You are viewing a plain text version of this content. The canonical link for it is here.

Posted to users@cxf.apache.org by António Mota <am...@gmail.com> on 2013/12/11 13:08:19 UTC

Websockets and SSE

Hi all.

Does CXF 3.0-M supports Websockets and Server-sent Events? I read the
migration but couldn't find any info.

Cheers.




* Melhores cumprimentos / Beir beannacht / Best regards *
*______________________________________________________*

*António Manuel dos Santos Mota <http://gplus.to/amsmota>*
*http://www.linkedin.com/in/amsmota* <http://www.linkedin.com/in/amsmota>
*______________________________________________________*

Re: Websockets and SSE

Posted by Sergey Beryozkin <sb...@gmail.com>.

Hi Aki
On 11/12/13 14:44, Aki Yoshida wrote:
> Hi,
> I worked on the camel transport using atmosphere for the server-side
> and async-http-client for the client-side.
> I got stuck at the point in naming the components or putting this
> behavior into one of the existing components.
> http://camel.465427.n5.nabble.com/DISCUSS-Re-update-of-websocket-td5742842.html
>
> For cxf, it makes sense to offer the websocket transport and somer
> people asked me already.
> I think, implementation-wise, it is straightforward.
> But we just have the same question of whether to have a separate
> transport or having it as a separate transport (e.g., ws) or an option
> in one of the existing ones (e.g, in http-hc for the client part and
> http or http-jetty or a new http-atmosphere for the server-side).
>
> I think both arguments have some points.
> Making a separate transport makes the code simpler to implement
> websocket functionality and also keeps its configuration simple.
> But the component is bound to one specific implementation and if
> someone wants a specific underlining implementation that is different
> from the one used, s/he will need an implementation specific component
> instead. But an implementation specific component that already
> provides its own transfer mode, adding a websocket mode makes it more
> complicated and so as its configuration.
>
> So, I would like to hear more opinions here.
>
May be http-websockets, with Jetty implementation by default and with 
the possibility to plugin Atmpshere, etc ?

Thanks, Sergey
> thanks.
> regards, aki
>
>
>
>
> 2013/12/11 Sergey Beryozkin <sb...@gmail.com>:
>> Hi
>>
>> On 11/12/13 12:08, António Mota wrote:
>>>
>>> Hi all.
>>>
>>> Does CXF 3.0-M supports Websockets and Server-sent Events? I read the
>>> migration but couldn't find any info.
>>>
>> We only have this JIRA so far
>> https://issues.apache.org/jira/browse/CXF-5339
>>
>> Aki is doing some work directly in Camel - please check it too;
>>
>> Though IMHO having it supported in CXF at some minimum level should also be
>> done eventually
>> Sergey
>>
>>> Cheers.
>>>
>>>
>>>
>>>
>>> * Melhores cumprimentos / Beir beannacht / Best regards *
>>> *______________________________________________________*
>>>
>>> *António Manuel dos Santos Mota <http://gplus.to/amsmota>*
>>> *http://www.linkedin.com/in/amsmota* <http://www.linkedin.com/in/amsmota>
>>> *______________________________________________________*
>>>
>>
>>
>> --
>> Sergey Beryozkin
>>
>> Talend Community Coders
>> http://coders.talend.com/
>>
>> Blog: http://sberyozkin.blogspot.com

Re: Websockets and SSE

Posted by Aki Yoshida <el...@gmail.com>.

Hi Sergey, Dan,
Thanks for your feedback.
I will try out a few things and try to get more clear feeling
concretely for the CXF case and how things will look.
regards, aki

