John Rose skrev:
>> About the slow down effect, It can be removed by allowing the two way
>> (John's one and Fredrik's one) to call a method handle:
>> MethodHandle.invoke() that performs a single pointer check and
>> MethodHandle.apply() that performs a less restrictive/more expensive
>> check.
I've seen two bugs like this before, and they both had the same cause:
assuming the second argument is at index 1 (for static methods, 2 for
instance methods). If a method takes a double and something else (e.g.
an int), then accessing the double works fine, but accessing the int
gives a verif
John Rose a écrit :
> On May 12, 2009, at 4:25 PM, Rémi Forax wrote:
>
>
>>> And it works like a charm :)
>>>
>>>
>> untile the next bug ...
>> The following code crash.
>> It doesn't crash if there is only one invokedynamic.
>>
>
> Perhaps it's a problem with converting the floating
