On Oct 28, 2006, at 13:35, Anne van Kesteren wrote:
On Fri, 27 Oct 2006 22:09:20 +0200, Simon Pieters
<[EMAIL PROTECTED]> wrote:
DOM3 Core says that they "must generate a fatal error during
serialization" (or, for the CDATA case, "the cdata section must be
splitted before the serialization"). Does that mean raise a
SYNTAX_ERR exception?
One idea would be to update DOM Level 3 Core to make sure you can
never get documents that are not serializable. I don't really know
if that's feasible though.
In that case, the HTML parsing section would need to be revised to
forbid element and attributes names that are not conforming XML 1.0 +
Namespaces local names, to forbid non-XML characters in character
data and attribute values and to forbid "--" in comments. Personally,
I'd welcome such a change, since it would truly make text/html an
alternative infoset serialization for a subset of XML 1.0.
Non-browsers that use XML tools to process HTML5 will have to enforce
those constraints anyway in one way or another. Current text/html
browsers don't, though.
Or did you mean that browsers would not enforce XML 1.0
serializability if the DOM was created by parsing text/html? Would
you then throw an exception if a subtree is imported from such a DOM
into a DOM that enforces serializability?
The exposure of CDATA sections in the DOM is, IMO, a design flaw in
the DOM. I wouldn't mind serializing them as normal character data.
--
Henri Sivonen
[EMAIL PROTECTED]
http://hsivonen.iki.fi/
