How specifically has this problem been fixed? I'm looking at the samples/faults example in v1.1 and I can see now that a RemoteException (not AxisFault) is being thrown, and that it has a bean typeMapping in the wsdd. But what control does a developer have over this mapping? This should be evident in the documentation, no?
I'm glad to see the improvement, but the fact is that the v1.1 docs don't list the fact that this has changed, let alone give much of a clue as to how it works beyond the exception class name going into faultCode and the statement "This is an area which causes plenty of confusion, and indeed, the author of this section is not entirely sure how everything works". I'm concerned that I may be building on top of other hacks I've made based on previous versions when in fact Axis has matured since its early days. This highlights the importance of a CHANGES.txt file or something similar with each release, as well as more coherent documentation (written by whomever does know how things work, before they move on from the project). -- Dan Kamins On Tue, 15 Jul 2003 04:43:54 -0700 (PDT), Davanum Srinivas wrote: >Dan, > >This problem has been fixed already in 1.1 > >thanks, >dims > >--- Dan Kamins <[EMAIL PROTECTED]>�wrote: >> >>I faced the same problem (can't throw app-specific exceptions) and >>solved it by subclassing >>AxisFault for all of my exceptions. �It allows me to explicitly set >>text for the actor, code, >>details, etc. fields of the SOAP fault fairly easily in the >>constructors of the exceptions. >> >>That being said, I am somewhat unhappy with this, because it >>prevents me from doing what seems >>proper, which you (Peter) said in a previous post: >> >>>keep my business logic interfaces (and the exceptions >>>which belong to it) as independent from the underlying >>>RPC technology as possible >> >>A nicer solution might be some form of type-mapping serialization >>setup as is done with bean >>mapping and serialization of other classes. �AFAIK this does not >>exist, and if nobody responds >>to say it does, I second it as a nice feature request, and you >>should enter it in. >> >>After all, avoiding vendor lock-in is one of the reasons people >>choose open source to begin >>with. >> >>-- >>Dan Kamins >> >>On Tue, 15 Jul 2003 10:15:28 +0200, Peter Landmann wrote: >>>Hi, >>> >>>Once again, as no one replied: Did really no one else face this >>>problem up >>>to now? Don't you use exceptions in your service interfaces? >>> >>>I took a look on the JAX-RPC specification 1.0 now: It defines >>>that >>>service >>>specific exceptions must extend java.lang.Exception - and not >>>something >>>like RemoteException or AxisFault. This means that Axis 1.1 >>>doesn't >>>comply >>>with the specification in this point, unfortunately. Should I >>>write >>>a bug >>>report? >>> >>>Regards, >>>Peter >>> >>>__________________________________________ >>>Peter Landmann >>>e-Business Services >>>System Engineer >>>Tel.: +49 89 898157-85 >>>Fax: +49 89 898157-49 >>>mailto:[EMAIL PROTECTED] >>>__________________________________________ >>>PEGAS systemhaus gmbh >>>Rudolf-Diesel-Str. 1 >>>D-82166 Gr�felfing >>>http://www.pegas.com/ >>>http://www.e-integrator.de/ >>>__________________________________________ >>> >>> >>> >> >> > > >===== >Davanum Srinivas - http://webservices.apache.org/~dims/ >
