RFR(S) : 8234304 : runtime/cds/appcds/javaldr tests should use driver mode

Igor Ignatyev igor.ignatyev at oracle.com
Mon Nov 18 22:14:13 UTC 2019


Ioi, Misha,

thanks for your review, pushed.

-- Igor

> On Nov 18, 2019, at 2:11 PM, mikhailo.seledtsov at oracle.com wrote:
> 
> +1
> 
> On 11/18/19 12:50 PM, Ioi Lam wrote:
>> Looks good to me. Thanks for fixing this.
>> 
>> - Ioi
>> 
>> On 11/18/19 11:58 AM, Igor Ignatyev wrote:
>>> http://cr.openjdk.java.net/~iignatyev//8234304/webrev.00/index.html
>>>> 30 lines changed: 0 ins; 5 del; 25 mod;
>>> Hi all,
>>> 
>>> could you please review this small patch which replaces 'main/othervm' w/ 'driver' in runtime/cds/appcds/javaldr tests?
>>> from JBS:
>>>> runtime/cds/appcds/javaldr tests do actual testing in child processes, the main test class makes preparations, starts JDK under test, and verifies its results. so there is no need for the main test class to be run in JDK under test w/ all external flags.
>>> the patch also replaces "X" w/ X.class.getName() in all the places these tests refer to a class, which made explicit @build action unnecessary.
>>> 
>>> webrev: http://cr.openjdk.java.net/~iignatyev//8234304/webrev.00/index.html
>>> JBS: https://bugs.openjdk.java.net/browse/JDK-8234304
>>> testing:
>>>   - runtime/cds/appcds/javaldr on macosx-x64 w/ all flags used in Oracle CI
>>>   - runtime/cds/appcds/javaldr on linux-x64,windows-x64,macosx-x64,solaris-sparcv9 in OOB
>>>   - runtime/cds/appcds/javaldr one by one on macosx-x64
>>> 
>>> Thanks,
>>> -- Igor
>> 



More information about the hotspot-runtime-dev mailing list