed the SublimeLinter package. It’s the most popular code linting framework for Sublime Text and I like it for its focus, simplicity, and performance.

A code linter is a program that analyses your source code for potential errors. Code linters are great at finding “mechanical” issues like syntax errors, structural problems, such as the use of undefined variables, and also best practice or code style violations.

SublimeLinter let’s you integrate code linting feedback into your editing environment. Setting up SublimeLinter gives you immediate feedback on your code right when you type it:

When you install SublimeLinter it doesn’t actually include any linter engines. It’s more like a “meta linter” that lets you integrate various command-line linter binaries like Flake8 (Python) or JSHint (JavaScript) under one roof.

The linter binaries do the real work. And that way, SublimeLinter can support more than just one programming language. If you’re doing any kind of full-stack web development, for example, you could install code linters for JavaScript, CSS, Ruby, Go, and Python.

SublimeLinter will then pick the right code linter to run on each file you’re editing. Any errors or warnings found by these separate linters would all be integrated with the same look and feel into your Sublime Text editor window by SublimeLinter.

And because we were using command line tools through SublimeLinter we were able to set up the same set of code style checks on our CI build server very easily. That way no badly formatted code could slip through the cracks ever again.

It made the whole team more productive. And it was great for morale: No more time wasted on nudging braces or juggling whitespace 🙂

Additional Resources & Links

Here are a couple of extra links to help you get set up with SublimeLinter. I listed the most common linter binaries and linter plugins so you can get started right away:

• My Sublime Python course
• SublimeLinter docs
• All official linter plugins for SublimeLinter
• JavaScript: JSHint, Flow, JSL, JSXHint, JSCS
• Ruby: Ruby (built-in), Rubocop
• Python: Flake8, Pylint, Pep8, Pyflakes
• PHP: PHP, PHPLint
• Go: GoLint, GoType
• Lua: LuaCheck, Lua (built-in)
• Haskell: Ghc
• C++: CppCheck, CppLint
• CSS: CssLint
• HTML: HTMLTidy
• Java: Java (built-in)
• Plaintext: Proselint