Whenever I talk to Rubyists I hear great things about method_added and method_missing. However, I got recently chided for using both in a project. The argument was that another library (such as rspec, for example) could overload the methods too and put the program in a strange state depending on which version of method_missing got called first.
I'm curious to know how often this happens. Is it really that dangerous to overload method_missing? Does anyone have real-world examples of woe arising from method_missing conflicts?
One thing to keep in mind that if you (re)define method_missing on a class, you're replacing any previous implementations of this method (from a Gem for example).
You can avoid this by creating a new class that inherits from the class you want to extend with method_missing.
However, this is normally not an issue as most Gems have their own classes.
The other thing to keep in mind is to always call super at the end of method_missing to not break the method invocation chain.
def method_missing(m, *args, &block)
# your code...
super
end
Maybe this Graphic of the Ruby Method Lookup Flow is also helpful.
Related
So, I'm currently learning about metaprogramming in Ruby and I want to fully understand what is happening behind the scenes.
I followed a tutorial where I included some of the methods in my own small project, an importer for CSV files and I have difficulties to wrap my hand around one of the methods used.
I know that the define_method method in Ruby exists to create methods "on the fly", which is great. Now, in the tutorial the method initialize to instantiate an object from a class is defined with this method, so basically it looks like this:
class Foo
def self.define_initialize(attributes)
define_method(:initialize) do |*args|
attributes.zip(args) do |attribute, value|
instance_variable_set("##{attribute}", value)
end
end
end
end
Next, in an initializer of the other class first this method is called with Foo.define_initialize(attributes), where attributes are the header row from the CSV file like ["attr_1", "attr_2", ...], so the *args are not provided yet.
Then in the next step a loop loops over the the data:
#foos = data[1..-1].map do |d|
Foo.new(*d)
end
So here the *d get passed as the *args to the initialize method respectively to the block.
So, is it right that when Foo.define_initialize gets called, the method is just "built" for later calls to the class?
So I theoretically get a class which now has this method like:
def initialize(*args)
... do stuff
end
Because otherwise, it had to throw an exception like "missing arguments" or something - so, in other words, it just defines the method like the name implies.
I hope that I made my question clear enough, cause as a Rails developer coming from the "Rails magic" I would really like to understand what is happening behind the scenes in some cases :).
Thanks for any helpful reply!
Short answer, yes, long answer:
First, let's start explaining in a really (REALLY) simple way, how metaprogramming works on Ruby. In Ruby, the definition of anything is never close, that means that you can add, update, or delete the behavior of anything (really, almost anything) at any moment. So, if you want to add a method to Object class, you are allowed, same for delete or update.
In your example, you are doing nothing more than update or create the initialize method of a given class. Note that initialize is not mandatory, because ruby builds a default "blank" one for you if you didn't create one. You may think, "what happens if the initialize method already exist?" and the answer is "nothing". I mean, ruby is going to rewrite the initialize method again, and new Foo.new calls are going to call the new initialize.
I don't know how to correctly phrase the title, I think the best way to explain this issue is just with code samples.
My goal
I want to define a meta method like this (in Rails 5):
class Post < ApplicationRecord
override_this_attribute_writer :some_attribute
end
The override_this_attribute_writer follows a common pattern, it overrides the original writer by doing some filtering on top of it. I find this way of overriding very convenient and clear.
First approach
module MyCommonModule
extend ActiveSupport::Concern
module ClassMethods
def override_this_attribute_writer(attribute_name)
alias_method :"#{attribute_name}_old=", :"#{attribute_name}="
define_method :"#{attribute_name}=" do |a_value|
# Do my stuff
send(:"#{attribute_name}_old=", a_value)
end
end
end
When doing this, I was getting an exception at the call of alias_method, because, apparently, the method I was trying to copy didn't exist (yet).
Second approach
module MyCommonModule
extend ActiveSupport::Concern
module ClassMethods
def override_this_attribute_writer(attribute_name)
define_method :"#{attribute_name}=" do |a_value|
# Do my stuff
send(:write_attribute, attribute_name, a_value)
end
end
end
I was expecting this not to work: if, when running the meta method, ActiveRecord hasn't created the attribute writer yet, this means that it will do it later and override the method that I just defined.
But surprisingly it worked! So I put my hands inside ActiveRecord (5.1.5) to find out more.
Dig into ActiveRecord 5.1.5
I wanted to ensure that what I did was safe and it wasn't just working by accident: I looked into the definition of method writer, and put binding.pry around the method.
