[Date Prev][Date Next] [Chronological] [Thread] [Top]

Duplicate entries



Hi.

I've been browsing the sources looking for possible flaws that allow
duplicate entries. I posted an ITS (#1219) to keep track of this 
debugging. I noticed something that I'm not sure it caused my problem,
but in my opinion it may be leaky. In back-ldbm/add.c there's a call to
dn2id with the ndn of the entry being added, to check if it's already
there. This function, in case the dn2id.dbX file cannot be opened,
returnd NOID, which is the magic number ldbm_back_add expects to 
start adding the entry. I think the error code should be separated
from the output of the routine. Is it OK if I go ahead or is anybody
playing with that code yet?

Pierangelo.