JDK 10 RFR of 8028544 and 8028546

[joe darcy]

On 1/26/2017 4:41 PM, Jonathan Gibbons wrote:
> Mostly OK.
>
> test/tools/javac/file/MultiReleaseJar/MultiReleaseJarTest.java
> I wonder whether in this test you should be adding a test case, 
> instead of changing 9 to 10.

I just kicked off a fresh test run. I'll revisit this test once that run 
completes.

The support for --release may be a bit odd until 9 is truly separate 
from 10, but we should make it as full as possible as soon as possible.

Thanks,

-Joe

>
> -- Jon
>
>
> On 01/26/2017 04:23 PM, joe darcy wrote:
>> Hello,
>>
>> All langtools tests should be passing now with the refined changes in
>>
>>     http://cr.openjdk.java.net/~darcy/8028546.2/
>>
>> Stuart, jdeprscan required some minor edits.
>>
>> As noted implicitly in the some of changes, some of the areas of the 
>> code affected on this webrev will need revision once there is 
>> distinct support for --release 9 and --release 10.
>>
>> Thanks,
>>
>> -Joe
>>
>>
>> On 1/26/2017 12:37 PM, joe darcy wrote:
>>> Hello,
>>>
>>> As part of JDK 10 ground breaking, I'm working on fixes to
>>>
>>>     JDK-8028544: Add SourceVersion.RELEASE_10
>>>     JDK-8028546: Add -source 10 and -target 10 to javac
>>>
>>> Current webrev available at
>>>
>>>     http://cr.openjdk.java.net/~darcy/8028546.1/
>>>
>>> With a few exceptions, the langtools tests are all passing with 
>>> these changes. I updated a few annotation processing tests from 
>>> using the vanilla AbstractProcessor to the 
>>> JavacTestingAbstractProcessor. The exceptions include the currently 
>>> failing tests
>>>
>>> FAILED: 
>>> tools/javac/T8009640/CheckRejectProfileBCPOptionsIfUsedTogetherTest.java 
>>>
>>> FAILED: tools/jdeprscan/tests/jdk/jdeprscan/TestRelease.java
>>> FAILED: tools/javac/file/MultiReleaseJar/MultiReleaseJarTest.java
>>>
>>> The failures seem related to the --release option. At least until 
>>> the Java SE 9 API is finalized, I think it is okay for --release 9 
>>> and --release 10 to mean the same thing and refer to the in-forest 
>>> state of the API. However, I haven't fully reverse-engineered how to 
>>> do that.
>>>
>>> Thanks,
>>>
>>> -Joe
>>>
>>
>