I'm trying to understand what is an FP-alternative to good old dependency injection from OOP.
Say I have the following app (pseudocode)
app() is where application starts. It allows user to register and list user posts (whatever). These two functions are composed out of several other functions (register does it step by step, imperatively, while list posts really composes them (at least this is how I understand function composition).
app()
registerUser(u)
validate(u)
persist(u)
callSaveToDB(u)
notify(u)
sendsEmail
listPosts(u)
postsToView(loadUserPosts(findUser(u)))
Now I'd like to test this stuff (registerUser and listPosts) and would like to have stubbed functions so that I don't call db etc - you know, usual testing stuff.
I know it's possible to pass functions to functions e.g
registerUser(validateFn, persistFn, notifyFn, u)
and have it partially applied so it looks like registerUser(u) with other functions closed over and so on. But it all needs to be done on app boot level as it was in OOP (wiring dependencies and bootstraping an app). It looks like manually doing this will take ages and tons of boilerplate code. Is there something obvious I'm missing there? Is there any other way of doing that?
EDIT:
I see having IO there is not a good example. So what if I have function composed of several other functions and one of them is really heavy (in terms of computations) and I'd like to swap it?
Simply - I'm looking for FP way of doing DI stuff.
The way to answer this is to drop the phrase "dependency injection" and think about it more fundamentally. Write down interfaces as types for each component. Implement functions that have those types. Replace them as needed. There's no magic, and language features like type classes make it easy for the compiler to ensure you can substitute methods in an interface.
The previous Haskell-specific answer, shows how to use Haskell types for the API: https://stackoverflow.com/a/14329487/83805
It's possible to add extensions to existing Swift object types using extensions, as described in the language specification.
As a result, it's possible to create extensions such as:
extension String {
var utf8data:NSData {
return self.dataUsingEncoding(NSUTF8StringEncoding, allowLossyConversion: false)!
}
}
However, what's the best naming practice for Swift source files containing such extensions?
In the past, the convention was to use extendedtype+categoryname.m for the Objective-C
type as discussed in the Objective-C guide. But the Swift example doesn't have a category name, and calling it String.swift doesn't seem appropriate.
So the question is: given the above String extension, what should the swift source file be called?
Most examples I have seen mimic the Objective-C approach. The example extension above would be:
String+UTF8Data.swift
The advantages are that the naming convention makes it easy to understand that it is an extension, and which Class is being extended.
The problem with using Extensions.swift or even StringExtensions.swift is that it's not possible to infer the purpose of the file by its name without looking at its contents.
Using xxxable.swift approach as used by Java works okay for protocols or extensions that only define methods. But again, the example above defines an attribute so that UTF8Dataable.swift doesn't make much grammatical sense.
I prefer having a + to underline the fact it contains extensions :
String+Extensions.swift
And if the file gets too big, you can then split it for each purpose :
String+UTF8Data.swift
String+Encrypt.swift
There is no Swift convention. Keep it simple:
StringExtensions.swift
I create one file for each class I'm extending. If you use a single file for all extensions, it will quickly become a jungle.
I prefer StringExtensions.swift until I added too much things to split the file into something like String+utf8Data.swift and String+Encrypt.swift.
One more thing, to combine similar files into one will make your building more faster. Refer to Optimizing-Swift-Build-Times
Rather than adding my comments all over the place, I'm surfacing them all here in one answer.
Personally, I take a hybrid approach that gives both good usability and clarity, while also not cluttering up the API surface area for the object that I'm extending.
For instance, anything that makes sense to be available to any string would go in StringExtensions.swift such as trimRight() and removeBlankLines().
However, if I had an extension function such as formatAsAccountNumber() it would not go in that file because 'Account Number' is not something that would naturally apply to any/all strings and only makes sense in the context of accounts. In that case, I would create a file called Strings+AccountFormatting.swift or maybe even Strings+CustomFormatting.swift with a formatAsAccountNumber() function if there are several types/ways to actually format it.
Actually, in that last example, I actively dissuade my team from using extensions like that in the first place, and would instead encourage something like AccountNumberFormatter.format(String) instead as that doesn't touch the String API surface area at all, as it shouldn't. The exception would be if you defined that extension in the same file where it's used, but then it wouldn't have it's own filename anyway.
If you have a team-agreed set of common and miscellaneous enhancements, lumping them together as an Extensions.swift works as Keep-It-Simple first level solution. However, as your complexity grows, or the extensions become more involved, a hierarchy is needed to encapsulate the complexity. In such circumstances I recommend the following practice with an example.
