I feel like I understand the difference between require and require_dependency (from How are require, require_dependency and constants reloading related in Rails?).
However, I'm wondering what should happen if I use some of the various methods out there (see http://hemju.com/2010/09/22/rails-3-quicktip-autoload-lib-directory-including-all-subdirectories/ and Best way to load module/class from lib folder in Rails 3?) to get all files loading so we:
don't need to use require_dependency all over the place in the application and
don't have to restart development servers when files in the lib directory change.
It seems like development performance would be slightly impacted, which is not that big of a deal to me. How would performance be impacted in a production environment? Do all of the files generally get loaded only once if you are in production anyway? Is there a better way that I'm not seeing?
If you could include some resources where I could read more about this, they would be greatly appreciated. Some blog posts said that this behavior changed recently with Rails 3 for autoreloading lib/* files and that it was contentious, but I didn't see any links to these discussions. It would be helpful for considering the pros/cons. Thanks!
The code reloader is disabled by default in production. So if you are calling require_dependency at the top of a file it is going to be executed only once.
The Rails 3 change you mentioned is really small. You can usually call Foo and it will be loaded from app/models/foo.rb automatically. Before it could also be loaded from lib/foo.rb. (These directories app/models and lib are called autoload paths.) Rails team decided to remove lib from autoload paths in the 3rd version. You can still put it back. But it is encouraged to leave in lib less frequenttly-changed and project-specific files. If you have something that does not belong to any of the default app subdirectories like app/models or app/controllers you don't have to put it in lib. You can add your own subdirectory. I have app/presenters, for example. There is a discussion on the old issue tracker if you want more info on that.
Related
In the past, Rails developers who had files in the lib/ directory were told to add the lib directory to the autoload paths, by adding a line like this to the config/application.rb:
config.autoload_paths << "lib"
However, the latest Rails guides say that this is now discouraged:
... using autoload_paths on its own in the past (before Rails 5) developers might configure autoload_paths to add in extra locations (e.g. lib which used to be an autoload path list years ago, but no longer is). However this is now discouraged for most purposes, as it is likely to lead to production-only errors. It is possible to add new locations to both config.eager_load_paths and config.autoload_paths but use at your own risk.
(emphasis mine)
So instead we are asked to require lib/ files. From the Rails guides:
Of course, using require in autoloaded files to load ordinary 3rd party libraries is fine, and Rails is able to distinguish their constants, they are not marked as autoloaded.
However, this means that changes in those required files won't be noticed and served in the next request. So in order to get that to happen for lib/ files, we must add them to the autoload_paths... which is discouraged above.
So what is the appropriate way to include lib files in your app, with automatic detection of changes in Rails 5, or Rails 6 with the classic autoloader enabled?
As per this discussion Rails 5: Load lib files in production you should put your libs under app/lib. Or not. There are different opinions about "appropriate" way.
In the sidekiq documentation, there is this quote about preferring to use /app/lib instead of /lib in Rails projects related to autoloading errors:
A lib/ directory will only cause pain. Move the code to app/lib/ and make sure the code inside follows the class/filename conventions.
Additionally, there is also:
Don't configure extra paths in autoload_paths or eager_load_paths. That's a hack; follow the conventions! Any directory underneath app/ may contain Ruby code, you don't need to explicitly configure anything.
My questions are:
Is there any truth to these statements that using /app/lib is better than /lib?
Is this only helpful for autoloading Rails-related objects (such as AR models, controllers, jobs, etc)? Or will it also help POROs?
Is there only a specific context in which these comments make sense?
In my experience app/lib is easier to use. You can literally stick in something like Class MathFunction and use it elsewhere (e.g. controllers or modules) with MathFunction.sqrRoot.
To use /lib you need to configure your Rails app with autoload_paths. autoload_paths also need some tweaking to work properly in production. Matz himself discourages autoload because it's in the process of being deprecated.
The only time I've needed to use the lib directory is for making custom rake tasks. Otherwise I stick to app/lib.
I'm developing a Rails engine, and so I've taken some looks on existing ones. I noticed that many of them have files in app, but also in lib and vendor.
It's clear to me that I should put any code that should be replaceable by the host app into the app folder (e.g. when having a model app/user.rb, the host app can easily have its own app/user.rb file and use this one instead of the engine one's).
