Review request for IMF classes and Locale related classes (original) (raw)

Stuart Marks stuart.marks at oracle.com
Fri Dec 2 17:41:13 PST 2011

The changes you made are good, thanks.

I did figure out how to deal with the static initializer in AllAvailableLocales:

     @SuppressWarnings({"unchecked"})
     Class<? extends LocaleServiceProvider>[] providerClasses =
             (Class<? extends LocaleServiceProvider>[]) new Class<?>[] {
         java.text.spi.BreakIteratorProvider.class,
         ...
     };

This is a bit uglier but it does reduce the scope of warnings suppression.

I think you might be able to get rid of the wildcards, too:

     @SuppressWarnings({"unchecked"})
     Class<LocaleServiceProvider>[] providerClasses =
             (Class<LocaleServiceProvider>[]) new Class<?>[] {
         java.text.spi.BreakIteratorProvider.class,
         ...
     };
     ...
     for (Class<LocaleServiceProvider> providerClass : providerClasses) {
         ...
     }

I'm not entirely sure why this works, but it seems to compile for me without warnings or errors....

s'marks

On 12/2/11 4:00 PM, Naoto Sato wrote:

Hi Stuart,

Thank you for your comments. Here is the updated webrev reflecting your suggestions. Please review: http://cr.openjdk.java.net/~naoto/7117469/webrev.01/ Naoto

On 12/2/11 3:32 PM, Stuart Marks wrote: Hi Naoto,

A couple comments. java/util/Currency.java -- The @SuppressWarnings covers the entire method. We're trying to use @SuppressWarnings with as narrow a scope as possible. Sometimes it's helpful to create a local variable declaration for this purpose; perhaps something like this will help: @SuppressedWarnings("unchecked") Set result = (Set) available.clone(); return result; You can probably do something similar in sun/util/LocaleServiceProviderPool.java in the getLocalizedObjectImpl() method. The suppression of warnings on the static AllAvailableLocales class is a bit of a puzzle. Maybe it's OK to leave the suppression for that entire class. s'marks

On 12/2/11 2:24 PM, Masayoshi Okutsu wrote: Looks good to me. Thanks, Masayoshi On 2011/12/02 10:56, Naoto Sato wrote: Hello,

Could you please review these two changesets for the WCD? One is for classes that belongs to input method framework: http://cr.openjdk.java.net/~naoto/7117465/webrev.00/ and the other is for (some of the) i18n related .util. classes: http://cr.openjdk.java.net/~naoto/7117469/webrev.00/ Thanks! Naoto

More information about the jdk8-dev mailing list