I had a class which talks to my back-end, called Server. It started to grow bigger to cover two different target apps. Some people like a large file but just logically split up with extensions. My preference is to keep each file relatively short so I chose the following solution. Server originally conformed to CloudAdapterProtocol and implemented all its methods. What I did was to turn the protocol into a hierarchy, by making it refer to subordinate protocols:
protocol CloudAdapterProtocol: ReggyCloudProtocol, ProReggyCloudProtocol {
var server: CloudServer {
get set
}
func getServerApiVersion(handler: #escaping (String?, Error?) -> Swift.Void)
}
In Server.swift I have
import Foundation
import UIKit
import Alamofire
import AlamofireImage
class Server: CloudAdapterProtocol {
.
.
func getServerApiVersion(handler: #escaping (String?, Error?) -> Swift.Void) {
.
.
}
Server.swift then just implements the core server API for setting the server and getting the API version. The real work is split into two files:
Server_ReggyCloudProtocol.swift
Server_ProReggyCloudProtocol.swift
These implement the respective protocols.
It means you need to have import declarations in the other files (for Alamofire in this example) but its a clean solution in terms of segregating interfaces in my view.
I think this approach works equally well with externally specified classes as well as your own.
Why is this even a debate? Should I put all my sub classes into a file called _Subclasses.swift. I think not. Swift has module based name spacing. To extend a well known Swift class needs a file that is specific to its purpose. I could have a large team that creates a file that is UIViewExtensions.swift that express no purpose and will confuse developers and could be easily duplicated in the project which would not build. The Objective-C naming convention works fine and until Swift has real name spacing, it is the best way to go.
I wish to replace the implementation of System.Web.Mvc.ActionMethodSelector as used by the FindAction method of ReflectedControllerDescriptor, but would like to take advantage of the existing implementation, ideally by deriving from ActionMethodSelector. However, because the class is marked as internal the only way I can see to do this 'properly' is to derive from ReflectedControllerDescriptor and implement FindAction by copying the code from ActionMethodSelector. I wish to avoid this however due to the quantity of code, and potential issues trying to keep it up to date with the framework.
I'm considering the following approaches:
Biting the bullet and copying the code
Using reflection so as to take advantage of the existing implementation
Are there any other approaches that I'm missing, better or otherwise?
I know it is a bit late to answer still I am giving it a try.... :)
I believe that you somehow want to tweak action method selection process in ASP.NET MVC. If my understanding is correct you can make use of custom ActionMethodSelectorAttribute by deriving from System.Web.Mvc.ActionMethodSelectorAttribute. Write your own custom logic in the custom selector and apply it on the top of the action methods. I believe in this way the action method selection process can be tweaked without disturbing the natural process.
If you wish you can visit these links: http://programersnotebook.blogspot.in/2014/02/aspnet-mvc-actionnameselector-and.html, http://programersnotebook.blogspot.in/2014/02/aspnet-mvc-actionnameselector-and_2.html
A rule in xText called "Component" will typically generate a class "Component" in the src-gen folder.
I would like to add additional methods to these classes without them being overridden every time I make minor changes to the DSL. What's the proper way to inject my own code into these classes and is there a way to make all classes extend my own root class instead of the default EObject?
Thanks in advance.
You basically have two choices:
You can use a IXtext2EcorePostProcessor to modify the EMF-model which Xtext inferred from your grammar. The actual code generation is done by EMF, so you have to fiddle your code through that bottleneck. The details are described in a blog of Christian Dietrich. This approach is only suitable for small modifications.
You can use the "generation gap pattern" (a.k. "implementation gap pattern") which allows you the write classes which derive from the generated model classes. Here you can add anything you want. The details are described in a blog of Heiko Behrens. This approach is better suited for large scale modifcations by inheritance.
You may of course mix the two approaches...
In the Inferring a JVM Model section of the Xtext documentation (http://www.eclipse.org/Xtext/documentation.html#_17) it starts by saying:
"In many cases, you will want your DSLs concepts to be usable as Java elements. E.g. an Entity will become a Java class and should be usable as such".
In the example above, how can I use the generated Entity class outside of xbase, i.e. in real Java code in a different project to the xtext one?
What I am essentially asking is if the Java classes created my by the model Inferrer can actually be used as real java classes, which can have their methods called and fields accessed from java code, in an altogether different project, and if so how this can be done?
My going through the documentation has lead me to fear that the generated "Java classes" are only Xbase types, only referenceabe in an xtext context, and are not therefore real java classes...
The Xbase compiler can compile all Xbase expressions to plain Java code usable everywhere where Java codes are available.
If you add your own elements to the language, you have to extend the generator to also support these elements - for this reason you define your own JVMModelInferrer.
The basic Xtext compiler then executes the JVMModelInferrer, calculates the JVM model that might (or might not) contain Xbase expressions as well; then this JVM model can be generated into a Java-compilable (thus Java-reusable) code.
If you want to test this functionality, simply generate the Xtext Domain Model example (available from the New... wizards in Xtext/Examples category), and evaluate the results: when you edit your domain model, Xtext automatically generates the usable Java code (if the required dependencies are set).