On Dec 8, 2011, at 4:46 PM, Sergey Beryozkin wrote:
>>>
>>
>> I don't think this is specified, should it be? A 400 makes more sense for a MBR.
>
> It appears it does make sense more, especially given that
> 400 is mandated for the JAXB one in case of empty payloads.
>
> I was thinking that may be failing to read does not necessarily mean it is a problem with a client request, but I could not think of any realistic case, may be a large correct payload causes a problem for a default InputStream provider but it seems hypothetical a bit; may be the spec should note that a default jaxb provider *SHOULD* as opposed to MUST throw 400, but I won;t insist
Please file a JIRA for this.
-- Santiago