Detecting when a method is chained - ruby

I am looking to do some method chaining. I have the following code:
class MyClass
attr_accessor :handler
def do_a
puts 'i just did a'
self.handler = 'a'
self
end
def do_b_if_a
puts 'i just did b' if handler == 'a'
end
end
So the following works:
irb > test = MyClass.new
=> #<MyClass:0x007fa44ced9a70 #handler=nil>
irb > test.do_a
'i just did a'
irb > test.do_a.do_b_if_a
'i just did a'
'i just did b'
What I DONT want to work is when I call do_a the first time it sets the handler, which means now do_b_if_a can be called at any time. But I only want it to be called when it is chained with do_a, how do I do that?

In general, you don't want to care (and in most cases, you don't even get to know) if your methods are called in a certain way. That way lies madness. Magic call sequences make for a hell of a time debugging and testing.
What you could do, though...instead of having do_a return self, wrap a decorator around it that defines do_b, and return the decorator. At that point, your original MyClass can't do_b, cause it doesn't know how. But the thingie returned from do_a can.
(Now, you can still say like a = test.do_a and then a.do_b, but you can't really get around that without parsing the Ruby code yourself.)

I think you're tracking state in the wrong place. You'd be better off with something similar to ActiveRecord's query interface, for example:
class MyClass
attr_accessor :handler
def do_a
puts 'i just did a'
with_handler 'a'
end
def do_b_if_a
puts 'i just did b' if handler == 'a'
end
private
def with_handler(h)
o = dup
o.handler = h
o
end
end
That way you always have an instance of MyClass but you have a throw-away copy that keeps track of its history. This is similar to cHao's approach but it doesn't need an extra decorator class as MyClass can decorate itself.

As cHao said, you shouldn't try to method call patterns. But I can also think of another way to do this:
def do_a
puts "i just did a"
do_b_if_a(true)
end
def do_b_if_a(did_a=false)
puts "i just did b" if did_a
end
do_a # prints "i just did a"
do_b_if_a # does nothing
Sure, you can call do_b_if_a(true), but then it just makes it more flexible ;)

Doesn't get you all the way there, but you could define do_b_if_a as a singleton method that gets added when to_a is called:
class MyClass
def do_a
puts 'i just did a'
self.tap{|s| def s.do_b_if_a; puts 'i just did b' end}
end
end
You could take a look at Kernel.caller if you want to take certain actions based on the call stack. However, you might get punched in the face by another developer if you don't have a good reason for doing so, lol.

Related

Naive aspect implementation in ruby

I am trying to make a simplistic implementation of AOP in ruby. I was able to implement before and after advices, I got stuck with around advice.
This is the target class that is going to be advised:
class MyClass
def method
puts "running method"
end
end
This is the Aspect class to instantiate objects capable of making advices:
class Aspect
def advise(class_name, method, type, &block)
class_name.send(:alias_method, :proceed, :method)
class_name.send(:define_method, :method) do
case type
when :before
yield
proceed
when :after
proceed
yield
when :around
yield(proceed) # * proceed is the old version of the method
end
end
end
end
(*) Yield should execute the block around MyClass#proceed on the current object when method is invoked.
Creating the target and the aspect:
mc = MyClass.new
a = Aspect.new()
Invoking the method without advising it:
puts mc.method
Advising MyClass#method with around:
a.advise(MyClass, :method, :around) do |proceed|
puts "First"
proceed # this is not working *
puts "Last"
end
puts mc.method
(*) I am not being able to pass something to identify the call of proceed, that is the invocation of the old method without the advice.
The output should be:
First
running method
Last
In Ruby, a method call looks like this:
receiver.method(arguments)
Or, you can leave off the receiver if the receiver is self.
So, to call a method named proceed on some receiver, you would write
receiver.proceed
However, in your implementation, you don't keep track of what the receiver should be, so since you don't know the receiver, you simply cannot call the method.
Note that there are lots of other problems with your approach as well. For example, if you advise multiple methods, you will alias them all to the same method, overwriting each other.
I believe there are two things going wrong here.
This section of code
when :around
yield(proceed) # * proceed is the old version of the method
end
Calls the block given to advise providing the output of the proceed method as an argument.
So your output probably looks something like:
running method
First
Last
This block
a.advise(MyClass, :method, :around) do |proceed|
puts "First"
proceed # this is not working *
puts "Last"
end
Just evaluates the argument given as proceed. If a method is given it does not call it. So taking problem 1 into consideration in your case the original definition of method (aliased to proceed) returns nil (output of return) which will be passed as the value to the proceed argument in the block when yielded. the block ends up evaluating to something like
puts "First"
nil
puts "Last"
mc.method is called.
To address the second part, you may want to consider using send. Because the inner workings of your aspect may not be known to your code that calls it. It may change over time, so what ever calls Aspect.advise shouldn't make assumptions that the original method will still be accessible. Instead, it should take an argument (the new method name) and send it to the object. Making the block passed to advise:
a.advise(MyClass, :method, :around) do |aliased_method_name|
puts "First"
send(aliased_method_name)
puts "Last"
end
And adjusting the around item added to your class when advise is called to the following:
when :around
yield(:proceed) # * proceed is the old version of the method
end
If you do both of these things, your around section will calls the provided block, using the symbol for the new alias for the overridden method.
N.B.: This approach won't work for methods that require any arguments.
This is what I did. In the definition of Aspect#advise now I use a Proc, like this:
when :around
yield Proc.new { proceed }
end
And when calling the method to advise MyClass#method with :around parameter I use this:
a.advise(MyClass, :method, :around) do |original|
puts "First"
original.call
puts "Last"
end
I got:
First
running method
Last
Here's the fixed version that will work for arguments, and avoid clobbering.
class Aspect
##count = 0
def self.advise(class_name, method, type=nil, &block)
old_method = :"__aspect_#{method}_#{##count += 1}"
class_name.send(:alias_method, old_method, method)
class_name.send(:define_method, method) do |*args, &callblock|
case type
when :before
yield
send(old_method, *args, &callblock)
when :after
send(old_method, *args, &callblock)
yield
when :around, nil
yield lambda {
send(old_method, *args, &callblock)
}
end
end
end
end
class Foo
def foo(what)
puts "Hello, #{what}!"
end
end
Aspect.advise(Foo, :foo) do |y|
puts "before around"
y.yield
puts "after around"
end
Aspect.advise(Foo, :foo, :before) do
puts "before"
end
Aspect.advise(Foo, :foo, :after) do
puts "after"
end
Foo.new.foo("world")
# before
# before around
# Hello, world!
# after around
# after

