I often need a client bundle and some i18n-ed messages in presenters and views.
I would like to know which is the best way to get them : Injection or Singleton?
Solution 1: Up to now, I used to get the messages using a Singleton :
public interface MyMessages extends Messages{
String key1();
String key2();
...
class Instance {
private static MyMessages instance = null;
public static MyMessages getInstance() {
if (instance == null) {
instance = GWT.create(MyMessages.class);
}
return instance;
}
}
}
FooView.java :
MyMessages.Instance.getInstance().key1();
Solution 2: Would it be better to get it with an injection like this ?
private MyMessages i18n;
#Inject
public FooView(MyMessages i18n){
this.i18n=i18n;
}
The second solution seems cleaner to me but I sometimes get stuck when I need a non-empty constructor which uses some i18n strings:
#Inject
private MyMessages i18n;
public Bar(Foo foo){
/*
* do something which absolutely requires i18n here.
* The problem is that injectable attributes are called
* after the constructor so i18n is null here.
*/
foobar();
}
First, client bundles and I18N messages, while not singleton themselves, share their state with all their instances, so that once compiled to JavaScript and optimized by the compiler it's as if they were singletons. There are a few corner-case (IIRC, when using the WithLookup variants of I18N interfaces) but generally speaking it doesn't buy you anything explicitly treating them as singletons.
So the question basically becomes whether to use GWT.create() explicitly or have the instance injected. I'd say it's a matter of taste, but also technically GWT.create() doesn't play nice with non-GWTTestCase unit-tests.
Finally, as for your latest question, I suppose that by "non-null constructor" you mean that it takes values that aren't dependencies (i.e. value objects); in which case you should probably use assisted-injection rather than constructing the object yourself and then injecting its members (as an aside: how are you injecting the members then?)
Related
I was kind of lazy and used to use almost entirely field injections. I was just providing empty constructor, put my #Inject fields I everything was looking nice and simple. However field injection have its trade-offs so I've devised some simple rules that help me to decide when to used field and when to use constructor injections. I will appreciate any feedback if there is mistake in my logic or if you have additional considerations to add.
First some clarification in order to be on the same page:
Constructor injection:
#Inject
public SomeClass(#Named("app version") String appVersion,
AppPrefs appPrefs) {...
Same with the field injection:
public class SomeClass {
#Inject
#Named("app version") String mAppVersion;
#Inject
AppPrefs appPrefs;
Rule 1: MUST use field injection if I don't control creation of the object (think Activity or Fragment in Android). If some (non-dagger aware) framework is creating my object and handles it to me I have no choice but to inject it manually after I receive the instance.
Rule 2: MUST use constructor injection if the class is/may be used in another project that does not use Dagger 2. If the other project(s) do not use Dagger they cannot use DI so the user have to create the object the "old" way using new.
Rule 3: PREFER constructor injection when working with class hierarchies because it is easier to create unit tests.
Clarification:
Considering the following structure that uses field injection:
package superclass;
public class SuperClass {
#Inject
HttpClient mHttpClient;
...
}
.
package differentpackage;
public class SubClass extends SuperClass {
public SubClass() {
}
}
When I am creating unit test for SubClass in directory test/java/differentpackage I have no choice but to bring up the entire DI infrastructure in order to be able to inject the HttpClient. In contrast, if I was using constructor injection like this:
public class SuperClass {
private final HttpClient mHttpClient;
#Inject
public SuperClass(HttpClient httpClient) {
mHttpClient = httpClient;
}
}
in my unit test I could simply:
HttpClient mockHttp = mock(HttpClient.class);
Subclass tested = new Subclass(mockHttp);
// tests
So basically now I am in the other extreme: I tend to rely mostly on constructor injections and use field injections only when 'Rule 1' applies.
The only 'problem' that I have with the constructor injects is that for 'end' classes constructors sometimes become quite overloaded with parameters and they look verbose and ugly like this:
#Inject
public ModelMainImpl(#Named("app version") String appVersion,
AppPrefs appPrefs,
LoginPrefs loginPrefs,
#ForApplication Context appContext,
NetworkInfoProvider networkInfoProvider,
AndroidEventPoster androidEventPoster,
Session session,
ForgeExchangeManager exchangeManager,
HttpFunctionality httpFunctionality,
#Named("base url") String baseUrl,
#Named("forge result producer") ResultProducer<ForgeExchangeResult> resultProducer
) {
Guys, what are your rules to choose between constructor and field injects? I am missing something, are there errors in my logic?