2013/12/11 Daniel Kulp <dk...@apache.org>:
>
> Honestly, I’d likely just stick it in Jetty and -hc as the implementation would be bound to those two implementations.   If at some point in the future there is enough interest in adding additional implementations, we could pull it out into a common module or something.
>
> That’s just my opinion though.   :-)
>
> Dan
>
>
> On Dec 11, 2013, at 9:44 AM, Aki Yoshida <el...@gmail.com> wrote:
>
>> Hi,
>> I worked on the camel transport using atmosphere for the server-side
>> and async-http-client for the client-side.
>> I got stuck at the point in naming the components or putting this
>> behavior into one of the existing components.
>> http://camel.465427.n5.nabble.com/DISCUSS-Re-update-of-websocket-td5742842.html
>>
>> For cxf, it makes sense to offer the websocket transport and somer
>> people asked me already.
>> I think, implementation-wise, it is straightforward.
>> But we just have the same question of whether to have a separate
>> transport or having it as a separate transport (e.g., ws) or an option
>> in one of the existing ones (e.g, in http-hc for the client part and
>> http or http-jetty or a new http-atmosphere for the server-side).
>>
>> I think both arguments have some points.
>> Making a separate transport makes the code simpler to implement
>> websocket functionality and also keeps its configuration simple.
>> But the component is bound to one specific implementation and if
>> someone wants a specific underlining implementation that is different
>> from the one used, s/he will need an implementation specific component
>> instead. But an implementation specific component that already
>> provides its own transfer mode, adding a websocket mode makes it more
>> complicated and so as its configuration.
>>
>> So, I would like to hear more opinions here.
>>
>> thanks.
>> regards, aki
>>
>>
>>
>>
>> 2013/12/11 Sergey Beryozkin <sb...@gmail.com>:
>>> Hi
>>>
>>> On 11/12/13 12:08, António Mota wrote:
>>>>
>>>> Hi all.
>>>>
>>>> Does CXF 3.0-M supports Websockets and Server-sent Events? I read the
>>>> migration but couldn't find any info.
>>>>
>>> We only have this JIRA so far
>>> https://issues.apache.org/jira/browse/CXF-5339
>>>
>>> Aki is doing some work directly in Camel - please check it too;
>>>
>>> Though IMHO having it supported in CXF at some minimum level should also be
>>> done eventually
>>> Sergey
>>>
>>>> Cheers.
>>>>
>>>>
>>>>
>>>>
>>>> * Melhores cumprimentos / Beir beannacht / Best regards *
>>>> *______________________________________________________*
>>>>
>>>> *António Manuel dos Santos Mota <http://gplus.to/amsmota>*
>>>> *http://www.linkedin.com/in/amsmota* <http://www.linkedin.com/in/amsmota>
>>>> *______________________________________________________*
>>>>
>>>
>>>
>>> --
>>> Sergey Beryozkin
>>>
>>> Talend Community Coders
>>> http://coders.talend.com/
>>>
>>> Blog: http://sberyozkin.blogspot.com
>
> --
> Daniel Kulp
> dkulp@apache.org - http://dankulp.com/blog
> Talend Community Coder - http://coders.talend.com
>

Re: Websockets and SSE

Posted by Daniel Kulp <dk...@apache.org>.

Honestly, I’d likely just stick it in Jetty and -hc as the implementation would be bound to those two implementations.   If at some point in the future there is enough interest in adding additional implementations, we could pull it out into a common module or something.   

That’s just my opinion though.   :-)

Dan


On Dec 11, 2013, at 9:44 AM, Aki Yoshida <el...@gmail.com> wrote:

> Hi,
> I worked on the camel transport using atmosphere for the server-side
> and async-http-client for the client-side.
> I got stuck at the point in naming the components or putting this
> behavior into one of the existing components.
> http://camel.465427.n5.nabble.com/DISCUSS-Re-update-of-websocket-td5742842.html
> 
> For cxf, it makes sense to offer the websocket transport and somer
> people asked me already.
> I think, implementation-wise, it is straightforward.
> But we just have the same question of whether to have a separate
> transport or having it as a separate transport (e.g., ws) or an option
> in one of the existing ones (e.g, in http-hc for the client part and
> http or http-jetty or a new http-atmosphere for the server-side).
> 
> I think both arguments have some points.
> Making a separate transport makes the code simpler to implement
> websocket functionality and also keeps its configuration simple.
> But the component is bound to one specific implementation and if
> someone wants a specific underlining implementation that is different
> from the one used, s/he will need an implementation specific component
> instead. But an implementation specific component that already
> provides its own transfer mode, adding a websocket mode makes it more
> complicated and so as its configuration.
> 
> So, I would like to hear more opinions here.
> 
> thanks.
> regards, aki
> 
> 
> 
> 
> 2013/12/11 Sergey Beryozkin <sb...@gmail.com>:
>> Hi
>> 
>> On 11/12/13 12:08, António Mota wrote:
>>> 
>>> Hi all.
>>> 
>>> Does CXF 3.0-M supports Websockets and Server-sent Events? I read the
>>> migration but couldn't find any info.
>>> 
>> We only have this JIRA so far
>> https://issues.apache.org/jira/browse/CXF-5339
>> 
>> Aki is doing some work directly in Camel - please check it too;
>> 
>> Though IMHO having it supported in CXF at some minimum level should also be
>> done eventually
>> Sergey
>> 
>>> Cheers.
>>> 
>>> 
>>> 
>>> 
>>> * Melhores cumprimentos / Beir beannacht / Best regards *
>>> *______________________________________________________*
>>> 
>>> *António Manuel dos Santos Mota <http://gplus.to/amsmota>*
>>> *http://www.linkedin.com/in/amsmota* <http://www.linkedin.com/in/amsmota>
>>> *______________________________________________________*
>>> 
>> 
>> 
>> --
>> Sergey Beryozkin
>> 
>> Talend Community Coders
>> http://coders.talend.com/
>> 
>> Blog: http://sberyozkin.blogspot.com

