as far as I recall the definition of a range,
"" is part of the range ("".."")
I verified with irb and
("".."") === ""
returns true
"" <=> "" returns 0 as expected,
"".succ => "" which is not quite conclusive for me, but see below
anyway it seems inconsistent that
("".."").to_a.include?("") returns false
while
("".."") === "" returns true
No I have mimicked the bahvior with
class NilClass
def succ; nil; end
def <=>(other); 0; end
end
Guess what
(nil..nil).to_a returns ??
[nil] as expected
the behavior you described clearly seems to be a bug.
Regards
Robert
···
On 3/29/06, vsv <gm.vlkv@gmail.com> wrote:
Could someone explain me, please, why Range with one element is
iterated inconsistently when this element is empty string:
$ irb --simple-prompt
>> ('a'..'a').to_a
=> ["a"]
>> (1..1).to_a
=> [1]
## but:
>> ("".."").to_a
=>
>> ("".."").each{|s| puts 'Inside'}
=> ""..""
Why to_a returns empty and each block never executed?
Is it bug or feature?
Any reasonable explanation?
thanks
Sergey
--
Deux choses sont infinies : l'univers et la bêtise humaine ; en ce qui
concerne l'univers, je n'en ai pas acquis la certitude absolue.
I initially thought it was because "".succ == "" which is not correct,
however I created a custom class that returns itself when succ is
called:
class BadSucc
include Comparable
attr_accessor :value
def <=>(other)
self.value <=> other.value
end
def succ
self
end
end
But this did not seem to work
irb(main):002:0> b = BadSucc.new
=> #<BadSucc:0x2dc49a8>
irb(main):003:0> b.value = 1
=> 1
irb(main):008:0> (b..b).to_a
=> [#<BadSucc:0x2dc49a8 @value=1>]
So it seems like it might be a bug to me or there is some other reason
why ("".."").to_a returns empty while (b..b).to_a doesn't.