I'm trying to wrap all instance methods of TestClass to perform code before and after an instance method is called. So far, this code is working:
module Wrapper
def wrap(*methods)
prependable_module = Module.new do
methods.each do |m|
define_method(m) do |*args, &block|
p 1
super(*args, &block)
p 3
end
end
end
prepend prependable_module
end
end
class TestClass
extend Wrapper
wrap :instance_method1
def instance_method1
p 2
end
end
TestClass.new.instance_method1 # => 1, 2, 3
I can call wrap with all method names as arguments. If I try to wrap all methods without listing them individually, I need to call it using instance_methods(false) at the bottom of the class definition.
class TestClass
extend Wrapper
def instance_method1
p 2
end
wrap(*instance_methods(false))
end
In Rails, all callback methods like before_action or after_create are usually called on top of the class definition. My goal is to call wrap on top of the class definition as well (without listing all methods individually). In this case, I can't call instance_methods(false) on top of the class definition, because at this point no method has been defined.
Thanks for your help!
Update
Thanks to Kimmo Lehto's approach I can wrap every instance method using the method_added hook. I don't want to prepend a new module for every defined method, so I add all overridden methods to the same module.
module Wrapper
def method_added(method_name)
tmp_module = find_or_initialize_module
return if tmp_module.instance_methods(false).include?(method_name)
tmp_module.define_method(method_name) do |*args, &block|
p 1
super(*args, &block)
p 3
end
end
def find_or_initialize_module
module_name = "#{name}Wrapper"
module_idx = ancestors.map(&:to_s).index(module_name)
unless module_idx
prepend Object.const_set(module_name, Module.new)
return find_or_initialize_module
end
ancestors[module_idx]
end
end
class TestClass
extend Wrapper
def instance_method1
p 2
end
end
tc = TestClass.new
tc.instance_method1 # => 1, 2, 3
You could use the Module#method_added hook to automatically wrap any methods that are added.
You will need some magic to not get a stack overflow from an infinite loop.
Another option is to use TracePoint to trigger the wrapping once the class has been defined. You can use the Module#extended to set up the tracepoint. Something like:
module Finalizer
def self.extended(obj)
TracePoint.trace(:end) do |t|
if obj == t.self
obj.finalize
t.disable
end
end
end
def finalize
wrap(*instance_methods(false))
end
end
Classes are usually not exactly "closed" unless you explicitly .freeze them so it's a bit of a hacky solution and will not trigger if methods are added afterwards. method_added is probably your best bet.
Related
When creating methods that yield, sometimes we want it to return an Enumerator if no block is given. The recommended way is basically return to_enum(:name_of_method, [args]) unless block_given?. However, it's a pain to have to type that for every method that does this. Ruby being ruby, I decided to create a make_enum method, similar to attr_accessor, which does this for me:
class Module # Put this in a mixin, but for the purposes of this experiment, it's in Module
def make_enum *args
args.each do |name|
old_method = instance_method(name)
define_method(name) do |*args, &block|
next to_enum(name, *args) unless block
old_method.bind(self).call(*args, &block)
end
end
end
end
Now I can use it like so:
class Test
def test
yield 1
yield 2
end
make_enum :test
end
t = Test.new
t.test { |n| puts n }
# 1
# 2
t.test.to_a #=> [1, 2]
And it works! But it doesn't work if make_enum is before the method definition.
How can I get this method to work before defining a method, so that the following works? Perhaps I need to make use of method_added?
class Test
make_enum :test
def test
yield 1
yield 2
end
end
I don't know if it's a bad idea for it to be before the method, but my reason for thinking that it would be nice to do that is that it better matches the way we use attr_accessor and the like.
