[11u] 8215913: [Test_bug]java/util/Locale/LocaleProvidersRun.java failed on de_DE and ja_JP locale.

[Severin Gehwolf]

Hi Christoph,

On Tue, 2019-10-01 at 09:03 +0000, Langer, Christoph wrote:
> Hi Severin,
> 
> thanks for doing this backport. I think we should definitely do this for keeping parity.

Maybe. It really is just adding a if (isWindows()) {...} conditional
around the test code.

> Your webrev looks fine but doesn't contain the change below to
> LocaleProviders.sh. This should definitely be done since it is an
> important part of the original change although being part of
> LocaleProvidersRun.java instead of LocaleProviders.sh. Can you update
> the webrev? Then I can put it into our queue for regression testing.

The change below is the *antidelta* of what's been done in the original
changeset. JDK 11u's version of LocaleProviders.sh already has the
equivalent of LocaleProvidersRun.java change:

if [ "${DEFFMTLANG}" = "zh" ] && [ "${DEFFMTCTRY}" = "CN" ]; then
...

So no, I don't want to change it back. That's the reason why right now
the issue doesn't reproduce on JDK 11u as bug8027289Test method is only
called for zh_CN locale (you need the patch to reproduce). IMHO, if we
wouldn't backport this fix at all we should be fine as well.

This backport was a weird one. Thoughts?

Thanks,
Severin


> Thanks
> Christoph
> 
> > -----Original Message-----
> > From: jdk-updates-dev <jdk-updates-dev-bounces at openjdk.java.net> On
> > Behalf Of Severin Gehwolf
> > Sent: Montag, 30. September 2019 13:49
> > To: jdk-updates-dev at openjdk.java.net
> > Subject: [11u] 8215913: [Test_bug]java/util/Locale/LocaleProvidersRun.java
> > failed on de_DE and ja_JP locale.
> > 
> > Hi,
> > 
> > Please review this test stabilization patch. The issue doesn't actually
> > reproduce in 11u without this patch:
> > 
> > diff --git a/test/jdk/java/util/Locale/LocaleProviders.sh
> > b/test/jdk/java/util/Locale/LocaleProviders.sh
> > --- a/test/jdk/java/util/Locale/LocaleProviders.sh
> > +++ b/test/jdk/java/util/Locale/LocaleProviders.sh
> > @@ -352,7 +352,7 @@
> >  # testing 8027289 fix, if the platform format default is zh_CN
> >  # this assumes Windows' currency symbol for zh_CN is \u00A5, the yen
> >  # (yuan) sign.
> > -if [ "${DEFFMTLANG}" = "zh" ] && [ "${DEFFMTCTRY}" = "CN" ]; then
> > +if [ ! "${DEFFMTLANG}" = "en" ] && [ ! "${DEFFMTCTRY}" = "CN" ]; then
> >    METHODNAME=bug8027289Test
> >    PREFLIST=JRE,HOST
> >    PARAM1=FFE5
> > 
> > With it it does, since the test is actually called (and fails) with
> > de_DE locale. The patch didn't apply cleanly due to copyright hunk
> > fails (upper bound year is already at 2019) and LocaleProvidersRun.java
> > being LocaleProviders.sh in JDK 11u. It's one of the Oracle JDK 11
> > partity patches, if that helps gauge why I'm proposing this.
> > Personally, I wouldn't backport it, but I thought to get some feedback
> > whether we want this or not.
> > 
> > Bug: https://bugs.openjdk.java.net/browse/JDK-8215913
> > webrev: http://cr.openjdk.java.net/~sgehwolf/webrevs/JDK-
> > 8215913/01/webrev/
> > Original changeset: http://hg.openjdk.java.net/jdk/jdk/rev/76f7dbf458fe
> > 
> > Thoughts?
> > 
> > Thanks,
> > Severin