You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@ws.apache.org by Tim McClure <Ti...@GE-interlogix.com> on 2002/08/08 22:44:18 UTC
Recommendations for passing errors and exceptions
Are there any standard recommendations for passing application error
messages or results back through an XML-RPC command? I know the XML-RPC
supports the concept of a fault message but this seems to be regulated to
XML-RPC service levels exceptions. Any input would be appreciated.
Thanks,
Tim McCLure
GE-Interlogix
Re: Recommendations for passing errors and exceptions
Posted by Steffen Schwigon <sc...@webit.de>.
Tim McClure <Ti...@GE-interlogix.com> writes:
> Are there any standard recommendations for passing application error
> messages or results back through an XML-RPC command? I know the XML-RPC
> supports the concept of a fault message but this seems to be regulated to
> XML-RPC service levels exceptions. Any input would be appreciated.
I wrote a wrapper around the real application interface. The wrapper
covers all those things xml-rpc does not as I like.
It returns the same type of answer for each call, i.e., a hashtable
consisting of a faultcode, faultmessage and the real result. That
allowed a consistent complementary unwrap and error handling on the
client side beside the "usual" effect of no further API changes if the
result value has to be enhanced for any reason.
GreetinX
Steffen
--
Steffen Schwigon <sc...@webit.de>
Re: Recommendations for passing errors and exceptions
Posted by Steffen Schwigon <sc...@webit.de>.
Tim McClure <Ti...@GE-interlogix.com> writes:
> Are there any standard recommendations for passing application error
> messages or results back through an XML-RPC command? I know the XML-RPC
> supports the concept of a fault message but this seems to be regulated to
> XML-RPC service levels exceptions. Any input would be appreciated.
I wrote a wrapper around the real application interface. The wrapper
covers all those things xml-rpc does not as I like.
It returns the same type of answer for each call, i.e., a hashtable
consisting of a faultcode, faultmessage and the real result. That
allowed a consistent complementary unwrap and error handling on the
client side beside the "usual" effect of no further API changes if the
result value has to be enhanced for any reason.
GreetinX
Steffen
--
Steffen Schwigon <sc...@webit.de>