Whereas attr_ methods create instance methods newly, your make_enum modifies an existing method, which is rather similar to protected, private, and public methods. Note that these visibility methods are used either in the form:
protected
def foo; ... end
or
protected def foo; ... end
or
def foo; ... end
protected :foo
The latter two ways are already available with your make_enum. Especially, the second form is already possible (which Stefan also notes in the comment). You can do:
make_enum def test; ... end
If you want to do the first form, you should try to implement that in your make_enum definition.
I'm trying to create a method to dynamically do the following: (as I will have to implement this on about 30 different sets of sub-classes)
def t1
FooT1.new
end
def t2
FooT2.new
end
def t3
FooT3.new
end
Where there will be 2 variables in the method generation, the tab number(t1...tx) and the name of the class (Foo)
I tried the following, but I'm new to Ruby and can not get this working.
def method_generator(num_tabs, class_name)
1.upto(num_tabs) do |i|
define_method("t#{i}") do
"#{class_name}_t#{i}".new
end
end
end
Then call it in the sub-class like so:
method_generator(3, "Bar")
I'm aware I'm probably quite far off in implementing this, so any help is appreciated.
Just do as below :
def method_generator(num_tabs, class_name)
1.upto(num_tabs) do |i|
class_name.send(:define_method,"t#{i}") do
"#{class_name}_t#{i}".new
end
end
end
Module#define_method is a private method, thus you can't call it on the class_name like class_name.define_method(:name) do ..end, as private method call not allows explicit receiver. But to do so Object#send will help you, as this method is here for this kind of scenarios, where you can't call private method by explicit receiver.
Lets verify with an example, if this tricks works or not.
class Foo;end
def method_generator(num_tabs, class_name)
1.upto(num_tabs) do |i|
class_name.send(:define_method,"t#{i}") do
"#{class_name}_t#{i}".new
end
end
end
method_generator(3,Foo)
Foo.instance_methods(false)
# => [:t1, :t2, :t3] # see here 3 instance methods has been created of class Foo
I would like to access a class' name in its superclass MySuperclass' self.inherited method. It works fine for concrete classes as defined by class Foo < MySuperclass; end but it fails when using anonymous classes. I tend to avoid creating (class-)constants in tests; I would like it to work with anonymous classes.
Given the following code:
class MySuperclass
def self.inherited(subclass)
super
# work with subclass' name
end
end
klass = Class.new(MySuperclass) do
def self.name
'FooBar'
end
end
klass#name will still be nil when MySuperclass.inherited is called as that will be before Class.new yields to its block and defines its methods.
I understand a class gets its name when it's assigned to a constant, but is there a way to set Class#name "early" without creating a constant?
I prepared a more verbose code example with failing tests to illustrate what's expected.
Probably #yield has taken place after the ::inherited is called, I saw the similar behaviour with class definition. However, you can avoid it by using ::klass singleton method instead of ::inherited callback.
def self.klass
#klass ||= (self.name || self.to_s).gsub(/Builder\z/, '')
end
I am trying to understand the benefit of being able to refer to an anonymous class by a name you have assigned to it after it has been created. I thought I might be able to move the conversation along by providing some code that you could look at and then tell us what you'd like to do differently:
class MySuperclass
def self.inherited(subclass)
# Create a class method for the subclass
subclass.instance_eval do
def sub_class() puts "sub_class here" end
end
# Create an instance method for the subclass
subclass.class_eval do
def sub_instance() puts "sub_instance here" end
end
end
end
klass = Class.new(MySuperclass) do
def self.name=(name)
#name = Object.const_set(name, self)
end
def self.name
#name
end
end
klass.sub_class #=> "sub_class here"
klass.new.sub_instance #=> "sub_instance here"
klass.name = 'Fido' #=> "Fido"
kn = klass.name #=> Fido
kn.sub_class #=> "sub_class here"
kn.new.sub_instance #=> "sub_instance here"
klass.name = 'Woof' #=> "Woof"
kn = klass.name #=> Fido (cannot change)
There is no way in pure Ruby to set a class name without assigning it to a constant.
