If Namespaces processing is enabled then I believe that any
failure to conform to that specification is a fatal error.
Only things that depend on validation, like finding a colon
in ID/IDREF(S) would be recoverable errors.

Sounds like you have it right.

-Glenn



                                                                                       
                       
                    Andy Clark                                                         
                       
                    <andyc@apache.       To:     [EMAIL PROTECTED]           
                       
                    org>                 cc:                                           
                       
                                         Subject:     Re: Interested in getting 
changes into main Xerces      
                    10/18/2001            distribution                                 
                       
                    12:56 AM                                                           
                       
                    Please respond                                                     
                       
                    to                                                                 
                       
                    xerces-j-dev                                                       
                       
                                                                                       
                       
                                                                                       
                       



[EMAIL PROTECTED] wrote:
> I seem to have exposed a parser bug in Xerces 2_0_0_beta2.  Run the
> attached application through sax.Writer. Look at the attributes on the
> <xsl:element/> -- one missing, one replicated.

Thanks for the report. It appears that the namespace binder
in the pipeline was not correctly checking for the error
condition pointed to by Glenn in the namespace specification.
This caused some strange error to be reported using the most
recent code in CVS.

I have now added an error message and the code needed to
check for this error and emit the proper message. I report
this as a fatal error -- should it be non-fatal?

--
Andy Clark * IBM, TRL - Japan * [EMAIL PROTECTED]

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






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

Reply via email to