Hi Ulrich, I understand your requirements. The parser will validate the content of an element after it has processed all of its children (no content validation of parent is taking place as we process each child). Since the parser is validating the parent element, it is generating an error with the line/column of that parent element (closing tag). I agree that the real error happens at the first D-Tag, and it would be easier if the line/column no of the error message represented that tag. That's a nice feature to have, and it would require either changing the validation mechanism so that the content of an element is validated each time we process a new child, or store the line/column notinfo of those children and use that info when generating the error messages. We have to look into that and ensure that we do not impact performance or add processing overhead.
Khaled Ulrich Arnold wrote: > Hi Khaled ! > > Thanks for your response. I realize that the line-position regarding > "element invalid for content-model" is the same for both cases. The > parser actually generates it not at the end-tag of the root, but at the > end tag of the parent-element. I understand that the parser cannot be > sure of the error before the next higher end-tag. But still the real > error position is at the D-Tag (either opening or closing-tag). Other > parsers, e.g. XML-Spy by Altova, are able to point to the real > error-location, not to the point where the parser was sure it was an > error. I need the error-line to provide the error-position to the > end-user via a GUI, so I need the posion of the cause of the error. > > Uli > > -----Ursprungliche Nachricht----- > Von: Khaled Noaman [mailto:[EMAIL PROTECTED]] > Gesendet: Montag, 29. April 2002 16:21 > An: [EMAIL PROTECTED] > Betreff: Re: BUG 6875 > > The error message that the parser produces regarding the invalid content > model, is generated when the parser validates the content of the Root > element. That's why you are getting the line/column no of that element. > If you check both test cases, the parser is generating an invalid > content error using the same line/column no. In the first test case, the > parser has no idea of element X and so it generates an extra error > message. In the second case, element D is already known, and so the > parser does not generate an error message, and the error message is only > generated when validating the content of Root. So, I would not consider > this as a bug. > > Khaled > > --------------------------------------------------------------------- > 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]
