RFR (XS) : 8027124 : [TESTBUG] NonTieredLevelsTest: java.lang.RuntimeException: private TestCase$Helper(java.lang.Object) must be osr_compiled

Vladimir Kozlov vladimir.kozlov at oracle.com
Sun Dec 8 09:43:51 PST 2013


Good.

Vladimir

On 12/8/13 8:16 AM, Igor Ignatyev wrote:
> Vladimir,
>
> Thanks for review, I've added CompilerWhiteBoxTest::skipXcompOSR(), see updated webrev:
> http://cr.openjdk.java.net/~iignatyev/8027124/webrev.01/
>
> I reran compiler/tiered and compiler/whitebox tests in Xcomp and Xmixed modes
>
> On 12/03/2013 11:52 AM, Vladimir Kozlov wrote:
>> On 12/2/13 11:35 PM, Igor Ignatyev wrote:
>>> Vladimir,
>>>
>>>  > Why space needed "ed "?
>>> just to make sure that it's the begin of 'compiled mode'
>>
>> Okay.
>>
>>>
>>>  > I don't think we should pollute output with messages which does not
>>>  > help. Or this message is used to mark test passed?
>>>  >
>>> no, it doesn't mark test as passed, it's just a warning message. I
>>> will wrap it  w/ 'if (IS_VERBOSE)' statement.
>>
>> Okay.
>>
>>>
>>> It's the same code which is used in 'compiler/whitebox' tests (fix for
>>> JDK-8023452), so would you prefer me to change
>>> they in a similar way?
>>
>> Yes, please. Can you move checks and warning message into a separate
>> method and use it everywhere? Something like:
>>
>> if (CompilerWhiteBoxTest.skipTest(testCase)) {
>>    return;
>> }
>>
>> Thanks,
>> Vladimir
>>
>>>
>>> Thanks,
>>> Igor
>>>
>>> On 12/03/2013 01:30 AM, Vladimir Kozlov wrote:
>>>> Don't split the line:
>>>>
>>>> +         if (testCase.isOsr && CompilerWhiteBoxTest.MODE.startsWith(
>>>> +                 "compiled ")) {
>>>>
>>>> Why space needed "ed "?
>>>>
>>>> I don't think we should pollute output with messages which does not
>>>> help. Or this message is used to mark test passed?
>>>>
>>>> Add comment.
>>>>
>>>> thanks,
>>>> Vladimir
>>>>
>>>>
>>>> On 12/2/13 12:39 PM, Igor Ignatyev wrote:
>>>>> Hi all,
>>>>>
>>>>> Please review patch.
>>>>>
>>>>> Problem:
>>>>> OSR test cases in 'compiler/tiered' tests are not applicable in -Xcomp
>>>>> mode, since there is no way to provoke OSR compilation
>>>>>
>>>>> Fix:
>>>>> Added skipping of OSR test cases, if -Xcomp is enabled
>>>>>
>>>>> webrev: http://cr.openjdk.java.net/~iignatyev/8027124/webrev.00/
>>>>> jbs: https://bugs.openjdk.java.net/browse/JDK-8027124
>>>>> testing: compiler/tiered in -Xcomp, -Xmixed, -Xint and default mode


More information about the hotspot-compiler-dev mailing list