Request to backport 7102369/709446/7100592 and 7092186

[Omair Majid]

On 11/10/2011 05:51 PM, Dr Andrew John Hughes wrote:
> On 16:39 Thu 10 Nov     , Omair Majid wrote:
>> Hi,
>>
>> The recent security changes breaks some applications as rmiregistry
>> needs to be started with the java.rmi.server.codebase property set. I
>> would like to backport the following two changesets from jdk7u to
>> address this problem:
>>
>> http://hg.openjdk.java.net/jdk7u/jdk7u/jdk/rev/7ed2fd310470
>> http://hg.openjdk.java.net/jdk7u/jdk7u/jdk/rev/2d10aa9d772a
>>
>> Does anyone have any thoughts or concerns?
>>
>
> Can you explain a bit more about these patches and why they are
> needed/appropriate?  The bugs don't give much detail and one isn't
> even accessible.
>

Here is a bug report of a user who now has to start rmiregistry with 
-J-Djava.rmi.server.codebase= option for his code to continue working 
after the security update:
https://bugzilla.redhat.com/show_bug.cgi?id=751203

As you can see, one of the patches removes the need for explicitly 
setting this property (and reverts back to the pre-security-update 
condition). I dont know about you, but I think security fixes should not 
be causing any change in behaviour of valid applications (as far as 
possible).

Also, the jtreg test sun/tools/jstatd/jstatdExternalRegistry.sh has been 
failing ever since the security fixes were added; applying these two 
patches makes that test pass too.

> Have you posted these to IcedTea?

I was expecting IcedTea7 to get these when it merges OpenJDK 7u 
changesets. I will be adding this to IcedTea6 when it is approved for 
OpenJDK6. Of course, unless people want otherwise and I will be happy to 
accommodate that.

Cheers,
Omair