# Function inside a function

#1

I have a question about the timing of function evaluations in this case:

def function_a(x = 1)
x**2 + x
end

def function_b(any_function)
y = 4 + 5 #this would actually be more complex
function_a(y)
end

puts function_b(function_a)

I found that function_a(x = 1) is evaluated first, then
function_b(function_a). I found that 'def function_a(x = nil) returns an
error. Why does function_a need to be evaluated before it is called as
an argument in function_b?

PS - the reason I am writing things this way is to make function_b
accept any function (within reason).

#2

On Wed, Jan 28, 2009 at 4:49 PM, Jason L.

puts function_b(function_a)

I found that function_a(x = 1) is evaluated first, then
function_b(function_a). I found that 'def function_a(x = nil) returns an
error. Why does function_a need to be evaluated before it is called as
an argument in function_b?

The problem here is that when you write function_a, this is actually
calling function_a,which as you are not supplying any parameter uses
the default value of 1 you defined.

PS - the reason I am writing things this way is to make function_b
accept any function (within reason).

For sure there are many ways to achieve this, but a common way is to
use blocks and lambdas:

function_a = lambda {|x| x**2 + x}

def function_b
y = 4 + 5 #this would actually be more complex
yield y
end

or more explicitly:

def function_b (&blk)
y = 4 + 5
blk.call(y) # or blk[y]
end

puts function_b(&function_a)

Hope this helps,

Jesus.

#3

In ruby, function_a is a method call (with no arguments), and as a
call, it gets evaluated to whatever that function returns. If you
wanted to grab the function as a “pure function”, you’d use lambdas or
Procs like Jesús pointed.

In other words:

function_a = Proc.new {|x| …do something with x… }

returns a functional object. You can then call the method with
function_a.call (aliased [])

see http://www.ruby-doc.org/core/classes/Proc.html for details on Proc
objects

Greetings

On Jan 28, 9:49 am, Jason L. removed_email_address@domain.invalid

#4

Ruben M. wrote:

In ruby, function_a is a method call (with no arguments), and as a
call, it gets evaluated to whatever that function returns. If you
wanted to grab the function as a “pure function”, you’d use lambdas or
Procs like Jesï¿½s pointed.

In other words:

function_a = Proc.new {|x| …do something with x… }

returns a functional object. You can then call the method with
function_a.call (aliased [])

Or, if you really want to do this with def’d methods, you can use

obj.method(:function_a)

which will turn an existing method (bound to the instance ‘obj’) into a
Proc-like thing that you can pass around and call later.

def function_a(x = 1)
x**2 + x
end

def function_b(any_function)
y = 4 + 5
any_function.call(y) # or: any_function[y]
end

puts function_b(method(:function_a))

#5

Perfect. Thank you!