What is the point of using "send" instead of a normal method call?

as far as I understand 'send' method, this
some_object.some_method("im an argument")
is same as this
some_object.send :some_method, "im an argument"
So what is the point using 'send' method?
It can come in handy if you don't know in advance the name of the method, when you're doing metaprogramming for example, you can have the name of the method in a variable and pass it to the send method.
It can also be used to call private methods, although this particular usage is not considered to be a good practice by most Ruby developers.
class Test
private
def my_private_method
puts "Yay"
end
end
t = Test.new
t.my_private_method # Error
t.send :my_private_method #Ok
You can use public_send though to only be able to call public methods.
In addition to Intrepidd's use cases, it is convenient when you want to route different methods on the same receiver and/or arguments. If you have some_object, and want to do different things on it depending on what foo is, then without send, you need to write like:
case foo
when blah_blah then some_object.do_this(*some_arguments)
when whatever then some_object.do_that(*some_arguments)
...
end
but if you have send, you can write
next_method =
case foo
when blah_blah then :do_this
when whatever then :do_that
....
end
some_object.send(next_method, *some_arguments)
or
some_object.send(
case foo
when blah_blah then :do_this
when whatever then :do_that
....
end,
*some_arguments
)
or by using a hash, even this:
NextMethod = {blah_blah: :do_this, whatever: :do_that, ...}
some_object.send(NextMethod[:foo], *some_arguments)
In addition to everyone else's answers, a good use case would be for iterating through methods that contain an incrementing digit.
class Something
def attribute_0
"foo"
end
def attribute_1
"bar"
end
def attribute_2
"baz"
end
end
thing = Something.new
3.times do |x|
puts thing.send("attribute_#{x}")
end
#=> foo
# bar
# baz
This may seem trivial, but it's occasionally helped me keep my Rails code and templates DRY. It's a very specific case, but I think it's a valid one.
The summing briefly up what was already said by colleagues: send method is a syntax sugar for meta-programming. The example below demonstrates the case when native calls to methods are likely impossible:
class Validator
def name
'Mozart'
end
def location
'Salzburg'
end
end
v = Validator.new
'%name% was born in %location%'.gsub (/%(?<mthd>\w+)%/) do
# v.send :"#{Regexp.last_match[:mthd]}"
v.send Regexp.last_match[:mthd].to_sym
end
=> "Mozart was born in Salzburg"
I like this costruction
Object.get_const("Foo").send(:bar)

How to memoize a class method with META PROGRAMMING?

