I don\'t understand the following code:
ruby-1.9.1-p378 > puts \"nil is false\" unless nil
nil is false
=> nil
ruby-1.9.1-p378 > puts \"nil isn\'t
Ruby considers that false
and nil
are the only two "falsy" values, while everything else is "truthy". This is by definition and can not be modified (at least in MRI). This definition is used for all builtin operators like if
, unless
, while
, until
, cond ? if_truthy : if_falsey
, ||
, &&
, ...
Writing foo == bar
will always call the ==
method on foo
with bar
as an argument. By default, nil
, false
, true
and all other immediates like symbols, etc..., are only equal to themselves. This could be changed, though:
def nil.==(bar)
super || bar == false
end
puts "nil == false" if nil == false # => "nil == false"
In Ruby 1.9, you can also redefine the operator !
, so unless foo
is not necessarily the same as if !foo
or the contrary of if foo
:
def true.!
true
end
puts "True?" if true # => "True?"
puts "or not?" if !true # => "or not?"
Not that anybody would recommend doing anything like this...
In some languages, you can only use booleans in "if" statements. Trying to check whether a string or a number is true or false is just silly and meaningless so the language won't let you do it.
In Ruby, however, everything can be thought of as a boolean, even though it's really not. In fact, everything in Ruby is effectively true, except nil and false (IIRC). That doesn't mean that nil is actually EQUAL to false, just as it doesn't mean that the integer 45 is actually EQUAL to true. They're different, separate, things. But if you're going to treat nil like a boolean (i.e. use it in an if or unless) then it's as though it is false.
nil
is not equal to false by comparison over ==
because their semantic content is different (nil
is no information, false is a boolean value). However, if you try to evaluate nil
in a boolean context, it will be considered False, mostly for convenience' sake and idiomatic compatibility with other languages.
nil == a <=> nil != a <=> false
is going to be the case for virtually any value of a except for nil.
So you can only tell that 'nil is not true' and 'nil is nil'. This is as far as ruby goes in its ways.