RFR(S) : 8012337 : Change Whitebox implementation to make absence of method in Whitebox.class not fatal

Vladimir Kozlov vladimir.kozlov at oracle.com
Fri Apr 19 13:59:38 PDT 2013


Good.

Thanks,
Vladimir

On 4/19/13 1:45 PM, Igor Ignatyev wrote:
> Vladimir,
> thank you for your hint.
>
> I rewrote the test w/o shell using:
> http://cr.openjdk.java.net/~iignatyev/8012337/webrev.02/
>
> Best regards,
> Igor Ignatyev
>
> On 04/19/2013 02:37 AM, Vladimir Kozlov wrote:
>> Why not use WB technique (ClassFileInstaller) which compiles and copies
>> class file into directory you want so you can avoid using shell?
>>
>> Vladimir
>>
>> On 4/18/13 2:01 PM, Igor Ignatyev wrote:
>>> could someone with "reviewer" status look at that?
>>>
>>> Best regards,
>>> Igor Ignatyev
>>>
>>> On 04/17/2013 05:56 PM, Igor Ignatyev wrote:
>>>> Mikael,
>>>> Thank you for review.
>>>>
>>>> I have updated webrev:
>>>> - added checking for exception class
>>>> - added print warning on unexpected error
>>>>
>>>> http://cr.openjdk.java.net/~iignatyev/8012337/webrev.01/
>>>>
>>>> Best regards,
>>>> Igor Ignatyev
>>>>
>>>> On 04/17/2013 03:22 PM, Mikael Gerdin wrote:
>>>>> Igor,
>>>>>
>>>>> On 2013-04-17 12:11, Igor Ignatyev wrote:
>>>>>> Hi all,
>>>>>>
>>>>>> Please review patch.
>>>>>>
>>>>>> Problem:
>>>>>> If there is some method that presences in native part of whitebox,
>>>>>> but
>>>>>> absences in java side, we get 'j.l.NoSuchMethodError' exception
>>>>>
>>>>> The intended use of is of course that whitebox.cpp and WhiteBox.java
>>>>> are
>>>>> in sync since they are in the same hg repository.
>>>>>
>>>>> I understand that some feel that they need to copy WhiteBox to some
>>>>> other test suite and I find this change acceptable but unfortunate.
>>>>>
>>>>>>
>>>>>> Fix:
>>>>>> Batch registration of WB functions was replaced with one by one
>>>>>> registration
>>>>>>
>>>>>> Testing:
>>>>>> JPRT of sanity tests and tests that use WhiteBox: runtime/interned,
>>>>>> compiler/whitebox
>>>>>>
>>>>>> webrev: http://cr.openjdk.java.net/~iignatyev/8012337/webrev.00/
>>>>>
>>>>> It's too bad that you need a shell script here but I don't see any way
>>>>> around it.
>>>>>
>>>>> I'm ok with you pushing this.
>>>>>
>>>>> /Mikael
>>>>>
>>>>>> jbs: https://jbs.oracle.com/bugs/browse/JDK-8012337


More information about the hotspot-dev mailing list