CVS log for servers/slapd/back-bdb/dn2id.c

Up to [OpenLDAP] / servers / slapd / back-bdb

Request diff between arbitrary revisions

Keyword substitution: kv
Default branch: MAIN

Happy New Year

Remove lint (unused vars/labels/funcs/default:s)

ITS#6577 real fix: keep dn2id cursor open until we re-acquire lock
on entryInfo, otherwise BDB locks in reader txn are released too soon

happy belated New Year

Forced commit - previous (rev 1.169) was ITS#6474, not 6467

ITS#6467, use memcpy instead for strcpy when copying across array
boundaries. To account for stricter checks when using -D_FORTIFY_SOURCE=2
with GCC 4.5

dn2id_lock is no longer needed

More for ITS#6177

More for prev commit: delete bdb_cache_info(), must use cache_find_id().
Fixed cache.c_cursize counting.

ITS#6095 experimental fix: avoid purging other cases of in-use EntryInfos,
avoid using purged EntryInfos that have not yet been fully unlinked.

Don't underflow malloc'd memory in hdb_fix_dn()

ITS#5859 fix IDL cache for hdb DB with empty suffix

Update copyright notices

make messages more esplicative; fix style

Fix dn2idl when suffix = ""

ITS#5729 check for NULL bei_parent

silence warnings

Use read-only txn's instead of read lockers. Support BDB 4.4-4.7

Fix bei_state access
http://www.openldap.org/lists/openldap-devel/200805/msg00073.html

s/leave/func_leave/ to avoid OSF1 name clashing (ITS#5411)

Remove unused variables

Happy New Year (belated)

More for -q tool mode

Cleanup dup_compare

Drop unnecessary lock

Re-introduce hdb dup_compare function, default compares too much

More cleanup, hdb fixes

ITS#5262 must protect dn2id between the time it's read from the DB
and it's inserted into EntryInfo cache.

ITS#5262 additional cleanup

ITS#5262 use explicit cursor and locker in dn2id

More for BDB 4.6 compatibility

Quick and dirty BDB 4.6.18 support. Expect this to change again soon.

ITS#4838 fix fake root handling broken by prev commit

use dn2id index for extended filters using entryDN

Fix HDB - don't use IDL cache when it's not configured

Unify use of BDB lockers

happy new year

Happy new year!

Fix ITS#4263

ITS#4223 don't store negative results in IDL cache

In hdb_dn2idl_internal, skip cursor get when dkids == 1

ITS#4132 reset CACHE_ENTRY_ONELEVEL when complete. Fell out from a bad merge.

ITS#4088 force cursors to use same locker

Fix dn2id_add/del parent cache updates

More hdb optimizing - cache subtree IDLs, not just onelevels

More hdb tweaks, add radix sort code from mbackes@symas.com

More optimizing - try harder to avoid sorting

ITS#4029 Fix misplaced lines in prev commit

More back-hdb search optimization

Remove warning that HDB code is experimental and unsupported

Change diskNode.(n)rdn from unsigned char to char, matching their usage.

Fix another crasher if parent is NULL

Fix crasher in hdb_dn2id() when parent is NULL

Silence "unused <something>" warnings

Don't do pointer arithmetic on void*

Fix typo in prev commit

ITS#3773 additional fix

ITS#3773 fix RDN comparison

ITS#3688 don't sort range IDLs!

ITS#3647 back-hdb fixes, create unsorted subtree IDLs and sort later.

ITS#3647 fix hdb dn2idl SEGV

ITS#3616 use fake context entry if suffix is empty.

ITS#3638 Init ids to zero in bdb_dn2idl

ITS#3559 fix hdb dn2idl idl_cache usage

More for SLAP_TOOL_QUICK, disable locks as well

Byte-order independence for HDB: dn2id format changed again.

Start removing custom sort functions from hdb

Happy New Year!

Fix ITS#3424

Store entryIDs in big-endian order, no more custom sort function.
back-hdb still needs work.

Add SLAP_MR_ORDERED_INDEX - support for inequality indexing. Currently
only implemented for generalizedTime syntax.

trying to make names consistent

unifdef -UNEW_LOGGING

Make sure hdb gets signed chars for len comparison

SLAP_IDL_CACHE macro removed

cleanup

back-hdb fixes for ITS#3063

Revert to original code

unifdef SLAP_IDL_CACHE

Happy new year

Initial subordinate implementation (for chaining)

Notices

Fix prev commit setting NO_GRANDKIDS incorrectly

Fix HDB's use of IDL cache

Fix hdb cursor leak.

Fix alignment issues in hdb_dn2idl_internal

Fix back-hdb test006 failures due to bad return code in hdb_dn2idl

Cleanup prev commit

back-hdb caching updates, use EntryInfo cache for hdb_dn2idl.

