Re: [xml] Reading CDATA

[Hartmut Sbosny <hartmut sbosny gmx de>]

Hi Daniel,

On Thursday, 25. September 2008 17:05, Daniel Veillard wrote:
> On Wed, Sep 17, 2008 at 06:48:37PM +0200, Hartmut Sbosny wrote:
> > Hello,
> > I am fresh to libxml. I want to read an xml file containing
> > the part
> >     <data>
> >             <![CDATA[...]]>
> >     </data>
> >
> > Currently I use the xmlParseDoc() interface. My first try was to read the
> > <data> node string via xmlNodeListGetString(). This returns something
>
>   You just can't using that API.

Is this a principle limitation of the xmlParseDoc API or only an 
accidental API lack? Which API is suitable to read cdata? Sorry for my
stumbling asking, I am new to libxml and xml.


> > > From there I could in princple extract the pure "..." content by
> > > subtracting the trailing white space node string, but it seems to me
> > > there should exist an easier libxml way to read the CDATA content?
>
> wei:~/XML -> cat tst.xml
> <data>
>    <![CDATA[...]]>
> </data>
>
> wei:~/XML -> xmllint --debug tst.xml
> DOCUMENT
> version=1.0
> URL=tst.xml
> standalone=true
>   ELEMENT data
>     TEXT compact
>       content=
>     CDATA_SECTION
>       content=...
>     TEXT compact
>       content=
> wei:~/XML ->
>
>  navigate in the tree and grab the data as content-> from the second
> child of your containing element

I probably miss the point. Do you mean I should use the command line 
tool 'xmllint'? This would be rather inconvinient for me. I have ready (more 
or less) a C program - using the xmlParseDoc API - which reads an XML file, 
where the cdata content is only a small piece of. As a workaround to get 
the cdata I evaluate currently the three strings
        "___...___"
        "...___"
        "___"
(in the said symbolic meaning) which I get when I parse through the 
children of the <data> element using xmlNodeListGetString(). Is that 
something I should not rely on?

Thanks for your response
Hartmut