cxf-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Dan Diephouse <...@envoisolutions.com>
Subject Re: stacktrace in fault?
Date Tue, 24 Oct 2006 20:37:31 GMT
Glynn, Eoghan wrote:
>  
>   
>> -----Original Message-----
>> From: Dan Diephouse [mailto:dan@envoisolutions.com] 
>> Sent: 20 October 2006 18:06
>> To: cxf-dev@incubator.apache.org
>> Subject: Re: stacktrace in fault?
>>
>> James Mao wrote:
>>
>>     
>>> I'm thinking of adding a getStackTrace()/setStackTrace into Fault I 
>>> saw in XML binding we just set the stacktrace into the message, i 
>>> don't think that is correct way.
>>> Any idea?
>>>
>>> -james.
>>>
>>>
>>>       
>> I don't think we should be sending the stack trace along with 
>> the fault. 
>> We should log the stack trace on the server side and send 
>> information to the client that there was some error 
>> processing, but shouldn't necessarily tell them all the details.
>>
>> - Dan
>>     
>
> I'd disagree. In certain circumstances, I think sending on the wire the
> stack trace of the exception that gave rise to the fault would actually
> be very useful for debugging and problem reporting. It could lead to
> much quicker resolution of issues reported by client-side developers, as
> opposed to trawling thru' (potentially very large) server-side logs.
>
> I'd see this as analogous to the CORBA ExceptionDetailMessage service
> context, which is often used by Java ORBs to marshall the offending
> stack trace into the exception reply. As the content of the stack trace
> may be deemed "sensitive", say in the case of a secure server, this
> facility can usually be turned off via configuration.
>
> /Eoghan
>
>   
I would be happy with a configuration switch :-)

- Dan

-- 
Dan Diephouse
Envoi Solutions
http://envoisolutions.com
http://netzooid.com/blog


Mime
View raw message