harmony-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Nathan Beyer <ndbe...@apache.org>
Subject Re: svn commit: r769133 - in /harmony/enhanced/classlib/trunk/modules/text: make/exclude.common src/test/java/org/apache/harmony/text/tests/java/text/DecimalFormatSymbolsTest.java
Date Mon, 27 Apr 2009 22:47:15 GMT
On Mon, Apr 27, 2009 at 4:03 PM, Tim Ellison <t.p.ellison@gmail.com> wrote:
> Nathan Beyer wrote:
>> FYI - this change is related to this issue
>> https://issues.apache.org/jira/browse/HARMONY-6184 i logged.
>>
>> I found that the DecimalFormatSymbolsTest was failing because the
>> localization data from the RI and the localization data from ICU. In
>> this case, the major difference was the NaN property being "NaN"
>> instead of the RI's "\uFFFD".
>>
>> I logged the bug as a non-bug difference. I modified the test to
>> assert the data that was in ser file, rather than comparing it to
>> Harmony's version.
>>
>> If we're going to stick with ICU, this seems like the only path to
>> take, unless we wanted to modify the localization info for ICU
>> ourselves.
>
> I'm comfortable with sticking to ICU's version of localization.  I don't
> want to get into having to maintain that stuff ourselves, it's a full
> time project in itself.
>
> If there are any complaints about the localization differences between
> ICU and the RI then we can refer them to the ICU project and let them
> apply their expertise to the issue.

Sounds good. I suspect that some of these localization differences may
be causing other tests to fail as well. We'll need to keep an eye out
for tests that may depending on localization equality to assert
behavior.

-Nathan

>
> Regards,
> Tim
>
>

Mime
View raw message