RFR: 8229899: java.io.File.isInvalid() is racy

Aleksey Shipilev shade at redhat.com
Tue Aug 20 07:56:41 UTC 2019


On 8/20/19 1:14 AM, Arthur Eubanks wrote:
> Bug: https://bugs.openjdk.java.net/browse/JDK-8229899
> Webrev: http://cr.openjdk.java.net/~aeubanks/8229899/webrev.00/

Right, the new code turns data race benign. In older code, the first read of "status" could have
returned "!null", while the second read could have returned "null", breaking the check. Dusty little
corner in Java memory model, that.

Looks good.

-- 
Thanks,
-Aleksey



More information about the core-libs-dev mailing list