On Tue, 13 Feb 2007, Vincent Snijders wrote:
> Michael Van Canneyt schreef:
> >
> > On Tue, 13 Feb 2007, Graeme Geldenhuys wrote:
> >
> > > As a quick fix I wrote the following XSLT file to fix all those
> > > confusing xml "empty" elements.
> > > This XSLT file will rewrite all empty "descr", "errors" and "seealso"
> > > elements to the following format. <node />
> > >
> > > Here is how you can apply the XSLT under Linux.
> > > xsltproc -o output.xml remove_whitespace.xsl ticriteria.xml
> > >
> > >
> > > PS:
> > > Searching the internet it seems the XML preferred format for a empty
> > > element is a single tag with a '/' at the end. I guess we need to
> > > modify 'makeskel' then to output the correct/preferred format.
> >
> > No, that defeats the purpose of makeskel.
> >
> > The point is that the check for 'undocumented' in lazdoc checks for the
> > presence of a short/descr node. It should also check whether there is actual
> > text inside the node.
> >
>
> I haven't looked at the code, but this could be caused by the following.
>
> IIRC, in earlier versions, if there was only whitespace between the start tag
> and the end tag, no child nodes were generated. With fpc 2.1.1 for such
> 'empty' pieces a text node is generated, so LazDoc sees it as content.
That could also be the case. In fpdoc there is a piece of code to check for
whitespace-only content, that could be used to implement a more correct
test.
Michael.
_________________________________________________________________
To unsubscribe: mail [EMAIL PROTECTED] with
"unsubscribe" as the Subject
archives at http://www.lazarus.freepascal.org/mailarchives