Use constructor injection. if you can't, use property injection.
Rule 1 seems ok, like decorations or attributes you can use Property(field) injection.
Rule 2 seems ok, because who uses your class they have to follow your constructor. They may not know they have to intilaize your property also.
Rule 3 It's not just good for unit test. It's good for applying Single Responsibilty. It's easier to see your object graph.Otherwise you will hide it with property.
If we come in your question, yes there a lot of parameters in your constructor. But the solution is not property injection. You can refactor your code and use aggregate services
I have used the standard factory pattern method before to create instances of classes (implementing a specific interface) using a Factory class, which has a "create" method, that returns the right instance based on the parameter passed to it (example snippet given below):
public class SimpleFactory {
public static SimpleObjectInterface getSimpleObject(int data) {
SimpleObjectInterface toReturn;
switch(data) {
case 1:
toReturn = new OneSimpleObject();
break;
case 2:
toReturn = new TwoSimpleObject();
break;
default:
toReturn = new DefaultSimpleObject();
break;
}
return toReturn;
}
}
Now I am using Guice in my project for dependency injection. My question is how can I achieve something like the above using Guice? Which implementation instance is needed is decided at runtime based on some user input.
I have looked at Provider and #Named annotations. But I don't understand how exactly it will help me.
In general for the problem where you want a factory that injects most dependencies, but still allows some client-supplied deps, you would use Factories by Assisted Injection.
However in your case this would lead to conditional logic in your factory, which is probably not ideal (it is explicitly discouraged in Guice modules).
I think for your situation a MapBinder would be ideal, and you wouldn't need a factory at all, since you're only switching on data type and not building anything really. In your module you configure a map of int (in your case) keys to impls of SimpleObjectInterface. Then in your main runtime class you inject the map, and when you need an instance of a simple object and have int data available, you call get(data) on the injected map.
I don't have an IDE on this machine, so I can't test the code, but from memory it would be something like below:
In your module:
public class MyModule extends AbstractModule {
protected void configure() {
MapBinder<Integer, SimpleObjectInterface> mapbinder
= MapBinder.newMapBinder(binder(), Integer.class, SimpleObjectInterface.class);
mapbinder.addBinding(1).toClass(OneSimpleObject.class);
mapbinder.addBinding(2).toClass(TwoSimpleObject.class);
}
}
In your app code:
#Inject
private Map<Integer, SimpleObjectInterface> simpleObjectMap;
...
void applicationCode() {
...
Integer data = getData();
SimpleObjectInterface simpleObject = simpleObjectMap.get(data);
...
}
Only issue here is you can't have the "default" binding that you had in your switch statement. Not sure of the best way to handle that, maybe you could assign a default impl in your app code if the object is still null after trying to instantiate it from the map binder. Or you could go back to assisted inject with conditional logic, but it's not really "assisted" injection if the sole dependency is client supplied.
See also: Can Guice automatically create instances of different classes based on a parameter?
So I have a class like so:
public class HBaseUtil {
private final String fileName = "hbase.properties";
private Configuration config;
private HBaseUtil() {
try {
config = new PropertiesConfiguration(fileName);
} catch (ConfigurationException e) {
// some exception handling logging
}
}
// now some getters pulling data out of the config object
public static String getProperty(String fieldKeyName) {...}
public static String getColumnFamily(String fieldName) {...}
// ... some more getters
// NO setters (thus making this a read-only class)
}
Thus, basically I have for myself a Singleton class, that the very first time that it is put to use, sets up a configuration object, and then simply keeps listening for get calls. There are a number of problems with this class:
Unit testing the static methods within class HBaseUtil becomes difficult because of a tight-knit coupling between the Singleton and the configurations file.
What I really want is me being able to supply the filename/filename+path to the class so that it can go in there, read the configuration properties from that file and offer them to incoming read requests. One important note here though: I need this flexibility in specifying the properties file ONLY ONCE per JVM launch. So I certainly don't need to maintain state.
