On Wed, 28 Oct 2020 15:00:59 GMT, Jorn Vernee wrote:
>> Hi,
>>
>> Currently, if MethodHandles::permuteArguments is used with a reorder array
>> that is the wrong size, or one of the indexes in it is out of bounds of the
>> new type, we simply get the exception message:
>>
>> bad reorder
On Wed, 28 Oct 2020 14:30:30 GMT, Chris Hegarty wrote:
>> I've added some negative tests that test for the different failure
>> conditions.
>
> Thanks for adding additional test coverage @JornVernee.
>
> Writing a tight implementation of assertThrows is non-trivial - I'm not sure
> that the
> Hi,
>
> Currently, if MethodHandles::permuteArguments is used with a reorder array
> that is the wrong size, or one of the indexes in it is out of bounds of the
> new type, we simply get the exception message:
>
> bad reorder array [...]
>
> I think we can improve the exception message
On Wed, 28 Oct 2020 14:57:44 GMT, Jorn Vernee wrote:
>> Hi,
>>
>> Currently, if MethodHandles::permuteArguments is used with a reorder array
>> that is the wrong size, or one of the indexes in it is out of bounds of the
>> new type, we simply get the exception message:
>>
>> bad reorder
On Wed, 28 Oct 2020 13:32:19 GMT, Jorn Vernee wrote:
>>>
>>>
>>> Seems like a reasonable change. Is there an already existing test for "bad"
>>> permute args that could be expanded to discern the new exception message?
>>
>> There are several tests for permuteArguments, but none that
On Wed, 28 Oct 2020 11:52:17 GMT, Jorn Vernee wrote:
>> Seems like a reasonable change. Is there an already existing test for "bad"
>> permute args that could be expanded to discern the new exception message?
>
>>
>>
>> Seems like a reasonable change. Is there an already existing test for
> Hi,
>
> Currently, if MethodHandles::permuteArguments is used with a reorder array
> that is the wrong size, or one of the indexes in it is out of bounds of the
> new type, we simply get the exception message:
>
> bad reorder array [...]
>
> I think we can improve the exception message
On Wed, 28 Oct 2020 09:20:54 GMT, Aleksey Shipilev wrote:
>
>
> This looks okay to me. Someone from core-libs should take a look as well.
>
> As the follow-up, maybe reconcile that method returns normally only with
> `true`, and throws exceptions otherwise. There are some uses like
>
On Wed, 28 Oct 2020 10:19:58 GMT, Chris Hegarty wrote:
>
>
> Seems like a reasonable change. Is there an already existing test for "bad"
> permute args that could be expanded to discern the new exception message?
There are several tests for permuteArguments, but none that explicitly test
On Wed, 28 Oct 2020 09:20:54 GMT, Aleksey Shipilev wrote:
>> Hi,
>>
>> Currently, if MethodHandles::permuteArguments is used with a reorder array
>> that is the wrong size, or one of the indexes in it is out of bounds of the
>> new type, we simply get the exception message:
>>
>> bad
On Tue, 27 Oct 2020 12:43:47 GMT, Jorn Vernee wrote:
> Hi,
>
> Currently, if MethodHandles::permuteArguments is used with a reorder array
> that is the wrong size, or one of the indexes in it is out of bounds of the
> new type, we simply get the exception message:
>
> bad reorder array
Hi,
Currently, if MethodHandles::permuteArguments is used with a reorder array that
is the wrong size, or one of the indexes in it is out of bounds of the new
type, we simply get the exception message:
bad reorder array [...]
I think we can improve the exception message by splitting these
12 matches
Mail list logo
