[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: and nXML mode
From: |
Yuri Khan |
Subject: |
Re: and nXML mode |
Date: |
Mon, 9 Aug 2021 12:57:50 +0700 |
On Mon, 9 Aug 2021 at 09:22, Jean-Christophe Helary
<lists@traduction-libre.org> wrote:
> Is there a reason why nXML mode refuses to consider entities as legit
> in a document that starts with:
>
> <!DOCTYPE html>
> <html xmlns="http://www.w3.org/1999/xhtml">
If you view that as an XML document (which is what nXML deals with),
without any preconceived knowledge of HTML5, there is nothing to
suggest that is legit.
In XML, an entity can be defined inline within the doctype declaration:
<!DOCTYPE html [
<!ENTITY nbsp "&#a0;">
]>
or by reference to an external entity definition:
<!DOCTYPE html
PUBLIC "-//W3C//DTD XHTML 1.1//EN"
SYSTEM "http://www.w3.org/TR/xhtml11/DTD/xhtml11.dtd">
(In the HTML5 spec, this is referred to as “obsolete permitted DOCTYPE
string”, and the obsoletion is from the HTML5 point of view. I.e. if
you use an HTML5-aware parser, <!DOCTYPE html> is sufficient to
declare an HTML5 document.)
If you fetch that url, you will see that it references a number of
modules, and if you chase references far enough, you will get to
http://www.w3.org/MarkUp/DTD/xhtml-lat1.ent which contains this as its
first significant line:
<!ENTITY nbsp " " ><!-- no-break space = non-breaking
space, U+00A0 ISOnum -->
and that’s what makes a valid entity reference in an XHTML document.
(XML processors normally have some shortcuts, such as DTD pre-cached
in the so-called XML catalog, so that they don’t have to fetch them
from the network each time. XML catalog is keyed by the PUBLIC and/or
SYSTEM identifiers but not by the doctype root element name.)