[12] RFR: 8214170: ResourceBundle.Control.newBundle should throw IllegalAccessException when constructor of the resource bundle is not public.

[naoto.sato at oracle.com]

Hi Roger,

Thanks for the review.

On 11/27/18 9:58 AM, Roger Riggs wrote:
> Hi Naoto,
> 
> Look ok.
> 
> java/util/ResourceBundle.java: 3187
>    The comment isn't strictly correct since the block applies to named 
> (and open) modules too.

Comment modified as suggested:

http://cr.openjdk.java.net/~naoto/8214170/webrev.01/

> 
> [I think I would complete the conversion of the test to java to know
> that it was working and not wait for 8213127].

I am not sure it was working, as some other test cases do like:

java/io/File/GetXSpace.sh:
---
deny() {
   case "$OS" in
   Windows* ) chacl -d ${SID}:f $* ;;
   * )        chmod 000 $*         ;;
   esac
}
---

This implies chacl should be used in shell based test on Windows. I 
could do the same thing, but just did not do it in the to-be-eliminated 
shell file :-)

Dora (RE for 8213127) is aware of the Windows permission issue and fix 
it in 8213127 with a java equivalent test.

Naoto

> 
> Thanks, Roger
> 
> 
> On 11/26/2018 08:07 PM, Naoto Sato wrote:
>> Hi,
>>
>> Please review the fix to the following issue:
>>
>> https://bugs.openjdk.java.net/browse/JDK-8214170
>>
>> The proposed changeset is located at:
>>
>> http://cr.openjdk.java.net/~naoto/8214170/webrev.00/
>>
>> The existing logic to determine if there is a pubic constructor for 
>> the ResourceBundle class is incorrect. Moved the catch clause for 
>> NoSuchMethodException to handle it correctly. A shell based test was 
>> modified according to this (intentionally omitted the case for windows 
>> as chmod does not work - it will be addressed with the test case clean 
>> up (8213127).
>>
>> Naoto
>