bbatsov/rubocop

A robust Ruby code analyzer, based on the community Ruby style guide.

Thomas Volkmar Worm

@tvw

What is a reasonable default for Metrics/AbcSize? As soon as a method contains a few conditions, rubocop starts complaining. In order to satisfy rubocop, I had to split a method into two methods, though the code belongs together and makes much more sense in one method.

Also Metrics/LineLength: I think VT100s are absolutely history. My lines are most of the time below 80, but when using a logger, a line could be longer and splitting the line just to satisfy rubocop makes the code looking very poor.

Thomas Thomassen

@thomthom

Hi all. I'm working on a RuboCop extension - to aid SketchUp developers using our Ruby API. I'm currently reviewing the severity property of the custom departments we are adding. I was wondering if there was some insight to what the various levels available means?
I've seen the list in the code, and in the docs: https://github.com/rubocop-hq/rubocop/blob/master/manual/configuration.md#severity
But I'm curious to when to use what. In particular, is "error" supposed to be used only when RuboCop encounter an internal error processing?

I've been using error for one department, this one have a set of cops that check for some hard requirement in order to get a SketchUp extension submitted to our Extension Warehouse (basically our app store for extensions to SketchUp)

But is this severity level misleading?

Found some older github issues where it seemed to be somewhat unclear exactly what severity should be applied to the different type of reporters: rubocop-hq/rubocop#633

Steve Whittaker

@swhitt

Casper Christensen

@casperc91_gitlab

Hi. I have tried to search in the docs and on the web for a Rubocop rule to use <= and < instead of > or >= in comparison, example:

# bad
a > 2
3 >= b

#good
2 < a
b <= 3
4 < c && c <= 7

Does anyone know if this exist or not? Can't seem to find it.

andre francois

@francoisa

I have installed rubocop 0.62.0 in my rails application, but when I run "bundle exec rake new_cop[dept/name]"
I get the error : "Don't know how to build task 'new_cop'"
How do I create a new cop?

Tejas Bubane

@tejasbubane

@francoisa That rake task is for running inside the rubocop project directory itself. For contributing new cop.

https://github.com/rubocop-hq/rubocop/blob/master/tasks/new_cop.rake This will not run elsewhere

Brandon Weaver

@baweaver

Bit of an out there idea I'd be interested in a sanity check on. Javascript has the idea of codemods, I kinda want to try and make the same thing in Ruby. Rubocop already has a lot of work done on top of parser, including the NodePattern syntax and autocorrect capabilities.

Is there a possibility of generalizing the autocorrect capability into a code rewrite gem?

This gem does some of that already - https://github.com/jonatas/fast

Well, rather is there interest in doing that? I'd be willing to pitch in on it if so, but I'd rather not jump on an idea that's either already been done or isn't of interest to the contributors

MrAtheist

@MrAtheist

Hey all, would someone be able to address this issue for me? rubocop-hq/rubocop#6567 For me specifically the exclude in Allcops doesnt work at all. Wondering if it's the inheritance thats messing it up.

Oppen

@Oppen

Hi, I have a possibly dumb question.

I've seen in the changelog people refer to their PR. How can I have the PR address before sending it?

I mean, I'm working in my changeset, I then upload it to a branch on my fork and create a PR, but then I can't go back and change it to use the proper id afterwards.

(I could in a different commit, after creating the PR, but I've seen it done in a single one)

John Vandenberg

@jayvdb

@Oppen after you create the PR, you can add another commit to your branch and it will be included in the PR. You can also 'amend' (replace) the initial commit

Oppen

@Oppen

OK. I thought of amend, but that always carried weird consequences when I forced pushed.

Thanks.

Oppen

@Oppen

Hi all. I'm working in a performance related cop. Should I send it to rubocop, rubocop-performance or both?

Simon

@stoivo

I have a open PR rubocop-hq/rubocop#6649, Could someone take a look at it and give me some feedback

Lewis Jones

@LewisARJones_twitter

Hi, I use Rubocop to write custom cops within my companies code bases to enforce our domain and help aid big refactors. I've got blocked trying to write cops that check if methods are called on instances of our domain object. Basically ruby classes we have created.
For example I want to write a cop that creates an offence if anyone sends the method expiry_date to and instance of Quote object object.
Can anyone point in the direction of an example of how to write a cop that can check the class of the object getting the method sent to it?

Thomas Thomassen

@thomthom

@LewisARJones_twitter The dynamic nature of Ruby makes this hard. I've tried to do something similar for rubocop-sketchup - but I found no good way to do so. In some cases I've scanned the parent scope to find the initialisation of the instance. But the moment the instance is passed on to other methods it gets really difficult to track. In the end I fell back on doing simple naive checks like mentioned. Or make educated guesses on variable names.
You can see this being a challenge even for IDEs like RubyMine or VSCode with Ruby+Solargraph extensions. They tend to make use of YARD to pick up types from doc comments to aid the static analysic. But they also have to analyse the whole project to have chance. If you have good YARD comments you could try to leverage that in your custom cops.

Jônatas Davi Paganini

@jonatas

@LewisARJones_twitter any reason to not deprecate the usage of expiry_date in the Quote object? because it seems more reasonable than try to create a linter for it. Make the interface fail or warn developers about it. When they run tests they will know the api is no longer going to support it.

