Why does Red Hat configure OpenLDAP to use GDBM instead of Berkeley DB?
OpenLDAP authors recommend the Berkeley DB for the backend and (from the
list archives) say that it will give better performance.
The only explanation I can see is this entry in the .spec file changelog:
* Mon Jul 10 2000 Nalin Dahyabhai <firstname.lastname@example.org>
- switch to gdbm; I'm getting off the db merry-go-round
What is that suppoed to mean?
Applications linked against DB2 which link against OpenLDAP compiled with
DB3 will die strangely because the two libraries provide functions with
the same names (__db_open and friends).
It's more of a namespacing issue until we can safely sort this multiple-
version mess out ("safely" means without having things break unexpectedly,
i.e., a major release, when breakage is expected). Marking as deferred.