Re: ldbmcat bug (ITS#211)

[michael@roka.net]

> You're patch appears to be reverse patch that backs out, in part,
> the previous ldbmcat changes.  I believe the current ldbmcat is
> actually correct; however, I've forward this note to our resident
> BerkeleyDB developer.

Oops, indeed my patch is in reverse.  If I call ldbm_datum_free with DB2,
then I get a SIGSEGV.  Looking at the code around it, it looked like I
shouldn't call it, so I patched it accordingly and the result works.

Michael