If you're using MRI and want to write yourself a very small C extension, it would look something like this:
VALUE
force_class_name (VALUE klass, VALUE symbol_name)
{
rb_name_class(klass, SYM2ID(symbol_name));
return klass;
}
void
Init_my_extension ()
{
rb_define_method(rb_cClass, "force_class_name", force_class_name, 1);
}
This is a very heavy approach to the problem. Even if it works it won't be guaranteed to work across various versions of ruby, since it relies on the non-API C function rb_name_class. I'm also not sure what the behavior will be once Ruby gets around to running its own class-naming hooks afterward.
The code snippet for your use case would look like this:
require 'my_extension'
class MySuperclass
def self.inherited(subclass)
super
subclass.force_class_name(:FooBar)
# work with subclass' name
end
end
I'm trying to override a dynamically-generated method by including a module.
In the example below, a Ripple association adds a rows= method to Table. I want to call that method, but also do some additional stuff afterwards.
I created a module to override the method, thinking that the module's row= would be able to call super to use the existing method.
class Table
# Ripple association - creates rows= method
many :rows, :class_name => Table::Row
# Hacky first attempt to use the dynamically-created
# method and also do additional stuff - I would actually
# move this code elsewhere if it worked
module RowNormalizer
def rows=(*args)
rows = super
rows.map!(&:normalize_prior_year)
end
end
include RowNormalizer
end
However, my new rows= is never called, as evidenced by the fact that if I raise an exception inside it, nothing happens.
I know the module is getting included, because if I put this in it, my exception gets raised.
included do
raise 'I got included, woo!'
end
Also, if instead of rows=, the module defines somethingelse=, that method is callable.
Why isn't my module method overriding the dynamically-generated one?
Let's do an experiment:
class A; def x; 'hi' end end
module B; def x; super + ' john' end end
A.class_eval { include B }
A.new.x
=> "hi" # oops
Why is that? The answer is simple:
A.ancestors
=> [A, B, Object, Kernel, BasicObject]
B is before A in the ancestors chain (you can think of this as B being inside A). Therefore A.x always takes priority over B.x.
However, this can be worked around:
class A
def x
'hi'
end
end
module B
# Define a method with a different name
def x_after
x_before + ' john'
end
# And set up aliases on the inclusion :)
# We can use `alias new_name old_name`
def self.included(klass)
klass.class_eval {
alias :x_before :x
alias :x :x_after
}
end
end
A.class_eval { include B }
A.new.x #=> "hi john"
With ActiveSupport (and therefore Rails) you have this pattern implemented as alias_method_chain(target, feature) http://apidock.com/rails/Module/alias_method_chain:
module B
def self.included(base)
base.alias_method_chain :x, :feature
end
def x_with_feature
x_without_feature + " John"
end
end
Update Ruby 2 comes with Module#prepend, which does override the methods of A, making this alias hack unnecessary for most use cases.
Why isn't my module method overriding the dynamically-generated one?
Because that's not how inheritance works. Methods defined in a class override the ones inherited from other classes/modules, not the other way around.
In Ruby 2.0, there's Module#prepend, which works just like Module#include, except it inserts the module as a subclass instead of a superclass in the inheritance chain.
If you extend the instance of the class, you will can do it.
class A
def initialize
extend(B)
end
def hi
'hi'
end
end
module B
def hi
super[0,1] + 'ello'
end
end
obj = A.new
obj.hi #=> 'hello'
Background:
I have a module which declares a number of instance methods
module UsefulThings
def get_file; ...
def delete_file; ...
def format_text(x); ...
end
And I want to call some of these methods from within a class. How you normally do this in ruby is like this:
class UsefulWorker
include UsefulThings
def do_work
format_text("abc")
...
end
end
Problem
include UsefulThings brings in all of the methods from UsefulThings. In this case I only want format_text and explicitly do not want get_file and delete_file.
