On Thu, 2004-02-05 at 14:47, Andrew Welch wrote:
Regardless of the form of schema I use .NET creates
<ERROR a='data' ...></ERROR>
Rather than the expected form of
<ERROR a='data' .../>
The problem is compounded when <xsl:output indent='yes' /> is
used. In the first form the indentation whitespace makes the
instance document invalid against either form of schema. If
the second form were generated the indentation would not be an issue.
Really? So a schema can tell the difference between
<node></node>
and
<node/>
Is that really the case?
It ought to be, of course, but I don't think that's the
point. The OP seems to be saying that the validator is
objecting to the first form being indented, but not to
the second form being indented.
If that is so then I would guess the software is broken.
Indentation can only affect the context of the content
model of the *enclosing* element: once an element is
closed by its end-tag (NET or otherwise), basta, finito.
In any case, I don't see any indentation in the example
given.
More likely is that the element is declared as EMPTY,
which ought to mean what it says: no chance *ever* of
any content, because the special form of the start-tag
precludes it. Unfortunately this simple message failed
to go home, and <foo></foo> is taken to be the same as
<foo/>.
Those who remember what Schemas derived from may find it
ironic that it is Microsoft who may be breaking the data
model here.
///Peter
XSL-List info and archive: http://www.mulberrytech.com/xsl/xsl-list