[concurrency-interest] DefaultThreadFactory implementation questions

Roel Spilker R.Spilker at topdesk.com
Wed Jan 14 10:52:37 EST 2009


In the class j.u.c.Executors, there is a package private static class DefaultThreadFactory. This class synthesizes a name for all created threads. For debugging and monitoring purposes, I would like to be able to provide a custom namePrefix. This has lead me to two questions:
1) Is there a reason the ThreadGroup is determined in the constructor and the later used to create a new Thread? As far as I can see, Thread also has a constructor that takes a Runnable and a String name that would use similar code to determine the ThreadGroup. 
2) Do other people also feel the need to provide a name? Or is there another way to provide more debugging and monitoring information? If so, is it appropriate to add some API to do so? 
I think the API changes could be small. The smallest change would be to add two static methods to Executors:
public static ThreadFactory defaultThreadFactory(String namePrefix);
public static ThreadFactory privilagedThreadFactory(String namePrefix);
This would be enough, since the code in j.u.g that call defaultThreadFactory also have a counterpart that accepts a provided ThreadFactory. 
A more complete API would be to also add: 
public static ExecutorService newFixedThreadPool(int nThreads, String namePrefix); 
public static ExecutorService newSingleThreadExecutor(String namePrefix); 
public static ExecutorService newCachedThreadPool(String namePrefix); 
... and probably the same for the ScheduledExecutorService factory methods.
Also, ThreadPoolExecutor could get two extra constructors that accept a namePrefix.
The changes in the code would, for as far as I can see, be very small.
Roel Spilker

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://cs.oswego.edu/pipermail/concurrency-interest/attachments/20090114/1abe4785/attachment.html>

More information about the Concurrency-interest mailing list