I can see several possible solutions to this:
Somehow invoke the method directly on the module without including it anywhere
I don't know how/if this can be done. (Hence this question)
Somehow include Usefulthings and only bring in some of it's methods
I also don't know how/if this can be done
Create a proxy class, include UsefulThings in that, then delegate format_text to that proxy instance
This would work, but anonymous proxy classes are a hack. Yuck.
Split up the module into 2 or more smaller modules
This would also work, and is probably the best solution I can think of, but I'd prefer to avoid it as I'd end up with a proliferation of dozens and dozens of modules - managing this would be burdensome
Why are there lots of unrelated functions in a single module? It's ApplicationHelper from a rails app, which our team has de-facto decided on as the dumping ground for anything not specific enough to belong anywhere else. Mostly standalone utility methods that get used everywhere. I could break it up into seperate helpers, but there'd be 30 of them, all with 1 method each... this seems unproductive
I think the shortest way to do just throw-away single call (without altering existing modules or creating new ones) would be as follows:
Class.new.extend(UsefulThings).get_file
If a method on a module is turned into a module function you can simply call it off of Mods as if it had been declared as
module Mods
def self.foo
puts "Mods.foo(self)"
end
end
The module_function approach below will avoid breaking any classes which include all of Mods.
module Mods
def foo
puts "Mods.foo"
end
end
class Includer
include Mods
end
Includer.new.foo
Mods.module_eval do
module_function(:foo)
public :foo
end
Includer.new.foo # this would break without public :foo above
class Thing
def bar
Mods.foo
end
end
Thing.new.bar
However, I'm curious why a set of unrelated functions are all contained within the same module in the first place?
Edited to show that includes still work if public :foo is called after module_function :foo
Another way to do it if you "own" the module is to use module_function.
module UsefulThings
def a
puts "aaay"
end
module_function :a
def b
puts "beee"
end
end
def test
UsefulThings.a
UsefulThings.b # Fails! Not a module method
end
test
If you want to call these methods without including module in another class then you need to define them as module methods:
module UsefulThings
def self.get_file; ...
def self.delete_file; ...
def self.format_text(x); ...
end
and then you can call them with
UsefulThings.format_text("xxx")
or
UsefulThings::format_text("xxx")
But anyway I would recommend that you put just related methods in one module or in one class. If you have problem that you want to include just one method from module then it sounds like a bad code smell and it is not good Ruby style to put unrelated methods together.
To invoke a module instance method without including the module (and without creating intermediary objects):
class UsefulWorker
def do_work
UsefulThings.instance_method(:format_text).bind(self).call("abc")
...
end
end
Not sure if someone still needs it after 10 years but I solved it using eigenclass.
module UsefulThings
def useful_thing_1
"thing_1"
end
class << self
include UsefulThings
end
end
class A
include UsefulThings
end
class B
extend UsefulThings
end
UsefulThings.useful_thing_1 # => "thing_1"
A.new.useful_thing_1 # => "thing_1"
B.useful_thing_1 # => "thing_1"
Firstly, I'd recommend breaking the module up into the useful things you need. But you can always create a class extending that for your invocation:
module UsefulThings
def a
puts "aaay"
end
def b
puts "beee"
end
end
def test
ob = Class.new.send(:include, UsefulThings).new
ob.a
end
test
A. In case you, always want to call them in a "qualified", standalone way (UsefulThings.get_file), then just make them static as others pointed out,
module UsefulThings
def self.get_file; ...
def self.delete_file; ...
def self.format_text(x); ...
# Or.. make all of the "static"
class << self
def write_file; ...
def commit_file; ...
end
end
B. If you still want to keep the mixin approach in same cases, as well the one-off standalone invocation, you can have a one-liner module that extends itself with the mixin:
module UsefulThingsMixin
def get_file; ...
def delete_file; ...
def format_text(x); ...
end
module UsefulThings
extend UsefulThingsMixin
end
So both works then:
UsefulThings.get_file() # one off
class MyUser
include UsefulThingsMixin
def f
format_text # all useful things available directly
end
end
IMHO it's cleaner than module_function for every single method - in case want all of them.
