Re: Precision and error handling (was URL work in HTML 5)

Hi Eric,

You wrote:

> The idea of the HTML parser is that it always produces a DOM that can
> serialize out into a document conforming to the author spec.

No, that's not the idea at all. Consider the following document:

===============
<!doctype html>
<xmp>
hi Eric
===============

When parsed, it produces a DOM tree that looks like so:

#document
| <HTML>
|   <HEAD>
|   <BODY>
|     <XMP>
|       "hi Eric"

When serialized, the resulting document does not meet the author
conformance criteria in the spec.


Ted

Received on Monday, 8 October 2012 22:27:17 UTC