[concurrency-interest] concurrency errors in java.lang.Throwable
gergg at cox.net
Wed Aug 2 13:57:46 EDT 2006
Joe Bowbeer wrote:
> I noticed the same thing a short while ago and discussed it with the
> experts. A summary of that discussion follows.
> We thought it was a good, concise example of a thread-safe one-shot --
> but that it was overkill in this situation...
Typically causes are used in environments that are 'complicated'. Either the
method can't return an exception and it is wrapped and thrown, or a thread
encounters a problem and publishes information about that problem while
continuing to make whatever progress it can.
In either case, I think there is considerable information to be gained from
making sure that the cause is available. I think that there is some merit to
Dougs thoughts that there will eventually be a synchronized piece of code that
runs and forces the visibility.
This is one of the primary issues that developers in MP systems will fight with.
So, education is an important thing. But, I also think that it will be
important for core classes to provide a highlevel of correctness garantees which
will make sure that developers get the right information for diagnosing the
problems that they are having.
More information about the Concurrency-interest