Code review request: 8006564: Test sun/security/util/Oid/S11N.sh fails with timeout on Linux 32-bit

[Wang Weijun]

I've thought about that, but it means there is no way to check forward-compatibility. (Am I using the correct word? Deserialize jdk7 format with jdk6).

Max

在 2013-1-21，下午9:58，Alan Bateman <Alan.Bateman at oracle.com> 写道：

> On 21/01/2013 01:51, Weijun Wang wrote:
>> Please take a look at the webrev at
>> 
>>   http://cr.openjdk.java.net/~weijun/8006564/webrev.00/
>> 
>> Bug is
>> 
>>   http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=8006564
>> 
>> The S11N.sh test uses /java to try to access old JDKs. When the test is running at Russia, /java exists but is on the other side of the globe, and a timeout happens.
>> 
>> This fix has created a jdk6-style class and use it to check for compatibility. Not precise but enough for this test.
>> 
>> Thanks
>> Max
> It's good to see the dependency on /java/re removed but I'm not sure that I like the -Xbootclasspath/o:old because it is too tied into the implementation.
> 
> Have you considered just put the serial versions of the older releases into a byte array and have test use that? It would of course require having a way to easily re-create the byte array (should be clear instructions).
> 
> -Alan
>