8021257: com.sun.corba.se.** should be on restricted package list

[Seán Coffey]

Looks good to me too Alan. Thanks for taking care of this. After 
studying the class delegation code, it's now making sense to me why the 
initial package restriction failed.

regards,
Sean.

On 21/10/13 13:45, Chris Hegarty wrote:
> Very nice Alan.
>
> I have reviewed all the changes, and am happy to be listed as reviewer 
> for both changesets, though Mark or Sean should also probably take a 
> look at the corba changes.
>
> -Chris.
>
> On 21/10/2013 12:14, Alan Bateman wrote:
>>
>> Back in June, Sean Coffey had an attempt to add com.sun.corba.se.impl.**
>> to the restricted package list. Problems arose and the change had to be
>> backed out ([1] [2]]).
>>
>> Mark Sheppard and I have been looking at this one again and we would
>> like to make a second attempt at this. The proposed patch is here:
>>
>> http://cr.openjdk.java.net/~alanb/8021257/
>>
>> The main difference between this and the original patch is that we've
>> changed the RMI-IIOP code to statically reference the fallback/default
>> implementation rather than using reflection. Same thing for the ORB
>> initialization. We could have fixed the reflection usages but it didn't
>> seem worth it (from a modularity point of view then there isn't a
>> significant need at this time to have the API + default implementation
>> in separate modules, if such a need arises then we could change this to
>> use ServiceLoader).
>>
>> The other difference is that we are proposing to add com.sun.corba.se.**
>> rather than just the impl classes. The rational is that
>> com.sun.corba.se.** is not a documented and supported interface so there
>> shouldn't be anything using it directly. This change shouldn't have any
>> impact any folks using the endorsed standards mechanism to replace
>> javax.rmi.CORBA and org.omg.**.
>>
>> -Alan
>>
>> [1] http://hg.openjdk.java.net/jdk8/tl/jdk/rev/c4480e0d9f53
>> [2] http://hg.openjdk.java.net/jdk8/tl/jdk/rev/7cc35dd1885d