ITS#7926 - [18:19] <hyc> this is going to be a WONTFIX for 2.4, I think
[18:20] <hyc> reiniting the listeners array is probably too much of a pain. it will screw with all currently open connections
+
+[10:22] <hyc> and rewriting Bind and Exop result handling
+[10:22] <hyc> so that it's no longer a special case for overlays
+
+[23:34] <hyc> ok, so far heap profile shows that memory use during refresh is normal
+[23:35] <hyc> not wonderful, but normal. mem usage grows because we're recording the present list while receiving entries in the refresh
+[23:36] <hyc> I'm seeing for 1.2GB of data about 235MB of presentlist
+[23:36] <hyc> which is pretty awful, considering presentlist is just a list of UUIDs
+[23:36] <hyc> being stored in an avl tree
+[23:37] <hyc> a btree would have been better here, and we could just use an unsorted segmented array
+[23:42] <hyc> for the accumulation phase anyway. we need to be able to lookup records during the delete pphase
+[00:05] <hyc> this stuff seriously needs a rewrite
+[01:13] <hyc> 2.8M records x 16 bytes per uuid so this should be no more than 48MB of overhead
+[01:13] <hyc> and instead it's 3-400MB
+
+<hyc> we need to merge the rest of Ondrej's slapmodify/slapdelete patches
+<hyc> get lmdb incremental backup working
+<hyc> I need to publish a spec for PREPARE/2-phase commit in LDAP TXNs and integrate that into lmdb and back-mdb
+
+Update slapo-unique to have an option to disable manageDSAit (ITS#8245)
+
+Leftover bits from ITS#8294 -- Drop custom SHA2 code, use default crypto library SHA2 functions instead
+ May need to document minimum crypto library version requirements
+ Can also include sha2.c in slapd-sha2.c and make everything static
+
+Matching rules for olc* attrs, see also ITS#8341
+
+New syslog engine
+
+[12:45] <hyc> I meant regcomp and regexec
+[12:45] <hyc> mainly - compile a pattern, then fill in substitution parameters later
+[12:46] <hyc> yes but the idea is still based on numbered substitution parameters
+[12:47] <hyc> it's a pretty obvious approach, and with regex, time-tested.
+[12:48] <hyc> I may try to sneak it into OpenLDAP 2.5
+
+
+Overhaul overlays and backends, see ITS 6166
+
+(In relation to a discussion about slapo-chain)
+<hyc> anyway, the nicer ting to fix would be in 2.5, push all of the repl consumer code into its own overlay
+<hyc> in that case, updateref would be processed wherever the overlay was configured
+<hyc> so no longer tied to the frontend
+<hyc> it would also make it more feasible to have multiple different consumer configs in a single DB, each with their own provider URL (and thus their own updateref)
+<hyc> I would think we can get rid of the update ref directive entirely, just point all writes to that consumer's provider.
+
+Update dynlist so it works with search filters, so that autogroup can be deleted
projects / openldap / blobdiff