Just a quick update: The RI works perfectly with my code so I think I have interpreted the spec correctly. I working on a fix in MyFaces but its slow going due to my unfamiliarity with the EL implementation.
sean On 9/26/05, Craig McClanahan <[EMAIL PROTECTED]> wrote: > > > On 9/26/05, Sean Schofield <[EMAIL PROTECTED]> wrote: > > Craig, > > > > I am using a explicity defined converter and everything is converting > > fine through the Process Validations phase. Its during Update Model > > phase that things go funky. Converters don't matter at this point do > > they? > > No ... by that time, the converted value will have been stored with > setLocalValue(), which should be of the appropriate data type to be stored > during Update Model Values. > > > I think there might be a bug in MyFaces > > ( http://issues.apache.org/jira/browse/MYFACES-623). > When I examined > > the ValueBindingImpl using my debugger it showed that it was > > specifically trying to coerce the value to class associated with > > java.lang.String . > > Hmm ... what does the RI do? > > > It looks like PropertyResolverImpl in MyFaces is checking if > > StatusImpl is an instanceof Map when it should be checking the result > > of StatusImpl.getData(). If it were checking the right object then > > we'd be ok (because the class that it tries to coerce to is the > > object's class.) > > Yah, the base object should be of type StatusImpl for the *first* property > resolution (dialog.data) but not for the second (data['foo']). > > > sean > > Craig > > --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