But I'm unsure, when I have to put stuff in to lib, and when into vendor? I thought, that in vendor, I should only put "external" code from other developers or projects, that I want to use in my project, and in lib I put my own additional libraries that I'm actually working on in the project. But why, for example, does WiceGrid put stuff into its wice_grid/vendor/assets directory? It doesn't look to me like external code, but code that is developed only for WiceGrid and hence should be in the lib directory?
Update
While experimenting a bit, I noticed that all code in the lib folder is not reloaded while developing the engine (I guess the same is the case for the vendor directory), so I should put them into a folder within app, but where exactly?
For example, I have a file lib/iq_list_controller.rb that holds some class and instance methods for ApplicationController which I mix into it in engine.rb like this:
initializer "wice_grid_railtie.configure_rails_initialization" do |app|
ActiveSupport.on_load(:action_controller) do
extend IqList::Controller::ClassMethods
include IqList::Controller::InstanceMethods
end
end
Where should I put this file so Ruby properly finds it?
Regarding the development reloading issue, if the lib folder is a natural home for your files, then add it to Rails' load path with something like:
module MyEngine
class Engine < ::Rails::Engine
config.autoload_paths << File.expand_path("../../lib", __FILE__)
end
end
Regarding the specific case of vendor assets, it seems reasonable to place your engine's assets in app/assets, where they will be found by the host Rails app.
if you want things to be autoloaded then put them in /app. Otherwise I think that anything in /lib should be manually required. I generally believe that autoloading the lib folder is bad practice.
your basic code for MVC comes in app folder.now suppose you have a generic method/module which is frequently used,so to implement DRYness for your code...all generic and common utilities can be simply put in lib folder.
theres a beautiful explaination when to use lib and using lib folder in rails
Since the latest Rails 3 release is not auto-loading modules and classes from lib anymore,
what would be the best way to load them?
From github:
A few changes were done in this commit:
Do not autoload code in *lib* for applications (now you need to explicitly
require them). This makes an application behave closer to an engine
(code in lib is still autoloaded for plugins);
As of Rails 2.3.9, there is a setting in config/application.rb in which you can specify directories that contain files you want autoloaded.
From application.rb:
# Custom directories with classes and modules you want to be autoloadable.
# config.autoload_paths += %W(#{config.root}/extras)
# Autoload lib/ folder including all subdirectories
config.autoload_paths += Dir["#{config.root}/lib/**/"]
Source: Rails 3 Quicktip: Autoload lib directory including all subdirectories, avoid lazy loading
Please mind that files contained in the lib folder are only loaded when the server is started. If you want the comfort to autoreload those files, read: Rails 3 Quicktip: Auto reload lib folders in development mode. Be aware that this is not meant for a production environment since the permanent reload slows down the machine.
The magic of autoloading stuff
I think the option controlling the folders from which autoloading stuff gets done has been sufficiently covered in other answers. However, in case someone else is having trouble stuff loaded though they've had their autoload paths modified as required, then this answer tries to explain what is the magic behind this autoload thing.
So when it comes to loading stuff from subdirectories there's a gotcha or a convention you should be aware. Sometimes the Ruby/Rails magic (this time mostly Rails) can make it difficult to understand why something is happening. Any module declared in the autoload paths will only be loaded if the module name corresponds to the parent directory name. So in case you try to put into lib/my_stuff/bar.rb something like:
module Foo
class Bar
end
end
It will not be loaded automagically. Then again if you rename the parent dir to foo thus hosting your module at path: lib/foo/bar.rb. It will be there for you. Another option is to name the file you want autoloaded by the module name. Obviously there can only be one file by that name then. In case you need to split your stuff into many files you could of course use that one file to require other files, but I don't recommend that, because then when on development mode and you modify those other files then Rails is unable to automagically reload them for you. But if you really want you could have one file by the module name that then specifies the actual files required to use the module. So you could have two files: lib/my_stuff/bar.rb and lib/my_stuff/foo.rb and the former being the same as above and the latter containing a single line: require "bar" and that would work just the same.
P.S. I feel compelled to add one more important thing. As of lately, whenever I want to have something in the lib directory that needs to get autoloaded, I tend to start thinking that if this is something that I'm actually developing specifically for this project (which it usually is, it might some day turn into a "static" snippet of code used in many projects or a git submodule, etc.. in which case it definitely should be in the lib folder) then perhaps its place is not in the lib folder at all. Perhaps it should be in a subfolder under the app folder· I have a feeling that this is the new rails way of doing things. Obviously, the same magic is in work wherever in you autoload paths you put your stuff in so it's good to these things. Anyway, this is just my thoughts on the subject. You are free to disagree. :)
UPDATE: About the type of magic..
