[webbeans-dev] @New
by Pete Muir
Since the change in the spec defining that there is a new bean for
every bean, we lost the language that disallows:
@New @SomeOtherBinding Foo foo;
AFAICT it still makes sense to disallow this. So, should that language
be there?
We also lost the language that disallows specific declaration of a
Bean as having @New as a binding type.
--
Pete Muir
http://www.seamframework.org
http://in.relation.to/Bloggers/Pete
15 years, 7 months
[webbeans-dev] Context.get with mismatched scopes
by David Allen
Hi Gavin,
I was just writing some tests and realized the spec in section 8.1 does
not clearly state what happens if I use a context from one scope and a
bean from a different scope. Should this be a runtime exception?
Example:
@Dependent
public class MyBean {}
...
Bean myBean = manager.resolveByType(MyBean.class).iterator().next();
MyBean myBeanInstance =
manager.getContext(SessionScoped.class).get(myBean);
-------------
David Allen
david.allen(a)acm.org
15 years, 7 months
[webbeans-dev] package of @Model
by Mark Struberg
Hi!
The spec does not cover in which package the @Model should go.
The ri has it in the 'old' javax.webbeans package which has been killed a while ago.
The package section of the Spec doesn't cover it at all.
So in which package should @Model finally end?
txs and LieGrue,
strub
15 years, 9 months
[webbeans-dev] TCK Question
by Gurkan Erdogdu
Hi guys,
While executing the standalone TCK tests related with "Event and Observers", I tackled the one point.
Specification says that whenever the bean uses the @Fires annotation on its class field, like @Fires Event<T>, container provides the implicit bean component. But it does/may not say that you have to add this implicit bean component to the "Manager"'s beans bag. So when resolving the implicit event object with Manager#resolveByType method, our implementation returns no bean.
What we do is that whenever the field injection is annotated with @Fires, we create a new implicit bean object and add it into the bean Dependent context, but not add this implicit bean into the Manager's bag.
But in the TCK, it requires that Manager adds this implicit bean into its bag.
Actually, these are also applied for all implicit beans except the Manager and Conversation beans.
WDYT?
/Gurkan
15 years, 9 months
[webbeans-dev] TCK Tests - Dependent Context Activation
by Gurkan Erdogdu
Hi;
Specification states that depending context is active --> "while an observer method is invoked", but it does not state that where is the starting point for the activation. In the TCK tests, it is activated on the Manager#fireEvent.
But currently in the OpenWebBeans project, we activate the Dependent Context at the start of the Observer#notify method.
Which one is correct?
Thanks;
/Gurkan
15 years, 9 months
[webbeans-dev] Beta1 release
by Pete Muir
All,
I would like to release beta1 on 3rd March. Outstanding for this we
have:
* Dependent object destruction: Nik, what is the status of this?
* Conversations: Nik, what is the status of this?
* Hierarchical managers: I will work on this next week
* Disposal methods: Daniel Roth is working on a patch for this
* Injection into EJBs: I will work on this next week
* Injection into servlets: I will work on this next week
* TX observers: implemented, but has issues. David, if you can get
this done for the release that would be ace!
* Async observers: needs implementing. Again, David, if you can get
this done, that would be great...
* A number of small issues thrown up by the TCK. I will work on this
tomorrow
Please let me know if you are going to have issues finishing a
particular features (including TCK tests) for the 3rd and we can
either slip the feature or slip the release date.
Thanks!
--
Pete Muir
http://www.seamframework.org
http://in.relation.to/Bloggers/Pete
15 years, 10 months
Re: [webbeans-dev] [TCK] ContextTest#testGetContextWithTooManyActiveContextsFails
by Pete Muir
Hi Mark,
If you, or others on the OpenWebBeans project discover similar issues
with the TCK, please don't hesitate to create a JIRA :-)
We can then discuss there, and make a decision about how the TCK
should be. This also provides a record so that people can look back if
a similar question comes up later.
Thanks!
On 26 Feb 2009, at 17:42, Mark Struberg wrote:
> done:
>
> https://jira.jboss.org/jira/browse/WBTCK-20
>
> LieGrue,
> strub
>
>
> --- Pete Muir <pmuir(a)redhat.com> schrieb am Do, 26.2.2009:
>
>> Von: Pete Muir <pmuir(a)redhat.com>
>> Betreff: Re: [webbeans-dev] [TCK]
>> ContextTest#testGetContextWithTooManyActiveContextsFails
>> An: "Mark Struberg" <struberg(a)yahoo.de>
>> CC: webbeans-dev(a)lists.jboss.org
>> Datum: Donnerstag, 26. Februar 2009, 18:09
>> Hi Mark,
>>
>> Looks like a bug, please file a WBTCK :-)
>>
>> Thanks
>>
>> On 26 Feb 2009, at 16:42, Mark Struberg wrote:
>>
>>> Hi!
>>>
>>> I have a question on the
>>>
>> ContextTest#testGetContextWithTooManyActiveContextsFails()
>>>
>>> The spec says that 2 _active_ contexts may not be
>> registered at the same time.
>>>
>>> However, you assume an IllegalArgumentException to be
>> thrown if two active contexts exists while calling the
>> addContext() function (btw cannot find this in the Spec).
>>>
>>> On the other hand, the Spec requests to throw an
>> IllegalStateException for the getContext() if two active
>> contexts exist.
>>>
>>> Can we harmonise this?
>>> Or if ok the way it is, explain the background?
>>> This imho seams to interfere with the 'general
>> rules' of jsr-299 Exceptions you (or Gavin) mentioned in
>> a post a few months ago
>>>
>>> txs and LieGrue,
>>> strub
>>>
>>>
>>>
>>>
>>>
>>>
>>> _______________________________________________
>>> webbeans-dev mailing list
>>> webbeans-dev(a)lists.jboss.org
>>> https://lists.jboss.org/mailman/listinfo/webbeans-dev
>>
>> --
>> Pete Muir
>> http://www.seamframework.org
>> http://in.relation.to/Bloggers/Pete
>
>
>
--
Pete Muir
http://www.seamframework.org
http://in.relation.to/Bloggers/Pete
15 years, 10 months
[webbeans-dev] [TCK] ContextTest#testGetContextWithTooManyActiveContextsFails
by Mark Struberg
Hi!
I have a question on the
ContextTest#testGetContextWithTooManyActiveContextsFails()
The spec says that 2 _active_ contexts may not be registered at the same time.
However, you assume an IllegalArgumentException to be thrown if two active contexts exists while calling the addContext() function (btw cannot find this in the Spec).
On the other hand, the Spec requests to throw an IllegalStateException for the getContext() if two active contexts exist.
Can we harmonise this?
Or if ok the way it is, explain the background?
This imho seams to interfere with the 'general rules' of jsr-299 Exceptions you (or Gavin) mentioned in a post a few months ago
txs and LieGrue,
strub
15 years, 10 months
