Some more thoughts...
Let's assume we keep the current implementation. What does it implies
for a move or a rename operation ?
If we are trying to move or rename a simple entry having children, we
may have a hug problem. Let's draft a sample, where uper case letters
are AP and lowercase are normal entries. I will describe the hierarchy
using a file system notation
rootDSE/
A/
a1
a2
B/
b1
b2/
C/
c1
c2
Now, let's try to rename b2... It will impact the subentries under C,
which will be renamed. So it will also impact the selected entries under
this renamed subentry. In which order should be proceed with the
modifications ? Currently, the move is done by simply changing one
element in the RDN index. But renaming a subentry implies that we change
all the associated references. We know have to check in all the b2
children if there is an AP and some subentries, and for each of them,
modify the operational attributes. Until this is done, we have pending
references pointing to nowhere, because the pointed subentries have been
renamed.... Problem !
It's exactly the same for a move operation...
Not exactly a nice situation :/
--
Regards,
Cordialement,
Emmanuel Lécharny
www.iktek.com