As severin pointed out in his comment, the core "autoload a module mechanism" sure is part of Ruby, but the autoload paths stuff isn't. You don't need Rails to do autoload :Foo, File.join(Rails.root, "lib", "my_stuff", "bar"). And when you would try to reference the module Foo for the first time then it would be loaded for you. However what Rails does is it gives us a way to try and load stuff automagically from registered folders and this has been implemented in such a way that it needs to assume something about the naming conventions. If it had not been implemented like that, then every time you reference something that's not currently loaded it would have to go through all of the files in all of the autoload folders and check if any of them contains what you were trying to reference. This in turn would defeat the idea of autoloading and autoreloading. However, with these conventions in place it can deduct from the module/class your trying to load where that might be defined and just load that.
Warning: if you want to load the 'monkey patch' or 'open class' from your 'lib' folder, don't use the 'autoload' approach!!!
"config.autoload_paths" approach: only works if you are loading a class that defined only in ONE place. If some class has been already defined somewhere else, then you can't load it again by this approach.
"config/initializer/load_rb_file.rb" approach: always works! whatever the target class is a new class or an "open class" or "monkey patch" for existing class, it always works!
For more details , see: https://stackoverflow.com/a/6797707/445908
Very similar, but I think this is a little more elegant:
config.autoload_paths += Dir["#{config.root}/lib", "#{config.root}/lib/**/"]
In my case I was trying to simply load a file directly under the lib dir.
Within application.rb...
require '/lib/this_file.rb'
wasn't working, even in console and then when I tried
require './lib/this_file.rb'
and rails loads the file perfectly.
I'm still pretty noob and I'm not sure why this works but it works. If someone would like to explain it to me I'd appreciate it :D I hope this helps someone either way.
I had the same problem. Here is how I solved it. The solution loads the lib directory and all the subdirectories (not only the direct). Of course you can use this for all directories.
# application.rb
config.autoload_paths += %W(#{config.root}/lib)
config.autoload_paths += Dir["#{config.root}/lib/**/"]
As of Rails 5, it is recommended to put the lib folder under app directory or instead create other meaningful name spaces for the folder as services , presenters, features etc and put it under app directory for auto loading by rails.
Please check this GitHub Discussion Link as well.
config.autoload_paths does not work for me. I solve it in other way
Ruby on rails 3 do not automatic reload (autoload) code from /lib folder. I solve it by putting inside ApplicationController
Dir["lib/**/*.rb"].each do |path|
require_dependency path
end
If only certain files need access to the modules in lib, just add a require statement to the files that need it. For example, if one model needs to access one module, add:
require 'mymodule'
at the top of the model.rb file.
Spell the filename correctly.
Seriously. I battled with a class for an hour because the class was Governance::ArchitectureBoard and the file was in lib/governance/architecture_baord.rb (transposed O and A in "board")
Seems obvious in retrospect, but it was the devil tracking that down. If the class is not defined in the file that Rails expects it to be in based on munging the class name, it is simply not going to find it.
There are several reasons you could have problems loading from lib - see here for details - http://www.williambharding.com/blog/technology/rails-3-autoload-modules-and-classes-in-production/
fix autoload path
threadsafe related
naming relating
...
I'm attempting to reopen the String class in rails and add a bunch more methods for my app to use. Writing the code isn't a problem - my question is rather about where this code should go.
It doesn't make sense to me to reopen a class inside a different model file, because it really has nothing to do with any of the models specifically. I thought perhaps somewhere in config or lib would make sense, but I'm not particularly well versed with RoR yet.
To summarize, where would be the most logical place to define class-modifying code, and are there any implications depending on where/when the code is loaded?
The most logical place is probably in a file in the config/initializers directory. Any *.rb file you put in here will be automatically executed when rails boots. If you want, you could put them in a sub folder, so you could do something like config/initializers/extensions/*.rb.
I try keep these monkey-patches to a minimum, only when they are very clearly in the best interest of my code.
Lately I have preferred to keep the files organized in folders such as lib/monkey/string.rb, lib/monkey/hash.rb, etc. I then require all files in the lib/monkey folder in my environment.rb file.
# Load all monkey-patches.
Dir["lib/monkey/*.rb"].each {|monkeyfile| require monkeyfile}
This keeps all of my class modifying code isolated to one location, should a problem arise. I also enjoy the somewhat goofy naming, because it makes it stand out as something to be mindful of. Someone may have a better system, if so ... I'd love to hear about it!