Extra rubocop offenses when running rubocop on ruby/spec

[eregon]

From rubocop/rubocop#12724 (comment)
I noticed there are 3 more offenses when running with ParserEngine: parser_prism:

core/objectspace/define_finalizer_spec.rb:166:1: W: Lint/ImplicitStringConcatenation: Combine  and  into a single string literal, rather than using implicit string concatenation.
    RUBY

language/fixtures/bytes_magic_comment.rb:2:26: F: Lint/Syntax: invalid multibyte character 0xA0
(Using Ruby 3.3 parser; configure using TargetRubyVersion parameter, under AllCops)
$magic_comment_result = '你好'.bytes.inspect
                         ^^
language/fixtures/bytes_magic_comment.rb:2:27: F: Lint/Syntax: invalid multibyte character 0xBD
(Using Ruby 3.3 parser; configure using TargetRubyVersion parameter, under AllCops)
$magic_comment_result = '你好'.bytes.inspect
                           ^^

I guess there are due to some differences between parser gem and Prism::Translation::Parser.

[kddnewton]

Something very odd is happening with Parser::Source::Buffer. We're passing in a big5-encoded string, and it looks like it's transcoding it to utf-8. So then when it hits the multi-byte characters in the source, it has reencoded them as utf-8 which is invalid in big5. I think we need to tell it to please stop doing that.

[CKolkey]

I'm not sure if this is the right place to be tracking these, so apologies if that's the case, but I also found that the following cops were being triggered incorrectly when using implicit keyword values:

# Lint/UnusedMethodArgument

def foo(bar)
  Baz.new(bar:)
end

# Lint/UnusedBlockArgument

[1, 2, 3].each do |n|
  Foo.new(n:)
end

That said, it found these (incorrect) issues very quickly, so, nice work :D

[kddnewton]

Thanks @CKolkey I've opened #2563 to fix this. If you find more, feel free to open new tickets.

[andyw8]

Here's a minimal example that results in a false positive for Layout/ExtraSpacing due to the emoji:

# 👎
<<-RUBY.squish
  a
  b
RUBY