At 4:13 PM -0500 3/7/04, Neil Graham wrote:
>I seem to recall you having noted that Xerces, Crimson, and Oracle all
>concurred that endDocument() was optional under severe error conditions.
>If so, what parsers exhibit the opposite behaviour?
GNU JAXP, Piccolo, and the other AElfred derivatives
>I think the main thing though is that even in this very limited survey
>we've uncovered someone for whom this would be a backward-incompatible
>change. As regrettable as this inconsistency might be, and whatever the
>right answer is, we should surely not be introduce backward-incompatible
>behaviours in what wasn't even intended to be a minor patch release. To my
>mind, such things should be reserved for some future SAX 2.1.
Perhaps. The current situation is untenable though. Some hard choices
need to be made, sooner or later. As it stands, SAX is severely
underspecified, parsers do exhibit varying behavior, and programs are
not interoperable between parsers (and this is hardly the only area
where that is true). The failure to produce a test suite in
conjunction with the specification was a major error, but one I hope
to begin rectifying soon.
--
Elliotte Rusty Harold
Effective XML (Addison-Wesley, 2003)
http://www.cafeconleche.org/books/effectivexml
http://www.amazon.com/exec/obidos/ISBN%3D0321150406/ref%3Dnosim/cafeaulaitA
