[foreign] RFR: cleanup Pointer API
Sundararajan Athijegannathan
sundararajan.athijegannathan at oracle.com
Wed May 16 00:08:48 UTC 2018
Thanks for fixing it.
PS. And thanks for your comment on "public" :)
-Sundar
On 16/05/18, 1:58 AM, Maurizio Cimadamore wrote:
>
>
> On 15/05/18 19:09, Maurizio Cimadamore wrote:
>>
>>
>> On 15/05/18 17:44, Sundararajan Athijegannathan wrote:
>>>
>>> * Runtime class could be final and package-private. All static
>>> helpers methods could be package-private as well - because
>>> HeaderImplGenerator in the same package creates method handles on
>>> those methods.
>> good point, I'll make some experiments
> I've fixed this and pushed.
>
> Thanks
> Maurizio
>>>
>>> * Pointer.toString, .copy methods: "public" removed. Usually in
>>> APIs, we write "public" explicitly - even though interface methods
>>> are public by default, right?
>> Nope - look at Collection.java :-)
>>>
>>> +1 other than that.
>>>
>>> -Sundar
>>>
>>> On 15/05/18, 8:08 PM, Maurizio Cimadamore wrote:
>>>> Hi,
>>>> in preparation for the changes to the pointer API outlined in [1],
>>>> here's a cleanup that removes the methods from the Pointer
>>>> interface that are only required to support code generation.
>>>>
>>>> These methods are now moved into a new class, called
>>>> RuntimeSupport. I've tweaked the HeaderImplGenerator class to
>>>> perform indirect access to these method, through a bunch of static
>>>> MH fields (this is to avoid access errors due to module boundaries).
>>>>
>>>> Webrev:
>>>>
>>>> http://cr.openjdk.java.net/~mcimadamore/panama/pointer-cleanup/
>>>>
>>>> Maurizio
>>>>
>>>>
>>
>
More information about the panama-dev
mailing list