[jsr-314-open] Adding a description field to the JavaScript event payload

[Jim Driscoll]

On 6/11/09 11:37 AM, Andy Schwartz wrote:
> Yep, agree with Kito and Dan that we need to do whatever we can to
> reduce confusion when problems like this arise, so happy to see a more
> detailed error description made available. (Not sure that we need a
> description field for "events" - but definitely for errors.)

I agree that a description field may be overkill for an event.



> On a related note... So have we started to identify common problems that
> might lead to client-side errors such as the "MalformedXML" error? Are
> there certain issues/approaches that application developers should be
> aware of in order to avoid running into such errors? Are there some best
> practices that we should recommending to reduce the chance that such
> errors will be seen? If so, it would be good to collect this knowledge
> and make it available to our users.

The "render=false" error seems to be the major (user error) reason to 
see the error at this point.  All other occurances of that error that 
I've seen have been RI errors.

That doesn't mean that there aren't a lot of other ways that the user 
can make XML errors.  Once they start writing XML from the server back 
to the client, it's bound to happen sooner or later.

Since we have writers that enforce well formed XML, the majority of 
errors you'll see will then be updating or deleting fields that don't 
exist, or improperly CDATA escaping payload text.

But I think that for the standard components, we've found most of the 
common code path bugs.  *crosses fingers*

> Or are we seeing MalformedXML errors in cases where the application
> developer has done everything right, but something is going wrong at the
> framework/implementation/browser level?

Up until a month ago, bugs in the RI made me see that error frequently. 
  I think I've squashed all the common path errors in the RI that 
generated that error at this point.

> Kito -
>
> Were you able to track down the source of the MalformedXML error that
> hit during your demo?

I'd also be interested in hearing about that.  The RI errors seem to be 
mostly easy to fix, it's just a matter of finding them.

Jim