On Wed, 2006-10-25 at 08:47 +1000, Andrew Bartlett wrote:
> On Tue, 2006-10-24 at 18:37 -0400, simo wrote:
> > On Wed, 2006-10-25 at 07:31 +1000, Andrew Bartlett wrote:
> > > On Tue, 2006-10-24 at 20:20 +0000, vlendec@xxxxxxxxx wrote:
> > > > Author: vlendec
> > > > Date: 2006-10-24 20:20:39 +0000 (Tue, 24 Oct 2006)
> > > > New Revision: 19486
> > > >
> > > > WebSVN:
> > > > http://websvn.samba.org/cgi-bin/viewcvs.cgi?view=rev&root=samba&rev=19486
> > > >
> > > > Log:
> > > > Probably Coverity is wrong here, but this fixes their ID 317. Not sure
> > > > whether
> > > > to merge it to 4_0. I want it in 3_0 because it took a bit to persuade
> > > > myself
> > > > that el can not be NULL here.
> > > I can't see how it could ever be NULL. If ldb_msg_add_value returns
> > > LDB_SUCCESS, then an attribute by that name exists, and
> > > ldb_msg_find_element must return it.
> > If this kind of operation is used enough, we can think of creating
> > another support function that directly returns el on attribute creation
> > and use that function instead of the 2 ones here.
> I think that would be worthwhile. It would avoid looping over the
> attribute list twice.
Ok, changed ldb_msd_add_value() and ldb_msg_add_empty() to return the
element if a pointer to contain it is provided.
Testing and committing.
Samba Team GPL Compliance Officer