ValueType.findConstructor

Maurizio Cimadamore maurizio.cimadamore at oracle.com
Mon Jun 26 23:24:02 UTC 2017


Looks good - I noted you have removed the caching - that's probably ok 
since unbox is cached anyway, but dunno if filterReturnValue is costly 
(does that go through lambda form editor?) -  and whether it would make 
sense to cache that too.

Maurizio


On 26/06/17 22:23, Paul Sandoz wrote:
> Updated:
>
> http://cr.openjdk.java.net/~psandoz/valhalla/find-constructor/webrev/ 
> <http://cr.openjdk.java.net/%7Epsandoz/valhalla/find-constructor/webrev/>
>
> Paul.
>
>> On 26 Jun 2017, at 10:46, Paul Sandoz <Paul.Sandoz at oracle.com 
>> <mailto:Paul.Sandoz at oracle.com>> wrote:
>>
>> Hi Maurizio,
>>
>> It’s not naive, that code reflected a reduction on some hacking in an 
>> attempt to also explore construction with a default constructor 
>> (there is also a disconnect here between the VCC with a constructor 
>> and using vdefault with DVT). I think we should keep the constructor 
>> method and implement it as you suggest with a combinator. I’ll send 
>> out a simplified webrev.
>>
>> Separately, in my enthusiasm i may be going down a rabbit hole 
>> regarding re-construction, arguably it’s not in the spirit of 
>> minimal. Still at some point i think we should explore this notion of 
>> a more extensible wither for value types (but it also seems 
>> application to ref/data classes too).
>>
>> Paul.
>>
>>> On 23 Jun 2017, at 16:31, Maurizio Cimadamore 
>>> <maurizio.cimadamore at oracle.com 
>>> <mailto:maurizio.cimadamore at oracle.com>> wrote:
>>>
>>> Maybe a naive question - but do you need method handle builder (and 
>>> extra bytecode spinning) for constructor? Can't we combine the 
>>> result of findConstructor on the boxed class with the vunbox MH?
>>>
>>> Maurizio
>




More information about the valhalla-dev mailing list