RFR(XXXS): 8193407: jdk/hs fails Solaris slowdebug test-image build

Daniel D. Daugherty daniel.daugherty at oracle.com
Wed Dec 13 14:01:03 UTC 2017


On 12/13/17 4:01 AM, Erik Joelsson wrote:
> On 2017-12-13 03:19, Daniel D. Daugherty wrote:
>> On 12/12/17 9:16 PM, David Holmes wrote:
>>> Hi Dan,
>>>
>>> On 13/12/2017 12:13 PM, Daniel D. Daugherty wrote:
>>>> Greetings,
>>>>
>>>> I have a 1-liner fix to solve a Solaris slowdebug test-image build
>>>> problem:
>>>>
>>>>      JDK-8193407 jdk/hs fails Solaris slowdebug test-image build
>>>>      https://bugs.openjdk.java.net/browse/JDK-8193407
>>>>
>>>> Here is the context diff:
>>>>
>>>> $ hg diff
>>>> diff -r a576e1b6784d make/test/JtregNativeHotspot.gmk
>>>> --- a/make/test/JtregNativeHotspot.gmk    Tue Dec 12 14:14:06 2017 
>>>> -0500
>>>> +++ b/make/test/JtregNativeHotspot.gmk    Tue Dec 12 21:07:03 2017 
>>>> -0500
>>>> @@ -113,6 +113,7 @@
>>>> BUILD_HOTSPOT_JTREG_LIBRARIES_LIBS_libHandshakeTransitionTest := -lc
>>>>       BUILD_HOTSPOT_JTREG_LIBRARIES_LIBS_libHasNoEntryPoint := -lc
>>>>       BUILD_HOTSPOT_JTREG_LIBRARIES_LIBS_libReturnError := -lc
>>>> +    BUILD_HOTSPOT_JTREG_LIBRARIES_LIBS_libCNLookUp := -lc
>>>>   endif
>>>>
>>>>   ifeq ($(OPENJDK_TARGET_OS), linux)
>>>>
>>>>
>>>> David Holmes suggested the possible fix and I have tested the fix with
>>>> Solaris X64 'release', 'fastdebug' and 'slowdebug' builds.
>>>
>>> Thanks for that!
>>>
>>>> David is the first (R)eviewer. Since this affects a Compiler team
>>>> test, it would be good to hear from someone on the Compiler team.
>>>> Since it is a Makefile change, it would be good to hear from someone
>>>> on the build-dev at ... alias...
>>>
>>> I think that is a bit of overkill. I was going to say push this 
>>> under trivial rules!
>>
>> No problem. I'm just trying to be polite... :-)
>>
>> I'll go ahead and push it shortly... I'll wait long enough to clear my
>> email queue and we'll see if anyone else chimes in... :-)
>>
> I agree that it's trivial, but appreciate the cross post to build-dev 
> so that we are in the loop.

Thanks for the review!

Dan


>
> /Erik
>>
>>> Aside: I've started a separate email thread on why we need to list 
>>> -lc explicitly in the makefile.
>>
>> Yup. It would be good to get rid of that if we can...
>>
>> Dan
>>
>>
>>>
>>> Thanks,
>>> David
>>>
>>>> Thanks, in advance, for questions, comments, or suggestions.
>>>>
>>>> Dan
>>>>
>>
>



More information about the build-dev mailing list