RFR(s): 8170860: Get rid of the humanReadableByteCount() method in openjdk/hotspot

Dmitry Fazunenenko dmitry.fazunenko at oracle.com
Thu Dec 8 14:21:32 UTC 2016


Mikael,

 > Let's just go with your original webrev and perhaps file a follow-up 
RFE for factoring  out
 > printMemoryUsage in 10?

Sounds good.  I submitted 
https://bugs.openjdk.java.net/browse/JDK-8170923 for that change.

Thanks,
Dima

On 08.12.2016 17:00, Mikael Gerdin wrote:
> Hi Dima,
>
> On 2016-12-08 14:41, Dmitry Fazunenenko wrote:
>> Hi Mikael,
>>
>> On 08.12.2016 16:02, Mikael Gerdin wrote:
>>> Hi Dima,
>>>
>>> On 2016-12-08 13:47, Dmitry Fazunenenko wrote:
>>>> Hello,
>>>>
>>>> I'm looking for a couple of reviewers for a simple fix which 
>>>> eliminates
>>>> the code copied from Stack Overflow:
>>>>
>>>> http://cr.openjdk.java.net/~dfazunen/8170860/webrev.00/
>>>> https://bugs.openjdk.java.net/browse/JDK-8170860
>>>>
>>>> humanReadableByteCount() method was removed. A standard JDK class
>>>> NumberFormat is used instead. To make the output locale independent 
>>>> '_'
>>>> symbol is explicitly set as the group delimiter. New output will look
>>>> like:
>>>>
>>>> [allocate #1             ] init: 33_554_432, used: 5_253_664, comm:
>>>> 8_388_608, freeRatio ~= 37.4%
>>>> [allocate #2             ] init: 33_554_432, used: 8_399_392, comm:
>>>> 10_485_760, freeRatio ~= 19.9%
>>>> [allocate #3             ] init: 33_554_432, used: 11_545_120, comm:
>>>> 13_631_488, freeRatio ~= 15.3%
>>>> [allocate #4             ] init: 33_554_432, used: 14_690_848, comm:
>>>> 16_777_216, freeRatio ~= 12.4%
>>>> [allocate #5             ] init: 33_554_432, used: 17_836_576, comm:
>>>> 19_922_944, freeRatio ~= 10.5%
>>>> [allocate #6             ] init: 33_554_432, used: 20_982_304, comm:
>>>> 23_068_672, freeRatio ~= 9.0%
>>>> [allocate #7             ] init: 33_554_432, used: 24_128_032, comm:
>>>> 26_214_400, freeRatio ~= 8.0%
>>>> [allocate #8             ] init: 33_554_432, used: 27_273_760, comm:
>>>> 29_360_128, freeRatio ~= 7.1%
>>>
>>> This looks good but I'm wondering if it would make sense to move
>>> MemoryUsagePrinter to gc.testlibrary.Helpers, API could be something
>>> like:
>>>
>>> void printMemoryUsage(String label, MemoryUsage memusage) {
>>> ...
>>> }
>>
>> My original intention was moving the entire MemoryUsagePrinter class to
>> Helpers :)
>> But I thought that change will be more than just fixing the problem with
>> humanReadableByteCount() code.
>> And it would be rather an RFE.
>
> I see, that is probably an RFE.
> Let's just go with your original webrev and perhaps file a follow-up 
> RFE for factoring out printMemoryUsage in 10?
>
> Thanks
> /Mikael
>
>> But if nobody objects I will be glad to implement the bigger refactoring
>> now.
>>
>> I see new API as:
>>
>>    public static String formatNumber(Number n);
>>    public static void printMemoryUsage(String label, MemoryUsage mu);
>>
>> Thanks,
>> Dima
>>>
>>>
>>> If you feel that you have more urgent matters to work on then I'm fine
>>> with the change you proposed as well.
>>
>>
>>
>>>
>>> Thanks
>>> /Mikael
>>>
>>>>
>>>> Thanks,
>>>> Dima
>>




More information about the hotspot-gc-dev mailing list