Re: commit: ldap/libraries/libldap result.c

[Howard Chu <hyc@symas.com>]

Pierangelo Masarati wrote:
> Hallvard B Furuseth wrote:
> > Mm.  LDAP_OPT_NETWORK_TIMEOUT was implemented, though.  So the change
> > breaks code which uses that option.  And ldap_get_option() of the two
> > options should work the same way.
> >
> > I too would have preferred if ldap_get_option() did less mallocing
> > (since there would be less cleanup required afterwards), but I think the
> > right place to fix it would be in the Next Generation API we sometimes
> > mumble about.
>
> I think the best solution would be to change the way timeouts are
> stored, leave the old API in place (i.e. malloc stuff returned by
> ldap_get_option() for LDAP_OPT_*TIMEOUT, and add new LDAP_OPT_X_*TIMEOUT
> (call it whatever makes sense) that don't return malloc'ed stuff.

I've restored the malloc behavior for LDAP_OPT_*TIMEOUT. I wasn't too keen on 
adding new Option flags for retrieving the same setting, so I left that alone.

--
  -- Howard Chu
  Chief Architect, Symas Corp.  http://www.symas.com
  Director, Highland Sun        http://highlandsun.com/hyc
  Chief Architect, OpenLDAP     http://www.openldap.org/project/