[7u-dev] Request for approval for CR 7132378: Race in FutureTask if used with explicit set and get ( not Runnable ) (original) (raw)

Aleksey Shipilev aleksey.shipilev at oracle.com
Thu Jun 6 06:48:09 PDT 2013

Thanks Sean,

On 06/06/2013 05:39 PM, Seán Coffey wrote:

The code comments indicate that some behavioural change may be present with this fix :

+ * Revision notes: This differs from previous versions of this + * class that relied on AbstractQueuedSynchronizer, mainly to + * avoid surprising users about retaining interrupt status during + * cancellation races.

I think this is the corner case within the bounds of legal behavior for the implementation, as mentioned by Doug here:

http://mail.openjdk.java.net/pipermail/core-libs-dev/2012-January/009032.html

Also, there is no CCC for JDK 8 mentioning this change.

Was any further testing carried out other than FutureTask jtreg tests ?

I did the targeted test for FutureTask (similar to what the original bugreport was mentioned). Otherwise, I was relying on this code being tested within the JDK 8.

-Aleksey.

More information about the jdk7u-dev mailing list