[concurrency-interest] Propagation of signals to non-interrupted thread
davidcholmes at aapt.net.au
Mon Nov 14 17:15:28 EST 2011
Actually I was referring to the spec around wait/notify not Condition.await.
Note that "Implementation Consideration" is not the same as an
"Implementation Note" - this is still very much a part of the spec. These
items are called out specifically to the attention of implementors because
they need to think about them and if necessary document them.
From: Martin Buchholz [mailto:martinrb at google.com]
Sent: Monday, 14 November 2011 1:31 PM
To: dholmes at ieee.org
Cc: David M. Lloyd; Dr Heinz M. Kabutz; concurrency-interest at cs.oswego.edu
Subject: Re: [concurrency-interest] Propagation of signals to
The propagation here is nothing to do with poorly written code but a
work-around for lack of detail in the spec prior to JDK 5 that allowed
interrupted thread to consume a notification.
I rediscovered this spec:
* <p>An implementation can favor responding to an interrupt over
* method return in response to a signal. In that case the
* must ensure that the signal is redirected to another waiting
* there is one.
but this is an "Implementation Consideration" and it could be clearer that
a thread returning abnormally from await cannot cause a lost signal.
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Concurrency-interest