As I understand the question, you want to mix some of a module's instance methods into a class.
Let's begin by considering how Module#include works. Suppose we have a module UsefulThings that contains two instance methods:
module UsefulThings
def add1
self + 1
end
def add3
self + 3
end
end
UsefulThings.instance_methods
#=> [:add1, :add3]
and Fixnum includes that module:
class Fixnum
def add2
puts "cat"
end
def add3
puts "dog"
end
include UsefulThings
end
We see that:
Fixnum.instance_methods.select { |m| m.to_s.start_with? "add" }
#=> [:add2, :add3, :add1]
1.add1
2
1.add2
cat
1.add3
dog
Were you expecting UsefulThings#add3 to override Fixnum#add3, so that 1.add3 would return 4? Consider this:
Fixnum.ancestors
#=> [Fixnum, UsefulThings, Integer, Numeric, Comparable,
# Object, Kernel, BasicObject]
When the class includes the module, the module becomes the class' superclass. So, because of how inheritance works, sending add3 to an instance of Fixnum will cause Fixnum#add3 to be invoked, returning dog.
Now let's add a method :add2 to UsefulThings:
module UsefulThings
def add1
self + 1
end
def add2
self + 2
end
def add3
self + 3
end
end
We now wish Fixnum to include only the methods add1 and add3. Is so doing, we expect to get the same results as above.
Suppose, as above, we execute:
class Fixnum
def add2
puts "cat"
end
def add3
puts "dog"
end
include UsefulThings
end
What is the result? The unwanted method :add2 is added to Fixnum, :add1 is added and, for reasons I explained above, :add3 is not added. So all we have to do is undef :add2. We can do that with a simple helper method:
module Helpers
def self.include_some(mod, klass, *args)
klass.send(:include, mod)
(mod.instance_methods - args - klass.instance_methods).each do |m|
klass.send(:undef_method, m)
end
end
end
which we invoke like this:
class Fixnum
def add2
puts "cat"
end
def add3
puts "dog"
end
Helpers.include_some(UsefulThings, self, :add1, :add3)
end
Then:
Fixnum.instance_methods.select { |m| m.to_s.start_with? "add" }
#=> [:add2, :add3, :add1]
1.add1
2
1.add2
cat
1.add3
dog
which is the result we want.
After almost 9 years here's a generic solution:
module CreateModuleFunctions
def self.included(base)
base.instance_methods.each do |method|
base.module_eval do
module_function(method)
public(method)
end
end
end
end
RSpec.describe CreateModuleFunctions do
context "when included into a Module" do
it "makes the Module's methods invokable via the Module" do
module ModuleIncluded
def instance_method_1;end
def instance_method_2;end
include CreateModuleFunctions
end
expect { ModuleIncluded.instance_method_1 }.to_not raise_error
end
end
end
The unfortunate trick you need to apply is to include the module after the methods have been defined. Alternatively you may also include it after the context is defined as ModuleIncluded.send(:include, CreateModuleFunctions).
Or you can use it via the reflection_utils gem.
spec.add_dependency "reflection_utils", ">= 0.3.0"
require 'reflection_utils'
include ReflectionUtils::CreateModuleFunctions
This old question comes to me today when I am studing Ruby and found interesting so I want to answer with my new knowlege.
Assume that you have the module
module MyModule
def say
'I say'
end
def cheer
'I cheer'
end
end
then with the class so call Animal I can take cheer method from MyModule as following
class Animal
define_method(:happy, MyModule.method(:cheer))
end
This is so called unbound method, so you can take a callable object and bind it to another place(s).
From this point, you can use the method as usual, such as
my_dog = Animal.new
my_dog.happy # => "I cheer"
Hope this help as I also learned something new today.
To learn further, you can use irb and take a look at Method object.