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

Re: How the backend knows it's being searched by syncprov?



Pierangelo Masarati wrote:

Howard Chu wrote:


If we really need global knowledge of this control we can put the #define's and such back into slap.h and controls.c.
#define o_sync_mode ...
slap_cids.sc_LDAPsync


would probably be the most useful.


I see another problem: this control is dynamically registered, so o_sync_mode could yield unexpected results if the control is not registered but its presence is tested by some pice of code.

Yes, we could move its initialization / parser back into slapd/controls.c if necessary.


I guess only statically (i.e. always) registered controls should make it into the slap_control_ids structure, or a tri-state mech (present/not present/undefined) should be used for testing, where "undefined" could return the same value as "not present" but without crashing! (e.g. all "undefined" cids could point to slot [0] of o_controls, which should be intentionally left NULL for undefined controls?)

That sounds like a good idea.

--
 -- Howard Chu
 Chief Architect, Symas Corp.       Director, Highland Sun
 http://www.symas.com               http://highlandsun.com/hyc
 Symas: Premier OpenSource Development and Support