Looks fine AFAICT Stephen On 17 Dec 2013 09:23, "roger riggs" <roger.ri...@oracle.com> wrote:
> Hi Sherman, > > Thanks for the comments, corrected and updated the webrev: > http://cr.openjdk.java.net/~rriggs/webrev-time-serialization/ > > On 12/16/2013 1:00 PM, Xueming Shen wrote: > >> On 12/16/2013 09:02 AM, roger riggs wrote: >> >>> Please review these changes to java.time serialization. >>> The format of the serialized data is unchanged; deserialization >>> uses readObject instead of readResolve to flag invalid values. >>> >>> Webrev: >>> http://cr.openjdk.java.net/~rriggs/webrev-time-serialization/ >>> >>> Thanks, Roger >>> >>> || * >> *(1) test/java/time/tck/java/time/serial/TCKLocalDateSerialization.java* >> * >> -- import not necessary? >> > removed extra imports > >> >> (2) AbstractTCKTest.assertNotSerializable() >> >> -- the problem is that this test fails with current readResolve() >> implementation, as we chatted last Friday. So it might be hard to >> say it's a bulletproof test. But I guess it might not be worth to >> a proof there would be a loophole here if only define the readResolve() >> to throw the IOE. >> > Thanks, Yes, the test(s) should be more liberal and catch Exception instead > of only InvalidObjectException since any exception during deserialization > terminates serialization. > > Thanks, Roger > >