Lewis Jones

@LewisARJones_twitter

Thanks for your help @jonatas and @thomthom

Mailo Světel

@roolo

Hey guys, I just discovered AllCops are being rewritten by children configs, what is the right way to do it?

https://gitlab.com/lipoqil/rubocop-mailo/blob/459031ee5e101aa471dca90b202836f8d37025ca/rubocop.yaml#L1-7 If I use AllCops - Exclude in file which inherits from this one, the configs from gem are ignored 🥺

Rubocop 0.65.0

Mailo Světel

@roolo

https://github.com/rubocop-hq/rubocop/blob/master/manual/configuration.md#merging-arrays-using-inherit_mode FTW

Philipp Hansch

@phansch

Hi, is there some way to write a cop that detects usage of :[] on Hashes and suggest fetch instead? Specifically, how could I limit this to only work on Hash?

I suppose this is more difficult as I thought because of Ruby's dynamic nature..

i.e, I'm currently getting false positives on Dir[pat], some_array[x], etc

Jonas Arvidsson

@jonas054

@phansch The short answer is "No". You'd have to add functionality in RuboCop to guess the type of values, similar to what "intellisense"/auto-completion does in IDEs.

Colin Swenson-Healey

@chozandrias76

Has anyone else recently updated rubocop, added the rubocop-performance gem and found that rubocop no longer inspects any of the files you had it inspecting before?

It is inspecting only 2 files now when I run rubocop and I don't know which those are or why anything has changed

Lee-Jon

@lee-jon

Hey @chozandrias76 did you solve the issue? I have the same issue jumping from 47 -> 75

Lee-Jon

@lee-jon

I solved it by deleting the yml file and rewriting it line by line.

Vincent

@vincent-psarga

Hi all,
I'm having some difficulties finding a good solution to my current issue. In the project I work on (cucumber-ruby), I'm getting Rubocop errors on some string formatting. We have enforced annotated style for string formatting:

Style/FormatStringToken:
  EnforcedStyle: annotated

but we also use formatting to specify number of decimals to keep in stone functions, eg:

def format_duration(seconds)
  m, s = seconds.divmod(60)
  "#{m}m#{format('%.3f', s)}s"
end

And of course, rubocop complains that this is not annotated style:

Style/FormatStringToken: Prefer annotated tokens (like %<foo>s) over unannotated tokens (like %s).
        @io.print format_string(format('%.7f', stepdef_key.mean_duration), :skipped) + ' ' unless config.dry_run?

How to solve this kind of problem ? (hopefully I'm on the right channel :) )

Vincent

@vincent-psarga

ok, well after bit more research, that was pretty simple in fact: format('%<duration>.7f', duration: stepdef_key.mean_duration). I did not know the numbers formatting worked as the string ones :)

CJ Avilla

@w1zeman1p

Anyone happen to have a link handy for where I might learn about how to initialize and run rubocop directly in Ruby? I'm trying to build something to check snippets of ruby in some documentation so the source file is a view template not a ruby file and contains other languages. I want to pass a string in Ruby to some actor and have it return cops or some such.

For javascript, I'm doing something like

Eslintrb.lint(raw_code, :eslintrc)

I'll mostly be checking for Style cops

seems like SO suggestions are to write to file, then just use open3 https://stackoverflow.com/questions/58779791/programmatically-integrate-rubocop-formatter-with-stdin-stdout

halo

@halo

Hi! I have a question.

I have a rubocop.yml on Github, which I inherit_from in all my projects. Some projects are pure Ruby, some Rails, and some use RSpec.

It is fine to require: rubocop-rails in a Rails project, naturally.

But in a pure Ruby project I would not want to demand that the rubocop-rails gem is installed (neither should other developers have to install that gem, nor is it needed on CI).

So, how would I organize my shared rubocop config?

I cannot just specify any Rails/ cop in it, because that raises an error here

Similarly, the RSpec cops should only run in projects that use RSpec.

My idea is to have a rubocop.ruby.yml and a rubocop.rails.yml etc. on Github and inherit only the ones I need in my project.

Is that the recommended solution? I thought it was nice to have only one yml file with all cops, but I don't see how I could achieve that without the disadvantages I tried to describe.

Thank you very much.

Max Schwenk

@maschwenk

curious why ruby-parse considers a block that has multiple lines to be nested within a begin. I know this isn't a rubocop question, just trying to figure out how one would generically say:

this method call within a block of this type

is it easier to do that or

this block with this method call within

create_table :a_table do |z|
  z.goober :name
  z.string :col
end

create_table :a_table do |z|
  z.goober :name
end

parsed

(begin
  (block
    (send nil :create_table
      (sym :a_table))
    (args
      (procarg0
        (arg :z)))
    (begin
      (send
        (lvar :z) :goober
        (sym :name))
      (send
        (lvar :z) :string
        (sym :col))))
  (block
    (send nil :create_table
      (sym :a_table))
    (args
      (procarg0
        (arg :z)))
    (send
      (lvar :z) :goober
      (sym :name))))

Where communities thrive

People

Repo info

Activity