Here is what I was able to come up with:
Instead of a Singleton, I have a normal class with all static methods and no explicit constructor defined.
public class HBaseUtil {
// directly start with getters
public static String getProperty(Configuration config, String fieldKeyName) {...}
public static String getColumnFamily(Configuration config, String fieldKeyName) {...}
// ...and so on
}
And then, instead of using the class in my other code like such:
HBaseUtil.getProperty(String fieldKeyName)
I'd use it like so:
Configuration externalConfig = new PropertiesConfiguration("my-custom-hbase.properties");
HbaseUtil.getProperty(externalConfig, fieldKeyName)
My questions:
Am I even thinking in the right direction? My requirement is to have the flexibility in the class only ONCE per JVM. All that needs to be configurable in my project for this, is the location/contents of the HBase .properties file. I was thinking having a Singleton is overkill for this requirement.
What other better approaches are there for my requirement (stated in above point)?
Thanks!
Note: I've read this StackOverflow discussion, but now it's gotten me even more confused.
You should avoid all static methods and instead design a class which does not mandate its lifecycle: it can be a typical immutable POJO with a public constructor.
Then, when you need it as a singleton, use it as a singleton. For testing, use it in some other way.
Usually, dependency injection is the preferred avenue to solve these problems: instead of hard-coding a pulling mechanism for your configuration object, you have the object delivered to any class which needs it. Then you can decide late what bean you will deliver.
Since you are probably not using Spring (otherwise dependency injection would be your default), consider using Guice, which is a very lightweight and non-intrusive approach to dependency injection.
I'm new to OSGi and I'm interested in retrofitting some of my jars as OSGi bundles.
However I do not want to introduce additional dependencies to any osgi-specific libraries.
As such annotations are out of the question as are programmatic calls to bundle contexts and what not.
I have found a near match to my requirements in declarative services which allows me to expose my lower level bundles without impacting dependencies however at the higher level (where i actually need to consume the services) i'm still a bit stuck.
I understand that the component xml can be used to declare implementations of services (which i already use for my lower level jars) but also to inject service instances into a specific POJO.
Now my question: how do I get access to the osgi-managed POJO which has the services injected into it? Is it at all possible without introducing new dependencies or do I have to do it programmatically?
If the latter is the case can someone point me in the direction of some code to do it, in other words the component-equivalent of bundleContext.getServiceReference()?
UPDATE
To clarify, if you take the fifth part of this tutorial: http://www.vogella.com/articles/OSGiServices/article.html
He declares a component.xml file which uses reference binding to inject a service into the object QuoteConsumer.
Great, now how do I get an instance of QuoteConsumer that has the necessary services injected into it, I can't very well do "new QuoteConsumer()" right?
UPDATE2
Currently I am registering the instance created by osgi as a static variable which can be requested, I'm thinking this is not the best method especially because I can't set the constructor to private. (the latter would at least result in a true singleton)
Basically the Factory class has:
private void activate() {
instance = this;
}
UPDATE3
A full example of a factory:
public class Factory {
private static Factory instance;
public static Factory getInstance() {
if (instance == null)
instance = new Factory();
return instance;
}
private MyInterface implementation;
public void setMyInterface(MyInterface implementation) {
this.implementation = implementation;
}
public void unsetMyInterface(MyInterface implementation) {
implementation = null;
}
public MyInterface getMyInterface() {
if (implementation == null) {
ServiceLoader<MyInterface> serviceLoader = ServiceLoader.load(MyInterface.class);
Iterator<MyInterface> iterator = serviceLoader.iterator();
if (iterator.hasNext())
implementation = iterator.next();
else
implementation = new MyInterfaceStub();
}
return implementation;
}
#SuppressWarnings("unused")
private void activate() {
instance = this;
}
#SuppressWarnings("unused")
private void deactivate() {
instance = null;
}
}
Any client code can then do:
Factory.getInstance().getMyInterface();
and receive the OSGi loaded service, the SPI loaded one or a stub.
You can still manually set the service instance if necessary.
UPDATE4
To clarify further: this pattern is not meant for applications that are designed from the ground up to be run in an OSGi container but rather for low level libraries that have to run everywhere and even when on an OSGi container must not assume that all consumers are actually using OSGi.
You sound confused ... :-) A service is a replacement for static factories so your factory should not have to exist.
The whole idea of DS is that for each component:
wait until its dependencies are met
create an instance
bind the instance to its dependencies
call activate on the instance
register the instance as a service
So whenever you get a service managed by DS it already is injected (bound) with its dependencies. So as long as you stay with service dependencies you never need static factories ... The whole idea of service is that you do NOT have static factories and can only work with (injected) instances. One of the best parts of OSGi is that you rarely work with factories.
