<i18n dev> [8u] RFR: 8177776: Create an equivalent test case for JDK9's SupplementalJapaneseEraTest

Naoto Sato naoto.sato at oracle.com
Thu Apr 6 20:29:01 UTC 2017

Hi Brent, thank you for the review.

On 4/6/17 1:08 PM, Brent Christian wrote:
> Hi, Naoto
> On 4/5/17 2:14 PM, Naoto Sato wrote:
>> I revised the test case not to rely on shell script.
> Yay!  Hopefully this can also happen sometime for JDK 9+.

Sure. Will work on it.

>> http://cr.openjdk.java.net/~naoto/8177776/webrev.01/
> Looks fine to me, Naoto.  A few comments:
> * I presume additional @bug values will be added as other fixes are
> backported (e.g. 8054214).

Good point. Will add those bug ids on push.

> * On
>   73         Path dst = Paths.get("testjava").toAbsolutePath();
> This places "dst" within the scratch directory, then?  (I thought there
> was a jtreg system property for the scratch directory, but I can't find
> it so I think I'm mis-remembering).
> I agree with letting jtreg take care of cleaning up "scratch".

Yes, "dst" will be "scratch". Not aware of any system property to 
designate the output directory, other than "-w" command option to 
specify the JTwork dir.

> If anything goes wrong with copying of the JDK (e.g. full disk),
> hopefully it would be easy to diagnose, with an IOException with a full
> stack trace.
> * Have you confirmed (if it's practical to do so) that this test fails
> when expected (detects a bug)?

Yes. Without those backports, the test will fail as expected.


> Thanks,
> -Brent
>> On 3/30/17 2:10 PM, Naoto Sato wrote:
>>> Hello,
>>> Please review the changes to the following issue:
>>> https://bugs.openjdk.java.net/browse/JDK-8177776
>>> The proposed change is located at:
>>> http://cr.openjdk.java.net/~naoto/8177776/webrev.00/
>>> This is for backporting fixes for JapaneseEra related issues (8054214,
>>> 8173423). The original fixes in JDK9 included a test case,
>>> SupplementalJapaneseEraTest which is intended for the system property
>>> testing. The above proposed fix intends to adapt that test case into
>>> JDK8, where calendars.properties file is used instead of the system
>>> property. The test is pretty much identical to JDK9's. The difference is
>>> to deal with the calendars.properties file and removed some
>>> non-applicable cases in JDK8.
>>> Naoto

More information about the i18n-dev mailing list