Switch labels (null again), some tweaking

[John Rose]

On Apr 23, 2021, at 1:10 PM, Brian Goetz <brian.goetz at oracle.com<mailto:brian.goetz at oracle.com>> wrote:

I think the concern on the part of the null-fearers is not so much the O(n) scan (though that's a concern), as much as the subtle difference between `case Object o` and `default`.  They are almost identical, except in null.  No one likes having to carry that distinction around in their head.

I mildly disagree with this last point:  I think that
distinction will cause mild pain to learn, and zero
point zero zero pain to carry around for the rest
of one’s career.  I can feel the rule growing in my
own head: “default means old-school null rejection,
a null case label means null acceptance, if both
are present the null case label of course wins.”

Could this explicit null-ing be MaKing the New
nUlls Stand Out?

— John
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mail.openjdk.java.net/pipermail/amber-spec-experts/attachments/20210423/cd58609a/attachment.htm>