On 04.06.2010 16:51, Geoffrey Carman wrote:
> One of the cool things about IDM is that things sometimes just work, and
> you might not think it would, but it does.
> See my Structured GCV article, where you get different results in a
> string vs nodeset context. (VERY different, but GREAT different!)
> http://www.novell.com/communities/no...ion-values-idm
> So how about ParseDN?
> My colleague figured it would just work, if he for-eached through a
> nodeset, and parseDN'ed it in the nodeset definition.
> I kind of figured it might work as well, but it looks like he gets a
> single string.
> I guess that means ParseDN is more of a string only token? Might be
> interesting to list off the tokens that do things different in
> string/nodeset or only work in each one... Hmm...

I think it's completely correct that parse-dn doesn't create a node-set
in the for-each context

A better solution (in my opinion) is that token-parse-dn should allow
variables in the length and start attributes, that would make iterating
using a do/while loop over all nodes in a DN easier.

Maybe this should be suggested as an enhancement.