One remark about the requirement not to use annotations. The OSGi annotations are class time only, they do not create a runtime dependency. I strongly suggest to use them since they make services as lightweight as a class and are typesafe in contrast to XML.
One trick to use the annotations and not clutter your code is to create extend your implementation classes that you want to be an OSGi component and add the annotations on this class.
To access a service, you declare a reference to it from another component:
#Reference
public void setFoo(Foo foo) {
this.foo = foo;
}
You might find the Bndtools tutorial will help to clarify the concepts.
I'd say you are on the right track. You can use a static field if it is convenient.
The important thing is that you make the rest of your code deal with the QuoteConsumer appearing and disappearing. So, put in your activator the code to do what you need to do when the QuoteConsumer is available (register it in some field, call some initialization code, I don't know) and put in your deactivate the code you need to indicate that the QuoteConsumer is no longer available.
I better explain the question with an example.
I have an Interface Model which can be used to access data.
There can be different implementations of Model which can represent the data in various format say XMl , txt format etc. Model is not concerned with the formats.
Lets say one such implementation is myxmlModel.
Now i want to force myxmlModel and every other implementation of Model to follow Singleton Pattern.The usual way is to make myxmlModels constructor private and provide a static factory method to return an instance of myModel class.But the problem is interface cannot have static method definitions and a result i cannot enforce a particular Factory method definition on all implementation of Model. So one implementation may end with providing getObject() and other may have getNewModel()..
One work around is to allow package access to myxmlModel's constructor and create a Factory class which creates the myxmlModel object and cache it for further use.
I was wondering if there is a better way to achieve the same functionality .
Make a factory that returns
instances of your interface, Model.
Make all concrete implementations of the model package-private classes
in the same package as your factory.
If your model is to be a singleton, and you are using java
5+, use enum instead of traditional
singleton, as it is safer.
public enum MyXMLModel{
INSTANCE();
//rest of class
};
EDIT:
Another possibility is to create delegate classes that do all the work and then use an enum to provide all of the Model Options.
for instance:
class MyXMLModelDelegate implements Model {
public void foo() { /*does foo*/}
...
}
class MyJSONModelDelegate implements Model {
public void foo() { /*does foo*/ }
...
}
public enum Models {
XML(new MyXMLModelDelgate()),
JSON(new MyJSONModelDelegate());
private Model delegate;
public Models(Model delegate) { this.delegate=delegate; }
public void foo() { delegate.foo(); }
}
You can use reflection. Something like this:
public interface Model {
class Singleton {
public static Model instance(Class<? extends Model> modelClass) {
try {
return (Model)modelClass.getField("instance").get(null);
} catch (blah-blah) {
blah-blah
}
}
}
public class XmlModel implements Model {
private static final Model instance = new XmlModel();
private XmlModel() {
}
}
usage:
Model.Singleton.instance(XmlModel.class)
Actually, I don't like this code much :). First, it uses reflection - very slow, second - there are possibilities of runtime errors in case of wrong definitions of classes.
Can you refactor the interface to be an abstract class? This will allow you to force a particular factory method down to all implementing classes.
I used to ask myself the same question. And I proposed the same answer ;-)
Now I normally drop the "forcing" behavior, I rely on documentation.
I found no case where the Singleton aspect was so compelling that it needed to be enforced by all means.
It is just a "best-practice" for the project.
I usually use Spring to instanciate such an object,
and it is the Spring configuration that makes it a Singleton.
Safe, and so easy ... plus additionnal Spring advantages (such as Proxying, substituing a different object once to make some tests etc...)
This is more an answer to your comment/clarification to kts's answer. Is it so, that the real problem is not using the Singleton pattern but instead defining an eclipse (equinox) extension point schema that allows contributing a singleton?
I think, this can't be done, because everytime you call IConfigurationElement.createExecutableExtension you create a new instance. This is quite incompatible with your singleton requirement. And therefore you need the public default constructor so that everybody can create instances.
Unless you can change the extension point definition so that plugins contribute a ModelFactory rather than a model, like
public interface ModelFactory {
public Model getModelInstance();
}
So the extension user will instantiate a ModelFactory and use it to obtain the singleton.
If I guessed wrong, leave a comment and I delete the answer ;)