back-bdb2 <-> back-ldbm

[Hallvard B Furuseth <h.b.furuseth@usit.uio.no>]

It looks like a lot of extra work to maintain almost equal code in
back-ldbm and back-bdb2.  There are already differences that look like
they ought to be independent of the backend.

I think back-bdb2 should either be merged back into back-ldbm or merely
use a lot of the files from back-ldbm.  Most of the differences look
like they can be handled with macros, something like

    #if <back-bdb2>
    # define ldbm_thread_trylock        (void)
    # define ldbm_start_timing          bdb2i_start_timing
      ...
    #else
    # define ldbm_thread_rdwr_wtrylock  ldap_pvt_thread_rdwr_wtrylock
    # define ldbm_start_timing          (void)
      ...
    #endif

-- 
Hallvard