Question

Given a method:

def x(*a)

Can calling it like so:

 x('foo', y: 'bar', z: 'baz')

Ever differ from:

@yz = { y: 'bar', z: 'baz' }
x('foo', @yz)

I’m asking because I’ve found the I18n.translate method, defined as:

def translate(*args)
  options  = args.last.is_a?(Hash) ? args.pop : {}
  key      = args.shift
  backend  = config.backend
  locale   = options.delete(:locale) || config.locale
  handling = options.delete(:throw) && :throw || options.delete(:raise) && :raise # TODO deprecate :raise

  raise I18n::ArgumentError if key.is_a?(String) && key.empty?

  result = catch(:exception) do
    if key.is_a?(Array)
      key.map { |k| backend.translate(locale, k, options) }
    else
      backend.translate(locale, key, options)
    end
  end
  result.is_a?(MissingTranslation) ? handle_exception(handling, result, locale, key, options) : result
end

…behaves differently when called using the different approaches above. When passing in a variable instead of the inline hash, then locale is always set to config.locale and the built-in locale fallbacks mechanism is prevented from running (regardless of whether :locale is present in the hash variable.) Yet, in IRB tests, I can’t find a way for #x to tell the difference.

I’ve worked around this for the moment by copy-pasting the hash contents into each #translate call in the app affected, but sure would by nice to know what’s up here.

Answer 1

no, they are identical:

def x(*a)
  puts a.inspect
end

x('foo', y: 'bar', z: 'baz')
 => ["foo", {:y=>"bar", :z=>"baz"}]

@yz = { y: 'bar', z: 'baz' }

x('foo', @yz)
 => ["foo", {:y=>"bar", :z=>"baz"}]

Please note that the syntax you’re using will only work in Ruby 1.9 , not in 1.8