Jann Forrer wrote:
[...]
What is not clear to me is how the structure of the site is
determined in the JCR approach. For example is the sitetree of the
default publication determined by the design of the repository or
is it defined externally e.g. by a sitetree.xml file?
IMO that should be entirely up to the site manager.
The TreeSiteManager might store an extra XML file (which would be
a bad idea because of performance), use an extra node structure, or
use the implicit node structure of the repository.
Anyway,
<map:generate src="content/{pe:area}/sitetree.xml"/>
should be replaced with
<map:generate type="site-structure">
<map:parameter name="area" value="{pe:area}"/>
<map:parameter name="root" value="/foo/bar"/>
<map:parameter name="language" value="{pe:language}"/>
</map:generate>
This would use the SiteManager to produce an XML representation
of the site structure.
Can the data be stored in the repository anyhow and
are ardressed by the UUID in the second case? Is this even possible in a
JCR-approach?
I guess so. I discussed the node structure with Felix and Josias and
IIRC we came to the following conclusion:
- the nodes should be stored in a hierarchical manner which corresponds
to the URL space
- this results in an implicit site structure (equivalent to the
TreeSiteManager)
What about visibility or more generally attributes of a node(there is
still a bug in "create language version")?
Should we store attributes of a node as properties of a node in the
repository? This certainly make sense in the case where the structure is
defined by the repository.
That would be up to the site manager.
I guess using repo nodes for site nodes and properties for site node
attributes would be a reasonable approach.
-- Andreas
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
