Performance/ArraySemiInfiniteRangeSlice breaks code on Strings

[brushbox]

When running rubocop-performance over our code we noticed some autocorrections broke code.
The ArraySemiInfiniteRangeSlice rule was being applied to Strings but there is no String#drop method, leading to broken code.

See the minimal .rubocop.yml and sample.rb below for details.

Note: this also applies to non-obvious strings (returned from a method) and not just string literals.

The only workaround was to disable the rule.

---

Expected behavior

$ rubocop -a sample.rb 
Inspecting 1 file
.

1 file inspected, no offenses detected

And running sample.rb produces:

$ ruby sample.rb
ous string

Actual behavior

$ rubocop -a sample.rb
Inspecting 1 file
C

Offenses:

sample.rb:3:6: C: [Corrected] Performance/ArraySemiInfiniteRangeSlice: Use drop instead of [] with semi-infinite range.
puts 'an obvious string'[7..]
     ^^^^^^^^^^^^^^^^^^^^^^^^

1 file inspected, 1 offense detected, 1 offense corrected

And running sample.rb produces:

$ ruby sample.rb
Traceback (most recent call last):
sample.rb:3:in `<main>': undefined method `drop' for "an obvious string":String (NoMethodError)

Steps to reproduce the problem

1. Given the .rubocop.yml file:

   require:
     - rubocop-performance
   
   AllCops:
     NewCops: enable
     TargetRubyVersion: 2.7

2. And the sample.rb file:

   # frozen_string_literal: true
   
   puts 'an obvious string'[7...]

3. Run rubocop -a

• the "violations" will be auto-corrected.

4. Run ruby sample.rb

• the code will break as there is no String#drop method.

RuboCop version

$ rubocop -V
1.3.1 (using Parser 2.7.2.0, rubocop-ast 1.1.1, running on ruby 2.7.1 x86_64-darwin19)
  - rubocop-performance 1.9.0

[tejasbubane]

Although we could change the cop to skip literal strings, it still remains unsafe for variables which we cannot determine statically if they are arrays or strings.

I have raised a PR to mark this cop as unsafe.

[brushbox]

Thank you for looking into this!

…

On Wed, Nov 25, 2020 at 12:43 PM Koichi ITO ***@***.***> wrote: Closed #197 <#197> via #199 <#199>. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#197 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAADC7UJYXTDV7BWGBOAGD3SRROLTANCNFSM4T63I4NA> .