+1 :-) i don't like filters and don't like depending on the class-path
ordering

Scott O'Bryan schrieb:
> Mike, the reason for the API's I'm proposing is to solve a usecase we
> can't currently support.  Largely because of file uploads (but also some
> other things as well), it is impossible for two renderkits to somtimes
> even exist in the same webapplication as each other.
> 
> For instance, let's say you use Tobago for one page and Trinidad for
> another.  If Tobago exists first in your classpath and Trinidad exists
> second in your classpath, you will only ever be able to us the Tobago
> multi-part form handling.  By unifying these, it would still be up to
> the renderkit to have their own file-upload component, it's just the
> multi-part form handling could be done by the commons.
> 
> Furthermore, all of the Renderkits today have a filter implementation. 
> Renderkits like Trinidad and Tobago are getting away from these as much
> as possible.  Other then being difficult to configure, Filters do not
> port well over to Portal's.  Both Tobago and Trinidad have come up with
> their own methods for doing some of this filter logic inside of custom
> FacesContextFactories, but this work is error prone, non-trivial, and is
> subject to the class-path ordering issues I explained earlier.  By
> unifying the mechanism whereby this wrapping happens, we can concentrate
> on replacing as much of this legacy "filter" functionality as possible.
> 
> Scott
> 
> Mike Kienenberger wrote:
>> At least in terms of the validators/converters/etc (non-api) parts of
>> commons, it is not intended that Tobago or any other project depend on
>> these.   These are additional components made available to the end
>> users.   So it's true that some small subset of Tobago users won't be
>> able to use it.   The same is true for Trinidad and Tomahawk users.
>>
>> I don't think that's reason enough to require that these projects
>> support JSF 1.1.
>>
>> I still think the conglomeration of developer-targeted utilities/apis
>> and end-user-targeted components is a mistake, and I was under the
>> impression that each of these pieces was independent of each other.
>> If that's the case, there's no reason why the api/utils
>> developer-targeted projects can't maintain JSF 1.1 support while
>> leaving the component end-user-targeted projects at JSF 1.2.
>>
>> On Dec 5, 2007 5:08 PM, Bernd Bohmann <[EMAIL PROTECTED]> wrote:
>>  
>>> The tobago codebase will be 1.1 or 1.2 compatible as much as possible.
>>> If commons only supports 1.2 we are not able to use it in tobago until
>>> the 1.1 version is only in maintenance mode.
>>>
>>> I think it's ok if some parts of commons are 1.2 only.
>>> I think jdk 1.4 compatibility is not a requirement.(We can provide a
>>> retrotranslated version)
>>>
>>> As far I know some of the tobago users still using 1.1 jsf and not able
>>> to switch to 1.2 until end of 2008.
>>>
>>> Andrew Robinson schrieb:
>>>
>>>    
>>>> As the vote states, if -1, please provide a reason why 1.1 has to be
>>>> supported. An argument of why not is not enough.
>>>>
>>>> On Dec 5, 2007 2:14 PM, Bernd Bohmann <[EMAIL PROTECTED]>
>>>> wrote:
>>>>      
>>>>> -1
>>>>>
>>>>> I don't see any reason why a commons fileupload should not support 1.1
>>>>>
>>>>> Can someone define what commons API means?
>>>>>
>>>>> Is this just a subproject of commons like commons validator or commons
>>>>> converter?
>>>>>
>>>>> Scott O'Bryan schrieb:
>>>>>
>>>>>        
>>>>>> +1
>>>>>>
>>>>>> Mario Ivankovits wrote:
>>>>>>          
>>>>>>> +1
>>>>>>>            
>>>>>>>> Lets make the myfaces commons JSF API an official vote so we can
>>>>>>>> have
>>>>>>>> a fixed time frame on this decision
>>>>>>>>
>>>>>>>> +1 [ ] -- make JSF 1.2 the minimum requirement for the new myfaces
>>>>>>>> commons project
>>>>>>>> +0 [ ] -- you don't mind supporting a 1.1 trunk in addition to a
>>>>>>>> 1.2
>>>>>>>> trunk
>>>>>>>> -1 [ ] -- you feel that 1.1 should be required and why you feel
>>>>>>>> that
>>>>>>>> it is needed
>>>>>>>>
>>>>>>>> My vote: +1
>>>>>>>>
>>>>>>>> -Andrew
>>>>>>>>
>>>>>>>>
>>>>>>>>               
>>
>>   
> 
> 

Reply via email to