Nest host validation vs NestHost attribute performed by Lookup::defineHiddenClass

[John Rose]

(See previous message.). I think ignoring is simpler. Throwing an error is ok if you expect users to benefit from the discipline but in this case such benefit is in the noise and thus a more regular spec is the high order bit. 

> On Sep 24, 2019, at 8:31 PM, Mandy Chung <mandy.chung at oracle.com> wrote:
> 
> What about one refinement:  since hidden classes cannot participate in NestHost/NestMembers, I suggest to a stricter rule to disallow a hidden class containing NH/NM attribute (rather than ignoring it) which also forbids "nested nest".