Project

General

Profile

Actions

Bug #13781

closed

Should the safe navigation operator invoke `nil?`

Added by ioquatix (Samuel Williams) almost 5 years ago. Updated over 4 years ago.

Status:
Closed
Priority:
Normal
Assignee:
-
Target version:
-
[ruby-core:82236]

Description

In the following example:

class Later < BasicObject
	def initialize(&block)
		raise ::ArgumentError, "Block required" unless block
		
		if block.arity > 0
			raise ::ArgumentError, "Cannot store a promise that requires an argument"
		end
		@block  = block
	end
	
	def __resolve__
		@result ||= @block.call
	end
	
	def nil?
		__resolve__.nil?
	end
	
	def respond_to?(name)
	  __resolve__.respond_to?(name)
	end
	
	def method_missing(name, *args, &block)
		__resolve__.__send__(name, *args, &block)
	end
end

Person = Struct.new(:name, :age)

person = Later.new do
  nil # Person.new("Froob", 200)
end

puts person.nil?
puts person&.name

The code fails because person is a proxy object.

If safe navigation operator invoked nil? it should work. But, it's not clear exactly how the implementation should behave, or whether it's possible to implement this style of proxy.

Updated by marcandre (Marc-Andre Lafortune) almost 5 years ago

Short answer is "no".

Longer answer is:

  • is there an actual use case? I very much doubt there is one
  • BasicObject does not respond to nil?, so the safe operator would not work in that case?
  • other Ruby conditional (like if foo or foo ? bar : baz) do not call nil? or ==(nil), they simply do a straight comparison with nil and false. That's the way it should be for the safe navigation operator too.

Updated by shevegen (Robert A. Heiler) almost 5 years ago

is there an actual use case? I very much doubt there is one

This is what some filed issues appear to be - a mostly theoretical
view that does not appear to be likely to emerge.

When Hiroshi filed the request, I do not think that he had any
"proxy" object in mind - it was simply to avoid compound methods
check e. g:

if u && u.profile && u.profile.thumbnails && u.profiles.thumbnails.large

versus Activerecord

if u.try!(:profile).try!(:thumbnails).try!(:large)

To be honest, I actually find the .try! variant more readable than the
&. variant but this is not the topic of course (not sure why activerecord
uses the '!' there, it also makes the chain ugly, in my opinion; perhaps
I am way too picky).

https://bugs.ruby-lang.org/issues/11537

Actions #3

Updated by marcandre (Marc-Andre Lafortune) over 4 years ago

  • Status changed from Open to Closed
Actions

Also available in: Atom PDF