This is the result of the experiment:
For attributes that I did not override,
This line is called
Then the method is defined inside this module eval call
Finally, the newly created writer method is correctly called when performing object.attribute=
For attributes that I DID override,
My own method is defined before anything else (when the ActiveRecord writers aren't there yet
Then ActiveRecord calls the same line that handles writer creation, as in the previous example
The method gets (apparently) correctly created by ActiveRecord, since it passes again by this point
But now, surprisingly, when calling object.attribute= my own method is still called in place of the ActiveRecord one
So, this is what I don't understand: if ActiveRecord seems to be overriding my method but it doesn't, what prevents it from doing it?
My questions
What in the end I need to know is whether the fix I have done is actually a good practice (and robust) or it's at risk and it might break if in the future we do upgrades.
If you think that my fix is dangerous, would you be able to suggest a different way to achieve the same goal?
Calling super is even more idiomatic:
module MyCommonModule
extend ActiveSupport::Concern
module ClassMethods
def override_this_attribute_writer(attribute_name)
define_method :"#{attribute_name}=" do |value|
# do some stuff
super value
end
end
end
end
I am new to Ruby and I saw methods defined like:
def method_one
puts "method 1"
end
class MyClass
method_one
def method_two
puts "method 2"
end
end
The way method_one is used reminds me of Python decorators.The output of
c = MyClass.new
c.method_two
is
method 1
method 2
I have been trying to search for more information about this syntax/language feature in the Ruby documentation on the web but I don't know what keywords to search for.
What this is thing called?
TL;DR
This code doesn't do what you think it does. Don't do stuff like this.
Ruby's Top-Level Object
Ruby lets you define methods outside a class. These methods exist on a top-level object, which you can (generally) treat as a sort of catch-all namespace. You can see various posts like What is the Ruby Top-Level? for more details, but you shouldn't really need to care.
In your original post, method_one is just a method defined in the top-level. It is therefore available to classes and methods nested within the top-level, such as MyClass.
Methods in Classes
Despite what you think, the following doesn't actually declare a :method_one class or instance method on MyClass:
class MyClass
method_one
def method_two; end
end
Instead, Ruby calls the top-level ::method_one during the definition of the class, but it never becomes a class method (e.g. MyClass::method_one) or an instance method (e.g. MyClass.new.method_one). There might be a few use cases for doing this (e.g. printing debugging information, test injection, etc.) but it's confusing, error-prone, and generally to be avoided unless you have a really strong use case for it.
Better Options
In general, when you see something like this outside an academic lesson, the programmer probably meant to do one of the following:
Extend a class.
Add a singleton method to a class.
Include a module in a class.
Set up a closure during class definition.
The last gets into murky areas of metaprogramming, at which point you should probably be looking at updating your class initializer, or passing Proc or lambda objects around instead. Ruby lets you do all sorts of weird and wonderful things, but that doesn't mean you should.
I think you're a little mislead; the output of:
c = MyClass.new
c.method_two
is
#<MyClass:0x007feda41acf18>
"method 2"
You're not going to see method one until the class is loaded or if you're in IRB you enter the last end statement.
I would suggest looking into ruby's initialize method.
This question already has answers here:
Abstract Method in Ruby
(7 answers)
Closed 7 years ago.
I'm writing a base class for an interface.. I want all the inherited classes to implement a few methods, is there a way to make them?
I have a payment_method_base class that I will be inheriting from.
I want all my payment_method classes to implement the method kind() that will return a string say 'credit_card' or 'braintree' or 'paypal' or 'amazon_pay'...
Is there a way to make sure that classes that inherit from payment_method_base are forced to implement the method kind()
Keeping in mind that the creator of a new payment_method class may not know about these requirements before they create the class.
In java these are called abstract methods. I'm wondering if ruby has something like that? https://github.com/shuber/defined/blob/master/lib/defined.rb
---- Evolution
I am wondering if there have been fixes to the language that allow for abstract methods.
---- partial answer
This answer might be the clue to adding the hooks I need.
Is there a hook similar to Class#inherited that's triggered only after a Ruby class definition?
This doesn't quite work the way I expected it would. The call doesn't seem to be done at the right time. Even so it is how many gems handle it.
https://github.com/shuber/defined/blob/master/lib/defined.rb
SIMPLES ANSWER THAT I CAN'T SUBMIT BECAUSE THIS IS A REPEAT OF A QUESTION THAT HAD NO REAL ANSWER.
def self.abstract(*methods_array)
##must_abstract ||= []
##must_abstract = Array(methods_array)
end
def self.inherited(child)
trace = TracePoint.new(:end) do |tp|
if tp.self == child #modules also trace end we only care about the class end
trace.disable
missing = ( Array(##must_abstract) - child.instance_methods(false) )
raise NotImplementedError, "#{child} must implement the following method(s) #{missing}" if missing.present?
end
end
trace.enable
end
abstract :foo
Ruby doesn't really implement anything like interfaces as it's against it's philosophy but if you really want to simulate interfaces you can use some gem, for example https://github.com/djberg96/interface.
