Autocorrect for Performance/BigDecimalWithNumericArgument cop converts an existing code to an invalid code

[y-yagi]

Expected behavior

Autocorrect for Performance/BigDecimalWithNumericArgument cop converts an existing code to a valid code

Actual behavior

Autocorrect for Performance/BigDecimalWithNumericArgument cop converts an existing code to an invalid code.

Steps to reproduce the problem

Sinc #463, autocorrect for Performance/BigDecimalWithNumericArgument cop converts string argument of BigDecimal to numeric. But, if a value is float, that generates an invalid code.

For example, if we have BigDecimal("0.8"), autocorrect converts it to BigDecimal(0.8), but this is an invalid code.

require 'bigdecimal'

BigDecimal(0.8)
(irb):3:in `BigDecimal': can't omit precision for a Float. (ArgumentError)

BigDecimal(0.8)
           ^^^
        from (irb):3:in `<main>'
        from <internal:kernel>:187:in `loop'
        from /home/yyagi/.rbenv/versions/3.3.5/lib/ruby/gems/3.3.0/gems/irb-1.14.0/exe/irb:9:in `<top (required)>'
        from /home/yyagi/.rbenv/versions/3.3.5/bin/irb:25:in `load'
        from /home/yyagi/.rbenv/versions/3.3.5/bin/irb:25:in `<main>' 

RuboCop version

Include the output of rubocop -V or bundle exec rubocop -V if using Bundler. Here's an example:
You can see extension cop versions (e.g. rubocop-performance, rubocop-rails, and others) output by rubocop -V,
include them as well. Here's an example:

$ bundle exec rubocop -V
1.66.1 (using Parser 3.3.5.0, rubocop-ast 1.32.3, running on ruby 3.3.3) [x86_64-linux]
  - rubocop-factory_bot 2.26.1
  - rubocop-performance 1.22.0
  - rubocop-rails 2.26.1
  - rubocop-rspec 3.0.5
  - rubocop-rspec_rails 2.30.0

Additional Information

I understand that change is caused by #454. But, if a value is a float, I think BigDecimal("0.8") is still faster than BigDecimal(0.8, 0).

# frozen_string_literal: true

require "bundler/inline"

gemfile(true) do
  source "https://rubygems.org"

  gem 'benchmark-ips'
  gem 'bigdecimal'
end

require 'benchmark/ips'
require 'bigdecimal'
require 'bigdecimal/util'

Benchmark.ips do |x|
  x.report('float string new')  { BigDecimal("0.8") }
  x.report('float new')         { BigDecimal(0.8, 0) }
  x.compare!
end

$ ruby bug_report_rubocop.rb  

Fetching gem metadata from https://rubygems.org/.
Resolving dependencies...
ruby 3.3.5 (2024-09-03 revision ef084cc8f4) [x86_64-linux]
Warming up --------------------------------------
    float string new   409.997k i/100ms
           float new   272.709k i/100ms
Calculating -------------------------------------
    float string new      4.128M (± 1.4%) i/s  (242.27 ns/i) -     20.910M in   5.066891s
           float new      2.716M (± 2.9%) i/s  (368.25 ns/i) -     13.635M in   5.025915s

Comparison:
    float string new:  4127550.6 i/s
           float new:  2715531.7 i/s - 1.52x  slower

[Earlopain]

This should be fixed with #469, can you please update? It's already been released.

[y-yagi]

I missed it, thanks!