Re: LibXML2 Indentation

[James Henstridge <james daa com au>]

Ali Akcaagac wrote:

> On Wed, 2003-04-30 at 00:23, Havoc Pennington wrote:
>  
>
> > > <stringvalue>9</stringvalue>
> > >      
> > ...
> >    
> >
> > >  <stringvalue>
> > >   9
> > >  </stringvalue>
> > >      
> > The difference is that in the second case the string stored is "\n 9\n
> > " and in the first case it is "9".  If you didn't do this, you
> > couldn't store strings containing whitespace.
> >    
>
> Yes right, I could have thought about this one scenario myself and yes
> this is quite correct and maybe a bad choosen example from my side. But
> we could indent complete TAG's in general e.g.
>
> <?xml version="1.0"?>
> <gconf>
> <entry name="accuracy" mtime="14" muser="galaxy" type="string">
>  <stringvalue>9</stringvalue>
> </entry>
> </gconf>
>
> and this one gives us no sideeffects the Tag's are indented more human
> readable and the data is placed between the start and endtag saving us
> from having eiter $a, $d or both (e.g. on Windows that adds two bytes
> for linebreaks). Isn't this possible ?
In your example above, the child list of the <entry> element is the text 
node "\n  " followed by a <stringvalue> element followed by a text node 
"\n ".

For the white space not to matter, the application will need to handle 
the cases where it is present.  For example, if the application is 
expecting the first child node of <entry> to be the <stringvalue>, then 
it will fail if whitespace is included.

The only place in an XML file where whitespace truely is insignificant 
is within tags.  For instance, the following two fragments are equivalent:
 <foo bar="baz"> </foo>
and:
 <foo
 bar="baz"> </foo>

James.

--
Email: james daa com au
WWW:   http://www.daa.com.au/~james/