Code Review 7104209: Cleanup and remove librmi (native library)

David Holmes david.holmes at oracle.com
Mon Oct 24 22:18:46 UTC 2011


On 25/10/2011 12:47 AM, Chris Hegarty wrote:
> Thanks Mike,
>
> It was actually Alan that caught this. Crazy that someone ( at some time
> ) even thought this was a good idea!

Not only did someone think it a good idea but nobody else disagreed. 
Sometimes these things are lost in ancient history but it is always a 
concern when we make changes where we don't understand the reasons for 
the original code. Just a comment/observation - changes look fine to me.

David

> -Chris.
>
> On 24/10/2011 15:41, Mike Duigou wrote:
>> Looks good. Nice catch.
>>
>> Mike
>>
>> On Oct 24 2011, at 07:20 , Chris Hegarty wrote:
>>
>>> It appears that the only function of librmi is to export the VM
>>> function JVM_LatestUserDefinedLoader. It seems really silly to have a
>>> whole native library to perform such a trivial operation.
>>> JVM_LatestUserDefinedLoader is also used in java.io.ObjectInputStream.
>>>
>>> It would make more sense to put an entry point in sun.misc.VM, and
>>> have rmi, ois invoke it. This would completely eliminate the need for
>>> librmi.
>>>
>>> http://cr.openjdk.java.net/~chegar/7104209/webrev.00/webrev/
>>>
>>> -Chris.
>>



More information about the core-libs-dev mailing list