Exception Handling in Singleton

[Bottlenecked]

I have a little problem with the Exception Handling in Singleton. Perhaps I can get help from one of you? I have created the source code of the private constructor can throw an exception. I'm getting now in the declaration line of "test" the error message that said to catch the exception or pass. But how does it work? I also tried the initialization in a static block (then try / catch to make), but then I get the error the "test" may not be initialized. So please anyone knowing some solutions about this, help me.!!

[Alejandro]

You have just mentioned that you are little problem with the Exception Handling in Singleton. If you are trying some coding, then it would be better if you provide your coding.!! Because after checking the coding, it would be easy for us to rectify the errors. Also you said that you have source code of the private constructor which is throwing an exceptions. So no-one can guess what exactly problem is.!! So I insist you to provide the coding for the solution.

[Bottlenecked]

I am extremely sorry for wasting your time because of improper query. Here is my class:

public class TestClass ( 

private static final TestClass test = new TestClass (); 

private TestClass () throws Exception (Test 

// Do something here 

) 

public static TestClass getTestClass () ( 

return test; 

) 

)

Hope that you will able to help me.!!

[Alejandro]

Well, it might go like this, I think but overall not a good idea. In general, the singleton instance during the initial call of getInstance () - initialize method what your problem could in principle solve only once (if getInstance () throws Exception Test "will be declared" with). I could have learned it in detail in recent years, constructors that throw exceptions to avoid if possible and throw the code, the exceptions to outsource to other methods which, in fact, rational treatment (either in class can be done in or in the calling class). What exactly is to happen in your test class in the constructor, which could throw an exception?

[PARRISH]

Classes are controlling their single objects even now an anti-pattern. This has several reasons: first, it violates the separation of concerns, secondly, a lot of technical writing unit tests, and thirdly, the client code for the Singleton virtually un (unit) testable because it binds to the implementation. The fact that a singleton then a Globle variable is, I can see also at a disadvantage. As an alternative has proven Dependency Injection. If you want not just a container (Pico-, nano-, Spring, WHATEVER ...) to shoot at the class, is a viable workaround to introduce a factory, which controls the lifecycle of the instances to be published.

[Deabelos]

I think if the designation of class, method and exception that it is in the example is rather an example of didactic nature as a "serious" implementation. And the approach that such patterns also 'try out manually once outside of "convenient" frameworks like Spring, I think is completely welcomed. Unfortunately withhold 99% of the cellular network to be found in the article stated disadvantages and prevent IMHO an objective assessment, implementation now form one. I'm waiting to be weighed against each other on the support of the GoF book in the alternate implementation forms.

[kyosang]

I do not think that using a framework like Spring or the programmatic use of some Dependency Injection without Framework necessarily an "alternative implementation of the form" is, especially in light of the fact that Spring beans, after all, if not otherwise explicitly declared, even Singletons are. Especially when I look at Spring, I see in many places more of an enormously powerful framework that is based on "established" (?) Models and architectures, as an "alternative form of implementation." To that extent, I think nothing is wrong with books or articles that give readers once the basic foundations to this - which in my view, belongs to stumble again on a hand-built singleton. Otherwise, you may stumble on the Singletons in Spring.