RFR 8158123: NPE when the annotations is used in export-to of module-info

[Maurizio Cimadamore]

Hi,
the patch looks good.

I would add another test for a case like this:

module m1 { exports p to m2. at Anno foo; }

and, possibly, for other constructs in a module-info which accept module 
names.

 From a more general point of view, I note that the parser is using the 
qualified identifier production to get at a module name (probably 
because module names can contain dots?). I'll defer to Alex any 
clarification to how the module grammar is specified - but I note here 
that, implementation-wise, perhaps having another production-like method 
to parse module names could be a winning move - since I don't think 
module names have much to do with qualified identifiers.

Maurizio

On 06/06/16 21:56, Maurizio Cimadamore wrote:
> Hi Shinya
> Jan is on vacation for few days, I'll try to look at it tomorrow.
>
> Thanks
> Maurizio
>
> On 06/06/16 14:16, ShinyaYoshida wrote:
>> Hi Jan and Maurizio,
>> Could you review this?
>>
>> webrev:
>> http://cr.openjdk.java.net/~shinyafox/jigsaw/8158123/webrev/ 
>> <http://cr.openjdk.java.net/%7Eshinyafox/jigsaw/8158123/webrev/>
>>
>> bugs:
>> https://bugs.openjdk.java.net/browse/JDK-8158123
>>
>> Regards,
>> shinyafox(Shinya Yoshida)
>>
>> 2016-05-30 14:09 GMT+09:00 ShinyaYoshida <bitterfoxc at gmail.com 
>> <mailto:bitterfoxc at gmail.com>>:
>>
>>     Hi jigsaw-dev,
>>     I've filed the issue and created the patch for it.
>>
>>     Please review this:
>>
>>     webrev:
>>     http://cr.openjdk.java.net/~shinyafox/jigsaw/8158123/webrev/
>>     <http://cr.openjdk.java.net/%7Eshinyafox/jigsaw/8158123/webrev/>
>>
>>     bugs:
>>     https://bugs.openjdk.java.net/browse/JDK-8158123
>>
>>     Regards,
>>     shinyafox(Shinya Yoshida)
>>
>>
>