Code coverage statistics for OpenJDK

Tue Feb 3 08:49:23 UTC 2015

Hi,

There are links to jtreg tutorials off the quality wiki :
https://wiki.openjdk.java.net/display/quality/Tutorials

Rgds,Rory
On 03/02/2015 07:22, Mani Sarkar wrote:
> There's also a simple example of how to write your first test (OpenJDK) for
> a Java class in the same document I mentioned earlier.
>
> Its worth going through the index and updating yourself about various
> topics thereunder.
>
> Cheers,
> Mani
>
> On Tue, Feb 3, 2015 at 7:07 AM, Mani Sarkar <sadhak001 at gmail.com> wrote:
>
>> HI Richard,
>>
>> Great question - I don't the figures of how much of these are community
>> contributed but OpenJDK has been public since sometime and I believe there
>> would be contributions from the community. A couple of years ago we had
>> TestFest in London and a number of us contributed to the test suite.
>>
>> Test patches are always welcome as long as they don't duplicate another
>> test, bring value and exhibit good coverage - above all written in the
>> OpenJDK protocol (using JTReg, TestNG, etc...).
>>
>> To get started have a look at the tests, first get familiar with the test
>> suites in the respective repos:
>>
>> │   ├── ./hotspot/test
>> │   ├── ./jaxp/test
>> │   ├── ./jdk/test
>> │   ├── ./langtools/test
>> │   ├── ./nashorn/test
>>
>> (Not all repos have a test folder)
>>
>> I have also an detailed section *How to use JTReg… - Java Regression Test
>> Harness* on tests in the Adopt OpenJDK Intermediate & Advance experiences
>> document: http://bit.ly/1ckphOl,
>>
>> Use both these are a route to get started.
>>
>> Also there's a thread on how to get JCov coverage when running JTReg tests
>> -
>> http://mail.openjdk.java.net/pipermail/adoption-discuss/2015-January/000542.html
>>
>> You will need to make sure you download the latest JTReg built on our
>> Cloudbees farm, see
>> https://adopt-openjdk.ci.cloudbees.com/view/OpenJDK%20code-tools/, this
>> build has jcov embedded in it. This along with the above changes to your
>> jdk/test/Makefile should give you a local build with coverage reports.
>>
>> Identify a class or package that is simple and interesting ==> look for
>> all the relevant test suites ==> check the coverage and find the gaps, if
>> any ==> write test(s) ==> get coverage report by running tests (make test)
>> ==> iterate!
>>
>> When your test patches are ready for submission, announce here and we can
>> help you further with your submissions (there is a section in the guide
>> about patch submissions as well).
>>
>> I hope this helps.
>>
>> Regards,
>> Mani
>>
>> On Tue, Feb 3, 2015 at 5:45 AM, Richard Kolb <rjdkolb at gmail.com> wrote:
>>
>>> Hi Mani,
>>>
>>> How does one get involved in something this awesome ?
>>>
>>> What percentage of these tests are written by the community ?
>>>
>>> thanks,
>>> Richard.
>>>
>>> On 3 February 2015 at 00:26, Mani Sarkar <sadhak001 at gmail.com> wrote:
>>>
>>>> And it can only get better with everyone's involvement. ;)
>>>>
>>>> On Mon, Feb 2, 2015 at 12:52 PM, Richard Kolb <rjdkolb at gmail.com> wrote:
>>>>
>>>>> Awesome results for a non tdd program!
>>>>>   On 1 Feb 2015 12:45, "Martijn Verburg" <martijnverburg at gmail.com>
>>>>> wrote:
>>>>>
>>>>>> Hi all,
>>>>>>
>>>>>> Recently, one of our Adopt OpenJDK incubator projects (for reference,
>>>>> these
>>>>>> projects are outside of OpenJDK proper, the wiki holds the full list)
>>>>>> managed to get some what we think are accurate code coverage stats for
>>>>>> OpenJDK (jdk9-dev) tests. Thanks to Jonathan Gibbons from code-tools
>>>>>> (jtreg/jcov) and Adopt's John Oliver for getting this out!
>>>>>>
>>>>>> I think this is potentially useful for the OpenJDK quality group to
>>>>> report
>>>>>> alongside the existing weekly tests passing that Balchandra kindly
>>>>> pushes
>>>>>> out.  It can also be useful to OpenJDK contributors to have a guide on
>>>>>> making a change -> writing a test -> seeing code coverage improve.
>>>>>>
>>>>>> Obviously we want to:
>>>>>>
>>>>>> * Make sure the numbers are correct.
>>>>>>
>>>>>> * Make it clear in the report that this does not represent how well
>>>>> OpenJDK
>>>>>> / Java is actually tested (internally Oracle and others run a far more
>>>>>> comprehensive test suite).
>>>>>>
>>>>>> Do people feel this is this a good idea? If so, who's the right
>>>>>> person/people to analyse our results and ensure we're using jtreg and
>>>>> jcov
>>>>>> correctly?
>>>>>>
>>>>>> Cheers,
>>>>>> Martijn
>>>>>>
>>>>
>>>>
>>>> --
>>>> @theNeomatrix369 <http://twitter.com/theNeomatrix369>*  |  **Blog
>>>> <http://neomatrix369.wordpress.com>**  |  *LJC Associate & LJC Advocate
>>>> (@adoptopenjdk & @adoptajsr programs)
>>>> *Meet-a-Project - *MutabilityDetector
>>>> <https://github.com/MutabilityDetector>*  |  **Bitbucket
>>>> <https://bitbucket.org/neomatrix369>* * |  **Github
>>>> <https://github.com/neomatrix369>* * |  **LinkedIn
>>>> <http://uk.linkedin.com/pub/mani-sarkar/71/a77/39b>*
>>>> *Come to Devoxx UK 2015:* http://www.devoxx.co.uk/
>>>>
>>>> *Don't chase success, rather aim for "Excellence", and success will come
>>>> chasing after you!*
>>>>
>>>
>>
>> --
>> @theNeomatrix369 <http://twitter.com/theNeomatrix369>*  |  **Blog
>> <http://neomatrix369.wordpress.com>**  |  *LJC Associate & LJC Advocate
>> (@adoptopenjdk & @adoptajsr programs)
>> *Meet-a-Project - *MutabilityDetector
>> <https://github.com/MutabilityDetector>*  |  **Bitbucket
>> <https://bitbucket.org/neomatrix369>* * |  **Github
>> <https://github.com/neomatrix369>* * |  **LinkedIn
>> <http://uk.linkedin.com/pub/mani-sarkar/71/a77/39b>*
>> *Come to Devoxx UK 2015:* http://www.devoxx.co.uk/
>>
>> *Don't chase success, rather aim for "Excellence", and success will come
>> chasing after you!*
>>
>
>

-- 
Rgds,Rory O'Donnell
Quality Engineering Manager
Oracle EMEA , Dublin, Ireland