Syntax for calling super
Brian Goetz
brian.goetz at oracle.com
Mon Aug 27 07:20:28 PDT 2012
>>> K.super.m() already has an existing meaning with inner classes, just as
>>> K.this.m() does. There's a difference between searching for a type alone
>>> and searching for an object and then a type. Using the same notation is
>>> confusing in my view.
>>
>> Oh, I wasn't aware of that. That changes things.
The EG *was* aware of that, but decided K.super.m() was still the
superior choice despite the theoretical conflict. In particular:
- Using this syntax for super-calls to defaults will not change the
meaning of any existing code, since all existing code that uses
ClassName.super.m() would use it with a class, not an interface, and
there currently are no existing such super-calls in interfaces (since
there is currently no method bodies in interfaces.)
- The "old" syntax is so rarely used that almost no one even knows
about it. In informal inquiries, most people we showed it to said
"that's not legal code." (Most people here didn't know about it either.)
- Conflicts are expected to be so rare that we didn't think it was
worth choosing a lesser syntax for something that may actually be common.
As I've said, if someone comes up with something that is "obviously
better", we'll consider it. (But please, "obviously better" is a pretty
high bar, so let's not paint a bikeshed that's already been painted.)
More information about the lambda-dev
mailing list