2:59 a.m.
Correct.
On 12/22/2014 09:59 AM, Mark Struberg wrote:
While we are at it: This is actually true for ALL System Events,
right?
So we should not define it in PAT but somewhere more general imo.
LieGrue,
strub
> On Monday, 22 December 2014, 9:56, Jozef Hartinger <jharting(a)redhat.com>
wrote:
>> Yes, we should explicitly define those other cases as non-portable if it
> has not been done yet.
>
> On 12/22/2014 09:49 AM, Mark Struberg wrote:
>> well, it's just the tip of the iceberg. It would be way easier to just
> define any 'modification' as non-portable. That would be much more
> precise and easier as well. And really covers al the edge cases as well.
>> LieGrue,
>> strub
>>
>>
>>
>>
>>
>>> On Monday, 22 December 2014, 9:42, Jozef Hartinger
> <jharting(a)redhat.com> wrote:
>>> On 12/22/2014 09:33 AM, Mark Struberg wrote:
>>>> No you can NOT I fear.
>>>>
>>>> How do you make sure that the AnnotatedType someone added
> doesn't get
>>> changed later? We have no whatever control over it. If some programmer
> like to
>>> make it mutable, well then it is that way and we have no chance to
> detect that
>>> :(
>>>> Imo the only thing we can do is to point programmers to the fact
> that this
>>> is stupid -> non-portable behaviour.
>>> Correct. This is stupid. But allowing setAnnotatedType() to be called
>>> anytime does not help this at all.
>>>
>>> The fact that we throw the exception on setAnnotatedType() prevents a
>>> class of bad things from happening. At the same time it does not
> prevent
>>> the other one that you mentioned. The fact that we cannot prevent all
> of
>>> them does not mean that we should not try to prevent those we can.
>>>
>>>> LieGrue,
>>>> strub
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>> On Monday, 22 December 2014, 9:29, Jozef Hartinger
>>> <jharting(a)redhat.com> wrote:
>>>>> On 12/21/2014 09:47 PM, Mark Struberg wrote:
>>>>>> Hi!
>>>>>>
>>>>>> I just came across this little sentence in the spec
>>>>>>
>>>>>>
>>>>>>
>>>>>> 11.5.6 "If any ProcessAnnotatedType method is
> called outside
>>> of the
>>>>> observer method invocation, an IllegalStateException is
> thrown."
>>>>>> I don't believe such a limitation helps much. What
> about
>>> extensions who
>>>>> do a setAnnotatedType and change this instance in a later
> phase?
>>>>> Such as?
>>>>>> We have no whatever chance to prevent this anyway.
>>>>> We can prevent that by throwing the exception, rather than
> silently
>>>>> ignoring that the extension is doing something wrong.
>>>>>
>>>>>> So why not just say that if a CDI System Event gets
> modified
>>> outside of the
>>>>> method it gets injected into then non portable behaviour
> results.
>>>>>> LieGrue,
>>>>>> strub
>>>>>> _______________________________________________
>>>>>> cdi-dev mailing list
>>>>>> cdi-dev(a)lists.jboss.org
>>>>>> https://lists.jboss.org/mailman/listinfo/cdi-dev
>>>>>>
>>>>>> Note that for all code provided on this list, the
> provider
>>> licenses the
>>>>> code under the Apache License, Version 2
>>>>> (http://www.apache.org/licenses/LICENSE-2.0.html). For all
> other ideas
>>> provided
>>>>> on this list, the provider waives all patent and other
> intellectual
>>> property
>>>>> rights inherent in such information.
>>>>>