cxf-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Sergey Beryozkin <sberyoz...@gmail.com>
Subject Re: 3.1.0 regression from 3.0.4
Date Wed, 06 May 2015 14:56:33 GMT
Hi,
On 06/05/15 10:19, Romain Manni-Bucau wrote:
> 2015-05-06 11:13 GMT+02:00 Sergey Beryozkin <sberyozkin@gmail.com>:
>
>> Sure, having it configurable would be one option, will take care of it.
>> Another option is to register a custom provider.
>>
>
> yes this is more or less the same since you redefine the priitive provider
> but would be sad to have to impl this one yourself as a user ;)
I've just realized this provider is already configurable indirectly, it 
has consumeMediaTypes and produceMediaTypes properties, setting both to 
*/* will do it.

FYI, I've just updated PrimitiveTextProvider to support Boolean/boolean, 
etc only but not String. I had to do it in order to annotate 
PrimitiveTextProvider with text/plain (to start getting CXF-6307 
addressed), but String has to be supported for */* by default.

Thanks, Sergey
>
>
>> Thanks, Sergey
>>
>>
>>
>>
>>
>> On 06/05/15 09:59, Romain Manni-Bucau wrote:
>>
>>> I see,
>>>
>>> maybe something doesn't hurting much CXF would be to add a flag in the
>>> primitive provider. This way users could at least configure it to go back
>>> to the previous behavior.
>>>
>>>
>>> Romain Manni-Bucau
>>> @rmannibucau <https://twitter.com/rmannibucau> |  Blog
>>> <http://rmannibucau.wordpress.com> | Github <
>>> https://github.com/rmannibucau> |
>>> LinkedIn <https://www.linkedin.com/in/rmannibucau> | Tomitriber
>>> <http://www.tomitribe.com>
>>>
>>> 2015-05-06 10:54 GMT+02:00 Sergey Beryozkin <sberyozkin@gmail.com>:
>>>
>>>   Hi Romain
>>>>
>>>> Looks like you have a very nice set of tests which is impressive :-)
>>>>
>>>> That was a 'side-effect' of getting a spec requirement enforced. It is
>>>> about  section 4.2.6,
>>>> "java.lang.Boolean , java.lang.Character , java.lang.Number Only for
>>>> text/plain"
>>>>
>>>> Personally I think it is too restrictive but a TCK test is failing
>>>> against
>>>> CXF right now, which is partially related to this requirement, hence this
>>>> restriction needs to be enforced first for the final test solution to be
>>>> done.
>>>> Now that it is done it would be sensitive still continue to support
>>>> Boolean/etc for any media type by default - we've had some troubles with
>>>> the matching algorithm offering a more flexible sub-resource selection in
>>>> some of the edge cases by default earlier...
>>>>
>>>> Thanks, Sergey
>>>>
>>>>
>>>> On 06/05/15 09:23, Romain Manni-Bucau wrote:
>>>>
>>>>   Hi guys,
>>>>>
>>>>> org.apache.cxf.jaxrs.provider.PrimitiveTextProvider#isSupported changed
>>>>> when upgrading to also validate the media type.
>>>>>
>>>>> This sounds better but it is actually a regression since before it was
>>>>> supported with any media type.
>>>>>
>>>>> A solution could be to have it twice in the writer chain, once with this
>>>>> restriction at the beginning and another time without this restriction
>>>>> at
>>>>> the end as a fallback.
>>>>>
>>>>> Not sure it does worth a fix but was a small side effect I experienced
>>>>> upgrading.
>>>>>
>>>>> Romain Manni-Bucau
>>>>> @rmannibucau <https://twitter.com/rmannibucau> |  Blog
>>>>> <http://rmannibucau.wordpress.com> | Github <
>>>>> https://github.com/rmannibucau> |
>>>>> LinkedIn <https://www.linkedin.com/in/rmannibucau> | Tomitriber
>>>>> <http://www.tomitribe.com>
>>>>>
>>>>>
>>>>>
>>>>
>>>
>>
>


Mime
View raw message