I think this is satisfactory. +1
Brian On May 20, 2016, at 2:03 PM, Phil Race <philip.r...@oracle.com> wrote: > The jai_imageio code parses the version string before > it does anything with the result. And the only thing > it intends to do is tweak the registruy ordering or > remove itself if it thinks the JDK version is high enough. > So I don't see the harm in this versus not adding it to > the registry at all. > > Theoretically a different plugin that had a fatal error > as a result of this probably should not be registered > but there can't be existing examples of that on JDK 1.8 > and even if there were a problem on 9 they'll get > the exception printed .. > > -phil. > > On 05/20/2016 01:55 PM, Brian Burkhalter wrote: >> These statements >> >> map.put(provider.getClass(), provider); >> poset.add(provider); >> >> occur before the “ if (provider instanceof RegisterableService)” block. Do >> we really want the provider added to that map and set if the Throwable >> occurs? >> >> Brian >> >> On May 20, 2016, at 1:07 PM, Phil Race <philip.r...@oracle.com> wrote: >> >>> Please review the following workaround for :- >>> https://bugs.openjdk.java.net/browse/JDK-8154860 >>> >>> -phil. >>> >>> --- a/src/java.desktop/share/classes/javax/imageio/spi/ServiceRegistry.java >>> +++ b/src/java.desktop/share/classes/javax/imageio/spi/ServiceRegistry.java >>> @@ -770,7 +770,12 @@ >>> poset.add(provider); >>> if (provider instanceof RegisterableService) { >>> RegisterableService rs = (RegisterableService)provider; >>> + try { >>> rs.onRegistration(registry, category); >>> + } catch (Throwable t) { >>> + System.err.println("Caught and handled this exception :"); >>> + t.printStackTrace(); >>> + } >>> } >>> >>> return !present; >>> >> >
