RFR(XS) 8242384: sa/TestSysProps.java failed due to "RuntimeException: Could not find property in jinfo output: [0.058s][info][cds] Archive was created with UseCompressedOops"
Daniel D. Daugherty
daniel.daugherty at oracle.com
Wed Apr 8 20:23:29 UTC 2020
Also, even though you did not ask for an "RFR(T)", I think this is
a trivial fix. You don't have to wait for 24 hours and can push when
you are happy with your testing.
Dan
On 4/8/20 4:21 PM, Daniel D. Daugherty wrote:
> To reiterate:
>
> > Thumbs up.
>
> Dan
>
> On 4/8/20 4:07 PM, Chris Plummer wrote:
>> Should I go ahead and push this, or do you want to see another review?
>>
>> Chris
>>
>> On 4/8/20 12:31 PM, Chris Plummer wrote:
>>> It's possible, but I think given how our tests are run by CI it will
>>> never happen. Also in order to mess up the test the output would
>>> need to have an '=' in it. In general it's pretty easy to break our
>>> tests by turning on things like logging or sprinkling debugging
>>> printfs the hotspot or library code. I've run into that many times.
>>> So I think we mostly just focus on the tests passing when run in
>>> more "normal" fashion.
>>>
>>> Chris
>>>
>>> On 4/8/20 12:10 PM, Daniel D. Daugherty wrote:
>>>> Is it possible that some logging output that may occur after the
>>>> key string
>>>> ("-- listing properties --") is observed will confuse the test? I
>>>> suspect
>>>> not, but I wanted to ask to be sure...
>>>>
>>>> Thumbs up.
>>>>
>>>> Dan
>>>>
>>>> On 4/8/20 2:54 PM, Chris Plummer wrote:
>>>>> Hello,
>>>>>
>>>>> Please review the following:
>>>>>
>>>>> https://bugs.openjdk.java.net/browse/JDK-8242384
>>>>> http://cr.openjdk.java.net/~cjplummer/8242384/webrev.00/
>>>>>
>>>>> The failure was occurring when running the test with
>>>>> -Xlog:cds=debug, which produced a logging output line that was
>>>>> mistaken for a property. The test now skips all output until it
>>>>> see the start of the properties list.
>>>>>
>>>>> It now passes locally for me when -Xlog:cds=debug is used. I'm
>>>>> still waiting for the tier4 test run to complete.
>>>>>
>>>>> thanks,
>>>>>
>>>>> Chris
>>>>
>>>
>>
>
More information about the serviceability-dev
mailing list