-- 
Daniel Kulp
dkulp@apache.org - http://dankulp.com/blog
Talend Community Coder - http://coders.talend.com

Re: Websockets and SSE

Posted by Aki Yoshida <el...@gmail.com>.

Hi,
I worked on the camel transport using atmosphere for the server-side
and async-http-client for the client-side.
I got stuck at the point in naming the components or putting this
behavior into one of the existing components.
http://camel.465427.n5.nabble.com/DISCUSS-Re-update-of-websocket-td5742842.html

For cxf, it makes sense to offer the websocket transport and somer
people asked me already.
I think, implementation-wise, it is straightforward.
But we just have the same question of whether to have a separate
transport or having it as a separate transport (e.g., ws) or an option
in one of the existing ones (e.g, in http-hc for the client part and
http or http-jetty or a new http-atmosphere for the server-side).

I think both arguments have some points.
Making a separate transport makes the code simpler to implement
websocket functionality and also keeps its configuration simple.
But the component is bound to one specific implementation and if
someone wants a specific underlining implementation that is different
from the one used, s/he will need an implementation specific component
instead. But an implementation specific component that already
provides its own transfer mode, adding a websocket mode makes it more
complicated and so as its configuration.

So, I would like to hear more opinions here.

thanks.
regards, aki

2013/12/11 Sergey Beryozkin <sb...@gmail.com>:
> Hi
>
> On 11/12/13 12:08, António Mota wrote:
>>
>> Hi all.
>>
>> Does CXF 3.0-M supports Websockets and Server-sent Events? I read the
>> migration but couldn't find any info.
>>
> We only have this JIRA so far
> https://issues.apache.org/jira/browse/CXF-5339
>
> Aki is doing some work directly in Camel - please check it too;
>
> Though IMHO having it supported in CXF at some minimum level should also be
> done eventually
> Sergey
>
>> Cheers.
>>
>>
>>
>>
>> * Melhores cumprimentos / Beir beannacht / Best regards *
>> *______________________________________________________*
>>
>> *António Manuel dos Santos Mota <http://gplus.to/amsmota>*
>> *http://www.linkedin.com/in/amsmota* <http://www.linkedin.com/in/amsmota>
>> *______________________________________________________*
>>
>
>
> --
> Sergey Beryozkin
>
> Talend Community Coders
> http://coders.talend.com/
>
> Blog: http://sberyozkin.blogspot.com

Re: Websockets and SSE

Posted by Sergey Beryozkin <sb...@gmail.com>.

Hi
On 11/12/13 12:08, António Mota wrote:
> Hi all.
>
> Does CXF 3.0-M supports Websockets and Server-sent Events? I read the
> migration but couldn't find any info.
>
We only have this JIRA so far
https://issues.apache.org/jira/browse/CXF-5339

Aki is doing some work directly in Camel - please check it too;

Though IMHO having it supported in CXF at some minimum level should also 
be done eventually
Sergey
> Cheers.
>
>
>
>
> * Melhores cumprimentos / Beir beannacht / Best regards *
> *______________________________________________________*
>
> *António Manuel dos Santos Mota <http://gplus.to/amsmota>*
> *http://www.linkedin.com/in/amsmota* <http://www.linkedin.com/in/amsmota>
> *______________________________________________________*
>


-- 
Sergey Beryozkin

Talend Community Coders
http://coders.talend.com/

Blog: http://sberyozkin.blogspot.com