If this language feature is annoying to you, you are the problem. You đź‘Źare đź‘Ź the đź‘Ź reason đź‘Ź it đź‘Ź exists.
I worked in places where the developers loaded their code full of unused variables and dead code. It costs a lot of time reasoning about it during pull request and it costs a lot of time arguing with coworkers who swear that they’re going to need that code in there next week (they never need that code).
This is a very attractive feature for a programming language in my opinion.
PS: I’m still denying your pull request if you try to comment the code instead.
❗️EDIT: A lot of y’all have never been to programming hell and it shows. 🪖 I’m telling you, I’ve fixed bayonets in the trenches of dynamically typed Python, I’ve braved the rice paddies of CICD YAML mines, I’ve queried alongside SQL Team Six; I’ve seen things in production, things you’ll probably never see… things you should never see. It’s easy to be against an opinionated compiler having such a feature, but when you watch a prod deployment blow up on a Friday afternoon without an easy option to rollback AND hours later you find the bug after you were stalled by dead code, it changes you. Then… then you start to appreciate opinionated features like this one. 🫡
Warn in dev, enforce stuff like this in CI and block PRs that don't pass. Go is just being silly here, which is not surprising given that Rob Pike said
Syntax highlighting is juvenile. When I was a child, I was taught arithmetic using colored rods. I grew up and today I use monochromatic numerals.
Yeah, insisting on things like a variable being used will result in people using work arounds. It won't result in people not doing it.
Then, because people trust the language to police this rule, the work-arounds and debug code will get committed.
func main() {
test := true
}
Oops, golang doesn't like that.
func main() {
test := true
_ = test
}
Perfectly cromulent code.
If they really wanted to avoid people having unused variables, they should have used a naming convention. Any variable not prefixed by "_" or "_debug_" or whatever has to be used, for example. Then block any code being checked in that still contains those markers.
reading my code after being up for 18 hours and having my eyes glaze over trying to parse the structure of my monochromatic code but then I remember Rob Pike said syntax highlighting is juvenile so I throw my head against that wall for another 3 hours
I agree that golang is being dumb when you don't even have the option to tell it that this is a testing env or something. But the thing about syntax highlighting is not the same. One is about handholding the developer so much that it makes it even more difficult to develop, and the other is a completely optional feature that is so uselful and non intrusive that even wizardly editors like emacs use it.
What's a situation where you need an unused variable? I'm onboard with go and goland being a bit aggressive with this type of thing, but I can't think of the case where I need to be able to commit an unused variable.
You probably wouldn't be committing this, unless you're backing up a heavily WIP branch. The issue is that if you're developing locally and need to make a temporary change, you might comment something out, which then requires commenting another now-unused variable, which then requires commenting out yet another variable, and so on. Go isn't helping you here, it's wasting your time for no good reason. Just emit a warning and allow CI to be configured to reject warnings.
I will need it two minutes tops. If I don't use it by then, I'll delete it, especially if it gives a warning like Rust does. But this? It just gets in the way.
I have a use case in Powershell: my company has a number of scripts that are minimally but importantly customized per-location, and I have an otherwise unused "SiteId" variable where I keep the location name for that specific script for a quick sanity check when I'm looking them over for any reason. Not necessary, but useful to me. Probably wouldn't do the same thing in a compiled program, but I can at least see where someone might want something similar.
That's a problem with your workplace, not the language nor OP.
You could have a build setting for personal development where unused variables are not checked, and then a build setting for your CI system that will look for them. It gives you freedom to develop the way you want without being annoyed when you remove something just to test something, but will not merge your PR unless the stricter rules are met.
I concur, it is a problem with that workplace. (In this case, OP is just sharing a funny meme. I wouldn’t suggest this meme means they’re a problem. I could have made this meme and I love the feature.)
Developing on a team at a company is like the “Wild West.” What’s considered to be acceptable will not only vary from workplace to workplace, but it can also fluctuate as developers and managers come and Go. Each of them have their own unique personality with their own outlook on what “quality” code looks like. (And many of them do not care about code quality whatsoever. They just need to survive 1-2 years there, make management happy with speedy deliveries, and then they can move on to the next company with a 30% pay bump.)
Having experienced working with developers who frequently filled with code base with unused code while having no control over who will leave or join as a contributor to the code base, I think features like this make for a more sane development experience when you’re developing with a team of seemingly random people that you never personally invited to contribute to the code base.
will not merge your PR unless the stricter rules are met.
This doesn’t fly when you work in big corporate and the boss doesn’t care about the code meeting stricter rules. “A working prototype? No it’s not- that’s an MVP!Deploy it to production now and move onto the next project!”
Why in the world would you want to develop something that doesn't follow the coding rules required by your org, just so you can go back and fix everything before submitting a PR? That's just extra work.
Because you want to know if the first half of the code works at all before you write the whole second half.
Finding all the bits that will be used by the second half and changing the declarations to just expressions is a bunch of extra work. As is adding placeholder code to use the declared variables.
I'm having a hard time envisioning a situation where testing my code requires a bunch of unused variables. Just don't declare the variables until you've started writing the code that uses them...
Most of the time you don't write the code, you change it.
I had tons of situations where I wanted to test deleting a code block which just happened to use an imported library, which the compiler is now complaining about because it's no longer being used.
If that’s the problem, then I would just use something like goimports to auto fix the imports every time I hit save. I never even see those errors so they don’t bother me.
That's what warnings are for. The jokes about programmers ignoring warnings are outdated - we live in an age where CIs run linters and style checkers on pull requests, there is no reason for a CI to not automatically reject code that builds with warnings.
I mean, yeah that kind of stuff absolutely should not be in production. However, it's easy to see how it could be annoying while testing something while working on it. It being annoying doesn't make it a bad feature, just as finding it annoying doesn't make you a problem imo.
It is VERY annoying when you're hunting for a bug. For example, when you comment out a function call to debug an issue, the compiler might suddenly complains about unused variables, so you'll need to comment them out as well. Repeat several times and you'll start having an urge to smack the monitor.
It costs a lot of time reasoning about it during pull request and it costs a lot of time arguing with coworkers who swear that they’re going to need that code in there next week (they never need that code).
You should go to your team leader and ask them to enforce a coding standard. I agree with other commenters that said this should be a warning instead of an error.
I was working for a team that did quality control on the code of an entire financial group and it's still amazing to me the shit we let through.
I feel annoyed even having compiler warnings in my code and here we were downgrading errors into warnings so the code would go through, or adding rules exceptions for a program so the team responsible could push a hotfix to prod...
It's all shit. All the way down.