Re: (ITS#5035) Warning patrol

[h.b.furuseth@usit.uio.no]

I listed two "#7"s, let's call the second #8.

Items #3-5 and 7 done. (#5 and #7 in 2.3.37, forgot to update CHANGES.)

Pierangelo Masarati writes:
>h.b.furuseth@usit.uio.no wrote:
>> #1. slapd/frontend.c:frontend_init() has the statement (...)
>
> Probably a leftover of a time when slap_known_controls was a pointer

Likely a non-bug then.  Leaving it there as a reminder until someone
(me?) gets around to checking that the current code is OK.

>> #2.  The return value from op->o_callback->sc_cleanup() is never
>> used.  Should it be?  I noticed because the pcache_op_cleanup()
>> function does not return anything.  What should it return?
>
> two possibilities:
> a) ignore it: cleanup could occur too late, when response has already
> been returned to client; in this case a return code is there only
> because of the standard API
> b) handle it: cleanup return code could be usefully returned to caller
>
> if (a), then the return code can be safely ignored

If (a), I suggest the prototype should be changed to return void.
I'll leave that to whoever knows overlays well though.

For now, Ralf updated pcache_op_cleanup() to return SLAP_CB_CONTINUE,
and I've copied that to RE23.

-- 
Regards,
Hallvard