There are no abstract methods in ruby.
I suppose you could implement a method like this, and I have seen it from time to time:
def some_method
raise "This method should be over-ridden by a sub-class"
end
If a sub-class neglects to implement that method -- it won't be caught at 'compile time', because that doesn't happen in ruby. But if someone tries to call that method, which hasn't been implemented, you'll get an exception raised.
Of course, if you hadn't put that method (that does nothign more than raise) in the "abstract" superclass, and it was called on a sub-class without the sub-class implementing it... you'd still get a NoMethodException raised anyway. So the "abstract" method implementation that does nothing more than raise doesn't add much. Maybe a documentation of intent.
I think the most ruby-like way to handle this would probably be to provide a test suite of some kind, that another developer implementing a sub-class can easily run against their sub-class. The test suite will complain if the sub-class hasn't implemented things it's supposed to -- and it can even go beyond method signatures, and make sure the semantics of the sub-class are as expected too (you can't do that with just method signatures in Java!). I haven't actually seen anyone do that though, but have often thought it would make sense.
In ruby, I understand that module functions can be made available without mixing in the module by using module_function as shown here. I can see how this is useful so you can use the function without mixing in the module.
module MyModule
def do_something
puts "hello world"
end
module_function :do_something
end
My question is though why you might want to have the function defined both of these ways.
Why not just have
def MyModule.do_something
OR
def do_something
In what kind of cases would it be useful to have the function available to be mixed in, or to be used as a static method?
Think of Enumerable.
This is the perfect example of when you need to include it in a module. If your class defines #each, you get a lot of goodness just by including a module (#map, #select, etc.). This is the only case when I use modules as mixins - when the module provides functionality in terms of a few methods, defined in the class you include the module it. I can argue that this should be the only case in general.
As for defining "static" methods, a better approach would be:
module MyModule
def self.do_something
end
end
You don't really need to call #module_function. I think it is just weird legacy stuff.
You can even do this:
module MyModule
extend self
def do_something
end
end
...but it won't work well if you also want to include the module somewhere. I suggest avoiding it until you learn the subtleties of the Ruby metaprogramming.
Finally, if you just do:
def do_something
end
...it will not end up as a global function, but as a private method on Object (there are no functions in Ruby, just methods). There are two downsides. First, you don't have namespacing - if you define another function with the same name, it's the one that gets evaluated later that you get. Second, if you have functionality implemented in terms of #method_missing, having a private method in Object will shadow it. And finally, monkey patching Object is just evil business :)
EDIT:
module_function can be used in a way similar to private:
module Something
def foo
puts 'foo'
end
module_function
def bar
puts 'bar'
end
end
That way, you can call Something.bar, but not not Something.foo. If you define any other methods after this call to module_function, they would also be available without mixing in.
I don't like it for two reasons, though. First, modules that are both mixed in and have "static" methods sound a bit dodgy. There might be valid cases, but it won't be that often. As I said, I prefer either to use a module as a namespace or mix it in, but not both.
Second, in this example, bar would also be available to classes/modules that mix in Something. I'm not sure when this is desirable, since either the method uses self and it has to be mixed in, or doesn't and then it does not need to be mixed in.
I think using module_function without passing the name of the method is used quite more often than with. Same goes for private and protected.
It's a good way for a Ruby library to offer functionality that does not use (much) internal state. So if you (e.g.) want to offer a sin function and don't want to pollute the "global" (Object) namespace, you can define it as class method under a constant (Math).
However, an app developer, who wants to write a mathematical application, might need sin every two lines. If the method is also an instance method, she can just include the Math (or My::Awesome::Nested::Library) module and can now directly call sin (stdlib example).
It's really about making a library more comfortable for its users. They can choose themself, if they want the functionality of your library on the top level.
By the way, you can achieve a similar functionality like module_function by using: extend self (in the first line of the module). To my mind, it looks better and makes things a bit clearer to understand.
Update: More background info in this blog article.
If you want to look at a working example, check out the chronic gem:
https://github.com/mojombo/chronic/blob/master/lib/chronic/handlers.rb
and Handlers is being included in the Parser class here:
https://github.com/mojombo/chronic/blob/master/lib/chronic/parser.rb
He's using module_function to send the methods from Handlers to specific instances of Handler using that instance's invoke method.