[concurrency-interest] static initializers and threads

teknokrat teknokrat at yahoo.com
Thu May 25 08:12:05 EDT 2006

thanks a lot. that cleared things up.

Bart Jacobs wrote:
> The answer is in Section 12.4 of the Java Language Specification 
> <http://java.sun.com/docs/books/jls/third_edition/html/execution.html#12.4>. 
> In fact, T2 waits for the initialization to finish before executing the 
> constructor of A.
> There's no need to synchronize the static block. As you can read in 
> Section 12.4, the virtual machine performs the necessary synchronization 
> automatically.
> The only way to get a new object of A before the completion (normal or 
> abrupt) of the static initializer is if the object is created from 
> within the static initializer itself (directly or indirectly), i.e. by 
> the thread that is executing the static initializer. But note that this 
> includes nested class initializations; for example, if the static 
> initializer of A triggers initialization of some class B, which in turn 
> creates an object of A, then this object creation succeeds and you get 
> an object of A prior to completion of the static initializer of A. 
> However, note that such a circularity could also cause a deadlock, 
> specifically if some other thread is already initializing B, so this 
> would be buggy code to write in any case.
> (Note that the static block should of course be inside the class body.)
> Bart
> teknokrat wrote:
>> This a question on static initializers in Java. The answer may be in the 
>> language spec but I have not been able to find it. I have the following 
>> class;
>> static {
>> 	// do some initialization
>> }
>> class A {
>> 	public A(){}
>> }
>> I have two threads T1, T2 executing
>> A a = new A();
>> Assuming T1 enters the static initializer first, what happens to T2? 
>> Does it block waiting for the initialization to finish before being 
>> given a reference to a?
>> Should the static block be synchronized? Are there any conditions under 
>> which its possible to get a new object a but with the initialization 
>> incomplete?
>> thanks
>> _______________________________________________
>> Concurrency-interest mailing list
>> Concurrency-interest at altair.cs.oswego.edu
>> http://altair.cs.oswego.edu/mailman/listinfo/concurrency-interest
> Disclaimer: http://www.kuleuven.be/cwis/email_disclaimer.htm

More information about the Concurrency-interest mailing list