DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG 
RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT
<http://nagoya.apache.org/bugzilla/show_bug.cgi?id=19868>.
ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND 
INSERTED IN THE BUG DATABASE.

http://nagoya.apache.org/bugzilla/show_bug.cgi?id=19868

Exception handling in HttpClient requires redesign





------- Additional Comments From [EMAIL PROTECTED]  2003-07-13 22:47 -------
Laura, Oleg, Mike, et al,
Generally I like the way things are sounding with this, but having had a recent 
experience trying to 
track down the cause of an exception in Xerces I'm really not so keen on ever seeing a 
nested 
exception again - particularly if it's only added for conceptual niceness.

In particular, I don't like the idea of wrapping every IOException in a 
HttpTransportException, 
instead I would have dropped HttpTransportException and used IOException itself.  The 
few 
IOExceptions that are recoverable would have to be wrapped in a 
HttpRecoverableException but 
the vast majority of exceptions that people would be trying to track down the source 
of, would be 
plain ol' unwrapped exceptions.

It also seems to fit better with my view on exception handling that you don't wrap 
exceptions when 
possible.  In essence, an exception is thrown and works its way back up the stack 
until someone is 
prepared to deal with it.  That doesn't work so well when you consider throwing a 
different 
exception as "dealing with it".

I don't really mind either way, but I found it incredibly difficult to have to wade 
through long stack 
traces *and* work out which exception actually occurred and which was just created 
because the 
API didn't allow the original exception to be thrown.  In fact, I wound up having to 
grab the source 
to Xerces and rewrite parts of it's exception handling to work out what was going on 
(for the 
record, it was my fault).

Sorry for coming to the party late, I'm happy to be ignored on this and I forfeit all 
rights to 
complain. :)

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to