RFR(M): JDK-8165805 - [TESTBUG] CDS needs to support JVMTI CFLH - test development

Mikhailo Seledtsov mikhailo.seledtsov at oracle.com
Thu Sep 15 03:05:28 UTC 2016


I placed the comments below the header.

Misha

On 9/14/16, 7:19 PM, Ioi Lam wrote:
>
>
> On 9/14/16 7:15 PM, Daniel D. Daugherty wrote:
>> On 9/14/16 3:33 PM, Daniel D. Daugherty wrote:
>>> On 9/13/16 11:08 PM, Mikhailo Seledtsov wrote:
>>>> Thank you again for review. I have implemented all the review 
>>>> recommendations.
>>>> The update also contains new test case from Jiangli to test 
>>>> super/sub class pair, where super and sub classes belong to 
>>>> different packages.
>>>> Thank you Jiangli for the test case.
>>>>
>>>> Here is an updated webrev: 
>>>> http://cr.openjdk.java.net/~mseledtsov/8165805.02/
>>>
>>> test/runtime/SharedArchiveFile/serviceability/transformRelatedClasses/TransformInterfaceAndImplementor.java 
>>>
>>>
>>>     L29:  * @requires (vm.opt.UseCompressedOops == null) | 
>>> (vm.opt.UseCompressedOops == true)
>>>     L30:  * @requires vm.flavor != "minimal"
>>>         Should there be a comment somewhere explaining why this test
>>>         can't be run with -XX:-UseCompressedOops or with the minimal 
>>> VM?
>>>         I'm not completely up to speed to @requires style...
>>
>> You can pretty much ignore this comment. I just remembered that
>> JTREG header format doesn't easily allow you to add comments
>> above or below the "@foo" items...
>>
> I concur. A @comment tag would be nice to add to jtreg. Maybe 
> something like this?
>
> /*
>  * @comment blah blah blah
>  * @requires vm.flavor != "minimal"
>  *
>  * @comment blah blah blah
>  * @requires (vm.opt.UseCompressedOops == null) | 
> (vm.opt.UseCompressedOops == true)
>  */
>
>
> Or maybe
>
> /*
>  * @@ blah blah blah
>  * @requires vm.flavor != "minimal"
>  *
>  * @@ blah blah blah
>  * @requires (vm.opt.UseCompressedOops == null) | 
> (vm.opt.UseCompressedOops == true)
>  */
>
>
> - Ioi
>
>
>> Dan
>>
>>
>>>
>>>
>>>>
>>>> Thank you,
>>>> Misha
>>>>
>>>> On 9/9/16, 7:31 PM, Mikhailo Seledtsov wrote:
>>>>>     Please review the tests that accompany the changes for
>>>>>     "JDK-8078644 - CDS needs to support JVMTI CFLH"
>>>>>
>>>>>     The tests include several test cases that are produced by 
>>>>> combination
>>>>>     of transforming related classes: super/sub class and 
>>>>> interface/implementor.
>>>>>     The classes reside in CDS archive. Tests check to ensure that 
>>>>> correct
>>>>>     transformation is performed, classes are loaded from the 
>>>>> archive only when appropriate,
>>>>>     and that test process using the archive completes w/o 
>>>>> errors/crashes.
>>>>>
>>>>>
>>>>>     Details:
>>>>>
>>>>>     The main class is TransformRelatedClasses.java; start review 
>>>>> with this class.
>>>>>     Other classes are:
>>>>>       - TransformSuperAndSubClasses.java, 
>>>>> TransformInterfaceAndImplementor.java
>>>>>         the "actual tests", but really just test headers calling 
>>>>> into TransformRelatedClasses.java
>>>>>
>>>>>       - TransformTestCommon.java - common methods for 
>>>>> transformation tests
>>>>>
>>>>>       - TestEntry.java - an entry into combinatorial test table 
>>>>> representing a single test case
>>>>>
>>>>>       - CDSTestUtils.java, TransformUtil.java - utility classes 
>>>>> that are common  CDS
>>>>>         tests
>>>>>
>>>>>       - TransformerAgent.java - agent that performs transformation 
>>>>> (comes with .mf file)
>>>>>
>>>>>
>>>>>     JBS: https://bugs.openjdk.java.net/browse/JDK-8165805
>>>>>     Webrev: http://cr.openjdk.java.net/~mseledtsov/8165805.00/
>>>>>     Testing:
>>>>>         Ran these tests on a standard set of platforms thru 
>>>>> automated test system.
>>>>>
>>>>> Thank you,
>>>>> Misha
>>>>>
>>>
>>>
>>
>


More information about the hotspot-runtime-dev mailing list