On Tue, Aug 23, 2011 at 9:13 PM, Igor Vaynberg <[email protected]> wrote:
> On Tue, Aug 23, 2011 at 8:04 AM, Emond Papegaaij
> <[email protected]> wrote:
>> On Tuesday 23 August 2011 16:17:43 Igor Vaynberg wrote:
>>> On Tue, Aug 23, 2011 at 1:16 AM, Emond Papegaaij
>>>
>>> <[email protected]> wrote:
>>> > On Tuesday 23 August 2011 09:04:00 Igor Vaynberg wrote:
>>> >> On Mon, Aug 22, 2011 at 11:44 PM, Emond Papegaaij
>>> >>
>>> >> <[email protected]> wrote:
>>> >> > For example, objects managed by weld are not detached at the end
>>> >> > of a request.
>>> >>
>>> >> erm? what is supposed to happen when?
>>> >>
>>> >> -igor
>>> >
>>> > That's something I'm not entirely sure of. The current implementation
>>> > makes it cumbersome to work with models inside scoped objects (requests
>>> > can share the same objects instances, causing multi-threading issues).
>>> > I'm aware of the fact that is inherent to the way the scopes work, but
>>> > it would be nice to have wicket-weld provide a way to cope with this.
>>> > Putting it in 1.5 leaves no time to (try to) develop an api to deal with
>>> > this.
>>>
>>> not entirely sure what you mean. requestsope and conversation scopes
>>> are threadsafe. session and application scopes are not, but that is
>>> something you have to be aware of and deal with just like with any
>>> other IOC framework...
>>
>> Are you sure that conversation scopes are threadsafe? The same conversation
>> can be shared across multiple requests. When some of these requests happen
>> simultaneously, I expect the conversation scoped objects to be accessed in a
>> multithreaded way.
>> I do not yet know how to cope with this, but it is something I will be
>> investigating in the next couple of days. Perhaps, wicket-weld (or cdi) needs
>> its own representation for the conversation scope, just as it has its own
>> applicatoin, session and request objects.
>
> from the spec
>
> """The container ensures that a long-running conversation may be
> associated with at most one request at a time, by blocking
> or rejecting concurrent requests. If the container rejects a request,
> it must associate the request with a new transient conversation
> and throw an exception of type
> javax.enterprise.context.BusyConversationException..."""
>
>>> > Igor:
>>> > Did you move your code for wicket-weld to another place, or did you
>>> > delete it? We are really interested in this code, and are currently
>>> > adding it to our project. For now I've resurrected it at
>>> > github.com/papegaaij/wicket-weld. It would be nice if it is placed
>>> > somewhere more central, where it could be developed further to be
>>> > included in Wicket at some later time.
>>>
>>> i will be moving it into 42lines github repo today. i was going to
>>> move it to wicketstuff but that is not possible because weld artifacts
>>> are not in central yet.
>>
>> Great. Could you post the location on this list when you placed it there, so 
>> I
>> can clone it? For now, we at Topicus will be maintaining or own clone, to
>> allow us to follow our own release management, but it would be great if we 
>> can
>> sync it with an 'upstream' repo.
>
> will do....
>
>>
>> I'm still quite new to CDI and Weld. Do you know any good resources (books?)
>> on this? I've found the Weld documentation very useful, but it's more of a
>> reference guide, rather than a book.
>
> i didnt use a book, so dont know of any good ones out there. sorry.
ivaynberg writes books, he don't read books!
>
> -igor
>
>>
>> Emond
>>
>



-- 
Martin Grigorov
jWeekend
Training, Consulting, Development
http://jWeekend.com

Reply via email to