Re: [Springframework-rcp-dev] Bean methods that throw exceptions?

[Jan H. <jh...@sc...>]

Normally when any Exception pops up during the setting of a value, the
exception is caught by the DefaultFormModel.ValidatingFormValueModel
class and transformed to a bindingError. So if you get a error in your
message area, it probably came from there. Note that there are some
issues here: all listeners attached to the value change may also throw
exceptions that are handled in the same manner. I did suggest a change
here to provide an explicit binding exception that can be thrown in the
accessStrategy to separate the good from the bad (so to speak). Normally
this error should disappear if the value being set doesn't generate any
exception. Does this work as expected?

The getting of a value is another story. I don't think any handling is
available at that point so an exception will probably be caught by the
ExceptionHandler that you provide in your context.

The developer blog was started by Peter. He could add you as a writer to
the blog or we could post it in your name after a quick review. It would
be nice to see some more stuff explained there.

Additionally I've checked your project's website. It seems to be open
source so I'm thinking of checking it out. It would be easier to
actually see the problem than to write about it. I guess this isn't a
problem?

Kind Regards,
Jan

On Wed, 2008-01-30 at 11:27 +0200, Rogan Dawes wrote:
> Geoffrey De Smet wrote:
> > Forms are ussually meant for domain objects or transfer objects.
> > Just like it's pretty impossible to try to jpa-persist a (heavy) 
> > business service or DAO instance.
> > 
> > Wouldn't the use case you describe be better off if you
> > initialize the encodedBody before showing the form?
> > 
> > With kind regards,
> > Geoffrey De Smet
> 
> Let me give some more detail:
> 
> This is for WebScarab-NG, which is an HTTP testing tool (including an 
> intercepting proxy).
> 
> What I am modeling here is the HTTP Message (Request or Response), with 
> the following properties:
> 
> "startLine" ("GET http://example.com HTTP/1.0" or "HTTP/1.0 200 ok")
> "headers" (NamedValue[])
> "rawBody"
> "decodedBody"
> 
> "decodedBody" is the result of applying the various transforms listed in 
> the "headers" (e.g "Transfer-Encoding: chunked" or "Content-Encoding: 
> gzip") to "rawBody". I don't actually store "decodedBody", but rather 
> update "rawBody" by applying the various transforms. Unfortunately, the 
> process of applying the transforms can generate IOExceptions, since I 
> push the data through (Input|Output)Streams such as GZIPInputStream.
> 
> So, the idea is that the user can edit various properties of the HTTP 
> message before sending the message to the server, and this should 
> obviously include setting the "rawBody" (if desired) or "decodedBody" 
> (likely easier). Naturally, the two properties should remain in sync, 
> modulo the transforms.
> 
> I have implemented PropertyChangeSupport for this object, so various 
> forms can be notified when the fields they are listening to are changed 
> as a result of other fields being modified. This is all working very well.
> 
> The only problem now is what to do when the "(get|set)DecodedBody" 
> methods throw an Exception. My preference is simply to allow the user to 
> continue editing, with a warning/error message until no exception is thrown.
> 
> Thanks for your help.
> 
> Rogan
> 
> P.S. is there an existing place to present the results of some of the 
> advice that I have received on this list? e.g. how to do the 
> Binder/Binding to String<->byte[], how to actually bind variables to 
> JComponents correctly, etc? I don't have my own blog (although I did 
> find http://spring-rich.blogspot.com/)
> 
> > 
> > Rogan Dawes schreef:
> >> Hi folks,
> >>
> >> I am trying to deal with a derived property that can throw an exception 
> >> when set (or even when "get").
> >>
> >> The method signature is:
> >>
> >>    byte[] getEncodedBody() throws IOException;
> >>    void setEncodedBody(byte[] body) throws IOException;
> >>
> >> As suggested, there are some cases where invalid data will cause an 
> >> IOException to be thrown. When this happens, I get "methodInvocation" 
> >> showing up in my form message area.
> >>
> >> Is there some way of handling this exception, and propagating the 
> >> underlying exception to the user? This appears to be handled somewhere 
> >> in the BeanPropertyAccessStrategy or BeanWrapperImpl, although I can't 
> >> see exactly where.
> >>
> >> Any ideas?
> >>
> >> Rogan
> >>
> >> -------------------------------------------------------------------------
> >> This SF.net email is sponsored by: Microsoft
> >> Defy all challenges. Microsoft(R) Visual Studio 2008.
> >> http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
> > 
> > 
> > -------------------------------------------------------------------------
> > This SF.net email is sponsored by: Microsoft
> > Defy all challenges. Microsoft(R) Visual Studio 2008.
> > http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
> > _______________________________________________
> > Springframework-rcp-dev mailing list
> > Spr...@li...
> > https://lists.sourceforge.net/lists/listinfo/springframework-rcp-dev
> > 
> > 
> 
> 
> -------------------------------------------------------------------------
> This SF.net email is sponsored by: Microsoft
> Defy all challenges. Microsoft(R) Visual Studio 2008.
> http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
> _______________________________________________
> Springframework-rcp-dev mailing list
> Spr...@li...
> https://lists.sourceforge.net/lists/listinfo/springframework-rcp-dev


 **** DISCLAIMER ****
 http://www.schaubroeck.be/maildisclaimer.htm