EDIT: To be clear. This is a question about how to do something with meta programming. It's not about memoizing. Clearly there are better ways to memoize. The relevant methods have "memoize" in them just to illustrate their purpose.
I'm just toying around with meta programming, so please don't answer use a
#foo instance variable.
I have the following that tries to memoize both an instance and a class method
by overwriting the method definition from the running method..
class Obj
class << self
def meta_me; self; end
def class_memoize
puts "hard core calculating ..."
abc = "huge calculation result"
raise "broken here with infinite loop"
define_class_method "class_memoize" do
puts abc
abc
end
class_memoize
end
def define_class_method name, &blk
meta_me.instance_eval do
define_method name, &blk
end
end
end
def instance_memoize
puts "hard core calculating ..."
abc = "huge calculation result"
self.class.meta_me.send :define_method, :instance_memoize do
puts abc
abc
end
instance_memoize
end
end
o = Obj.new
o.instance_memoize
# hard core calculating ...
# huge calculation result
o.instance_memoize
# huge calculation result
The instance version works, but the class version does not.
I've left in an attempt at the class version for reference.
Ethics aside, It's much easier than you think. Your main issue is you're using the wrong thing for your meta_me method. Try this:
class Object
def metaclass
class<<self;self;end
end
end
This is a fairly common monkeypatch when doing metaprogramming in Ruby. Now it's easy to reimplement a method, dynamically:
class Obj
def self.class_memoize
puts "calculating..."
abc = "result"
metaclass.send(:define_method, :class_memoize) do
puts abc
abc
end
class_memoize
end
def instance_memoize
puts "calculating..."
abc = "result"
metaclass.send(:define_method, :instance_memoize) do
puts abc
abc
end
instance_memoize
end
end
As you can see, once you have the metaclass object, redefining the methods done the same way whether it's a class method or an instance method. Note that if you don't want to sully the whole namespace with a metaclass method, it's probably better just to use the class<<self;self;end idiom whenever you want to refer to it. You can call methods directly on it, like this:
(class<<self;self;end).send(:define_method, :foo){|bar| bar*23}
Note that the parens aren't truly needed, it just helps to contain the messiness that is the bare metaclass reference :) Hope this helps.

calling a super class' method with a twist

Suppose I have two classes like so:
class Parent
def say
"I am a parent"
end
end
class Child < Parent
def say
"I am a child"
end
def super_say
#I want to call Parent.new#say method here
end
end
What are the options to do that? I thought of:
def super_say
self.superclass.new.say #obviously the most straight forward way, but inefficient
end
def super_say
m = self.superclass.instance_method(:say)
m = m.bind(self)
m.call
#this works, but it's quite verbose, is it even idiomatic?
end
I am looking for a way which doesn't involve aliasing Parent.new#say to something else, which would make it unique in the method lookup chain (Or is that actually the preferred way?).
Any suggestions?
I tend to prefer using an alias. (I'm not quite sure I understand your objection to it.)
Example:
class Child < Parent
alias :super_say :say
def say
"I am a child"
end
end
Gives:
irb(main):020:0> c = Child.new
=> #<Child:0x45be40c>
irb(main):021:0> c.super_say
=> "I am a parent"
Your second solution (the bind()) is the one i would go for. It is verbose because what you are doing is highly unusual, but if you really need to do it -- that solution seems fine to me.

Problem with accessing superclass methods in method redefinitions

I am having a bit trouble to understand when "super" can be called and when not. In the below example the super method leads to a no superclass error.
class Bacterium
def eats
puts "Nam"
end
end
class Bacterium
def eats
super # -> no superclass error
puts "Yam"
end
end
b = Bacterium.new
b.eats
But this works:
class Fixnum
def times
super # -> works
puts "done"
end
end
5.times { |i| puts i.to_s }
Is 5 not just also an instance of Fixnum. And am I not redefining an existing method like in the Bacterium example above?
No, not really. Fixnum inherits from Integer class, and you are in fact overriding Integer#times, so super works, as it calls implementation from the parent.
In order to achieve something similar when monkeypatching, you should alias method before redefining it, and there call it by alias.
class Bacterium
alias_method :eats_original, :eats
def eats
eats_original # -> "Nam"
puts "Yam"
end
end
Class reopening is not a form of inheritance and super is of no use there.
Just as Mladen said, and you can check that with Class#superclass:
irb> Fixnum.superclass
=> Integer
And does Integer implement #times?:
irb> Integer.instance_methods.grep /times/
=> [:times]
Yes it does.
So, in a simplified way, we can say, that super invokes the method you are in of a superclass. In your case the superclass of a Bacterium is Object, which doesn't implement #eats.
I said this is very simplified, because look at this example:
module One
def hi
" World" << super()
end
end
module Two
def hi
"Hello" << super()
end
end
class SayHi
def hi
"!!!"
end
end
h = SayHi.new
h.extend(One)
h.extend(Two)
puts h.hi
#=> Hello World!!
Don't take to serious what I wrote here, it is actually tip of the iceberg of the Ruby object model, which is important to understand (I am still learning it) - then you will get most, or all of those concepts.
Use some Google-fu for "Ruby object model"...

Resources