Re: [xml] redicting parts of trees

[Kasimier Buchcik <kbuchcik 4commerce de>]

Hi,

On Tue, 2005-05-24 at 08:17 -0400, Daniel Veillard wrote:
> On Tue, May 24, 2005 at 02:07:07PM +0200, Kasimier Buchcik wrote:
> > Hi,
> >
> > On Sun, 2005-05-22 at 22:04 +0200, cazic gmx net wrote:
> > > Hi,
> >
> > This non-wrapper @parent argument, what exactly should be its purpose
> > (besides for the ns-normalization)?
> >
> > 1. @node was previously added to @parent, so @parent is just a flag to
> >    activate the non-wrapper functionality
> >
> > 2. unlink @node and add it to @parent, which restricts only to
> >    appending; add-sibling or add-prev-sibling would not work, unless
> >    specified somehow
> >
> > 3. unlink @node and don't link to @parent, linking will be done by the
> >    user afterwards at the desired position in the children of @parent
>
>   I'm thinking of 3/, however if the node is already linked to parent
> then don't unlink it.

If 3/, I see the following problems, if @node is linked to @parent
already:
1. If @node is an attribute, we will loose ability to remove it from
   the ID list, since the previous parent is not known.
   So the user will be responsible for removing IDs.
2. The user might have tried by himself to add an ID to the
   destination-doc (if this attribute was an ID) before adoption;
   this could be incorrect; were are remapping entity-references, so
   the value of the attribute can change during adoption.

> > 4. something different? 
>
> Daniel

Regards,

Kasimier