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

Re: removing global_schemacheck and SLAPD_LASTMOD flags



At 12:24 PM 5/11/2005, Howard Chu wrote:
>Kurt D. Zeilenga wrote:
>>These flags should always be TRUE.  If FALSE, slapd(8)
>>will likely not behave correctly (especially in the
>>case of global_schemacheck).
>>
>>I am about to add some manage DIT capabilities and
>>rather not have to deal with the possibility that
>>these flags are FALSE.  Hence, I propose axing
>>these flags.  Comments?
>I agree with global_schemacheck.

Okay, this is one that's causing most of my concern.

>Not sure which LASTMOD flag you're referring to, or what the problem there is.

Mostly SLAP_LASTMOD().

>re: schema checking, we currently disable it on backends for sparse replication reasons (proxycache, syncrepl). I think we could make this a little more granular; probably if we only skip the test for required attributes (to allow sparse/partial entries) everything else would still work.

I rather mark the entries as shadow/cache copies and use that
marker to disable checks (that need only be performed on
master copies) than to have backends (or whatever) disable
checks.

Kurt