Scope modifiers in interfaces [was Default methods on Map]

[David Holmes]

On 9/04/2013 10:28 PM, Stephen Colebourne wrote:
> While I disagree with the choice made (one of Java's strengths is a
> little bit of extra verbosity), I am happy if it is consistent. Based
> on the links I provided, clearly Project Lambda has a long way to go
> to meet that consistency, so it clearly hasn't been a firm
> recommendation up until now.

The examples you gave have all been fixed within the last couple of 
weeks. If you still see some please point them out but my searching 
shows that the only use of "public default" in the lambda repos is from 
java.time and:

./share/classes/java/security/Principal.java:    public default boolean 
implies(Subject subject) {
./share/classes/java/security/KeyStore.java:        public default 
Set<Attribute> getAttributes() {
./share/classes/javax/security/auth/Destroyable.java:    public default 
void destroy() throws DestroyFailedException {
./share/classes/javax/security/auth/Destroyable.java:    public default 
boolean isDestroyed() {

> java.time will of course be adapted to match.

Terrific.

Thanks,
David

>
> Stephen
>
>
> On 9 April 2013 12:57, David Holmes <david.holmes at oracle.com> wrote:
>> All of the lambda-related work has now moved (or is moving) to #1. No public
>> on interface methods as has been the recommendation since day one.
>>
>>> I find this inconsistency troubling. Its time for a firm
>>> recommendation to be made as to Oracle's preferred coding standard.
>>
>>
>> I agree because I see that java.time is using #2.
>>
>>
>>> I'm of the opinion that moving code to #3, explicit use of "public"
>>> will serve Java better in the long run. I find #1 particularly
>>> troubling, as it means a default method (which looks very like a
>>> normal method) looks like it is package scoped when I read the source
>>> code.
>>
>>
>> If/when non-public methods are allowed in interfaces then we should probably
>> make things explicit, in my opinion. I don't find anything troubling about
>> not having public of interface methods - default or abstract.
>>
>> David
>> -----
>>
>>
>>> (this affects the Map changes webrev)
>>> Stephen
>>>
>>>
>>>
>>> On 8 April 2013 19:07, Mike Duigou <mike.duigou at oracle.com> wrote:
>>>>
>>>> Hello all;
>>>>
>>>> This is a combined review for the new default methods on the
>>>> java.util.Map interface being added for the JSR-335 lambda libraries. The
>>>> reviews are being combined because they share a common unit test.
>>>>
>>>> http://cr.openjdk.java.net/~mduigou/JDK-8010122/0/webrev/
>>>>
>>>> 8004518: Add in-place operations to Map
>>>>    forEach()
>>>>    replaceAll()
>>>>
>>>> 8010122: Add atomic operations to Map
>>>>    getOrDefault()
>>>>    putIfAbsent()          *
>>>>    remove(K, V)
>>>>    replace(K, V)
>>>>    replace(K, V, V)
>>>>    compute()              *
>>>>    merge()                *
>>>>    computeIfAbsent()      *
>>>>    computeIfPresent()     *
>>>>
>>>> The * operations treat null values as being absent. (ie. the same as
>>>> there being no mapping for the specified key).
>>>>
>>>> The default implementations provided in Map are overridden in HashMap for
>>>> performance purposes, in Hashtable for atomicity and performance purposes
>>>> and in Collections for atomicity.
>>>>
>>>> Mike
>>>>
>>