[9] Review request: 8170702: Document that javafx.graphics needs explicit access to application main class

[Kevin Rushforth]

Mandy Chung wrote:
>> On Mar 3, 2017, at 10:36 PM, Kevin Rushforth <kevin.rushforth at oracle.com> wrote:
>>
>> [fixed subject line]
>>
>> Please review the following to document that javafx.graphics needs explicit access to the Application class.
>>
>> https://bugs.openjdk.java.net/browse/JDK-8170702
>> http://cr.openjdk.java.net/~kcr/8170702/webrev.00/
>>     
>
>   69  * containing package must be {@link Module#isExported(String,Module) exported}
>
> @linkplain instead?
>   

I was following the pattern in Module.java, etc., which uses a regular 
@link in similar cases.

>  239         StackTraceElement[] cause = Thread.currentThread().getStackTrace();
>
> Good candidate to use StackWalker API.
>   

This is pre-existing code (since JDK 7), and I don't want to change the 
implementation this late while fixing a doc bug. I will file a follow-on 
bug to consider improving this for JDK 10.

> Is @throws RuntimeException an existing behavior?  I’d think CNFE and InaccessibleAE might be more appropriate.
>   

Yes, this is the existing behavior and we are just documenting it. I 
agree that it might have been nicer to do something else, but the 
behavior w.r.t., exception is unchanged since JDK 8.

> line 209 “It must be a subclass of Application or a RuntimeException will be thrown.”
>
> I think this statement should be extended to cover if the class and its constructor are public and exported.
>   

Yes, this seems like another good place to document the restriction. 
I'll post a .01 version of the webrev with this update.

-- Kevin


> Mandy