JDK 8 RFR 6910473: BigInteger range

Joe Darcy joe.darcy at oracle.com
Wed Oct 30 19:35:11 UTC 2013

Hi Brian and Dima,

I've looked over the webrev.

The code changes look fine. From a correctness standpoint, the tests 
look good; however, I still have concerns about the resource usage even 
of the tests that don't try to run with an 8 Gb heap since they do 
allocate a single object that is about 67 Mb.

Given the point in JDK 8 and its test cycle and the variaty of devices 
SE 8 gets run and tested on, I'd be more comfortable making the 
remaining test files not jtreg tests in JDK 8 (e.g. "@test" =" "@ test") 
and filing a follow-up bug to make them jtreg tests in JDK 9 build 01.

These tests are a good example of a more general problem of how to 
select for tests that are resource intensive in memory or cpu usage. 
Enabling these as tests at the start of JDK 9 will leave sufficient time 
to work through implications of having the tests enabled by default.



On 10/30/2013 12:02 PM, Brian Burkhalter wrote:
> This change has been granted CCC approval. The final version of the webrev is here:
> http://cr.openjdk.java.net/~bpb/6910473/webrev.6/
> It would be good to have a definitive JDK 8 Reviewer approval of this, or to resolve any straggling objections should there be any.
> Thanks,
> Brian
> On Oct 24, 2013, at 3:21 PM, Brian Burkhalter wrote:
>> Please review at your convenience.
>> --- Issues ---
>> https://bugs.openjdk.java.net/browse/JDK-6910473
>> https://bugs.openjdk.java.net/browse/JDK-8021203
>> https://bugs.openjdk.java.net/browse/JDK-8021204
>> https://bugs.openjdk.java.net/browse/JDK-8022780
>> ---Webrev ---
>> http://cr.openjdk.java.net/~bpb/6910473/webrev.4/
>> This is more of a "ping" on a thread which was previously labelled "RFC."
>> http://mail.openjdk.java.net/pipermail/core-libs-dev/2013-October/022431.html
>> and for which there is one apparent approval
>> http://mail.openjdk.java.net/pipermail/core-libs-dev/2013-October/022229.html
>> As the CCC request for this is now in its final, almost (but not quite) approved state, it would be good either to definitely approve the webrev or raise objections as to why it should not be approved.
>> Thanks,
>> Brian

More information about the core-libs-dev mailing list