Hi Christoph, Not a locale expert here so you probably need to hear from someone else, but this looks like a reasonable approach to me. One minor comment is to perhaps use format() at lines 67-68 instead of println().
Thanks, Brian > On Mar 18, 2019, at 7:51 AM, Langer, Christoph <christoph.lan...@sap.com> > wrote: > > please review a fix for 8172695, where java/util/Scanner/ScanTest.java fails. > It reproducibly fails on one of our Mac servers, where the locale is set to > en_DE. > > The test in its current form checks if the test VM is run with a supported > default Locale by just checking the language to be one of “en”, “zh”, “ko”, > “ja”. But it turns out, that also the country part seems to be important. I > suggest to change the test case to test whether the VM’s default locale is > one of a restricted set of locales and if it isn’t, it shall try to run with > “ENGLISH”. > > I tested the set of US, UK, CHINA, KOREA, JAPAN and GERMANY to work on my > machine. Before pushing, I’d run it through the submit repo. > > bug: https://bugs.openjdk.java.net/browse/JDK-8172695 > <https://bugs.openjdk.java.net/browse/JDK-8172695> > webrev: http://cr.openjdk.java.net/~clanger/webrevs/8172695.0/ > <http://cr.openjdk.java.net/~clanger/webrevs/8172695.0/>