Another BDB_HIER e_name / e_nname fix

Restructured to avoid direct use of sl_malloc functions. Minimizes impact
of #ifdef USE_SL_MALLOC...

Don't use single malloc for BDB_HIER e->e_name + e->e_nname

Don't re-cache IDL's retrieved from the cache

Fix HDB search_aliases crash, cleanup search scope processing: call dn2idl
directly instead of going thru filter_candidates. Use hdb_ function prefix
in dn2id.c to avoid ambiguities...

Use DB_RMW when deleting dn2id elements

Fix byte-ordering in bdb_dup_compare

Change RDN compare funcs, we don't need lexical ordering

Subtree rename support for the cache

Function renaming, make internal funcs static, etc.

Omit bei_rdn field from EntryInfo

filter_candidate tweaks, search_stack tweaks

Trim dead BDB_HIER stuff

Mem context tweaks for bdb_dn2idl

Fix typo in prev commit

Moved IDL cache code to standalone functions.
Use IDL cache in bdb_dn2id_children.

Pass parent EntryInfo instead of parent DN to dn2id_add/delete.
Add bdb_cache_children, remove cache flag manipulation from dn2id_children

memory context tweaks

Added CACHE_ENTRY_NO_KIDS flag to bei_state to shortcut dn2id_children.

Hierarchical cache management.

Tweak previous commit

Fix subtree and matched search in back-hdb

Disallow multiple suffixes by default. Otherwise subtree index optimization
breaks. #define BDB_MULTIPLE_SUFFIXES if you want multiple suffix support.
If you use it, subtree indexing will slow down by factor of 2.

Happy new year

Remove casts of AVL function pointers.

Call node_find_cmp() with (void *)&ulong instead of (void *)ulong.
The old code assumed sizeof(long) == sizeof(void *), among other things.

Updated BDB_HIER for current API

use ldap_charray_*() instead of charray_*()

Moved slap_strcopy, slap_strncopy to lutil_strcopy, lutil_strncopy

Coverted LDAP_LOG macro to use subsystem ID int values instead of string values

DB_RMW flag to dn2id and id2entry

Changed be_[n]suffix, be_[n]suffixAlias to BerVarray. No more bvec's anywhere.

Fix ITS#1774, BDB keys need ulen and flags to be initialized because BDB
tries to retrieve both the key and the data.

Added LDAP_LOG messages to operational.c and fixed compile error in dn2id.c

Added LDAP_LOG messages

Changed dnParent to void instead of int. (It always returned success...)

Changed be_issuffix and dnParent to struct bervals

Missed one db.bv_val in Debug statement

Straighten out some argument passing... cache code need more cleanup,
uses deprecated routines...

Entry caching. Based on ITS#1545 from Jong Hyuk Choi, jongchoi@us.ibm.com

Cleanup while loops in dn2id_add/delete

another round replacing dn_parent ...

fix ber_*str renaming

Fix some debug statements

Update copyright statements

More cleanup

Misc cleanup

Changed dn_rdn/dn_rdnlen to struct berval

More struct berval fixes

Some more (incomplete) struct berval conversion

Moved bdb_strcopy to slap_strcopy

Minor cleanup

More struct berval DNs

Fixed suffix DN_SUBTREE again. Index was still being written even though
it was not being read.

Fix check for DN_SUBTREE on be suffix

Remove lint and misc MSVC updates

Fix previous commit...

Cleanup previous commit

Added dn_rdnlen. Fixed rdn leak in limits.c.

Converted all use of dn_parent to the in-place version. (Fixed a pdn leak
in back-sql along the way.)

Switched to dn_parent1, cleaned up dn2id_add/delete

Changed be_nsuffix from char ** to struct berval ** to eliminate strlen's

Some fixes for BDB_IDL_MULTI. Experimental back-hdb code.

Missed a spot where root DN_SUBTREE index was still getting written.

Fix for BDB_IDL_MULTI to work.

Experimental index management. Uncomment BDB_IDL_MULTI in back-bdb.h
to activate. Needs to check limits in bdb_idl_insert_key, will get to
that later.

Minor rearrangement, give pdn to dn2id_add to avoid getting it twice.

Don't maintain a DN_SUBTREE index for the backend suffix, just return
BDB_IDL_ALL. This has a huge impact on I/O and txn logs.

Added experimental support for DB_DIRTY_READ.

Fixed initialization bug.

BDB updates...  haven't had a chance to work on this for a while,
committin what I have so that someone else might be able to work it.

Work in progress codes.  !UNTESTED!

working modrdn...
(need to fix leaks on transaction retries)

Clean up logging... need to hand run idl_delete to find bug

Mostly working modify

Rough in extended ops, modify, and modrdn.
General code cleanup.

Got base entry...

Rough in passwd and referral routines
add diag and fix a few idl bugs

Add more diagnostics including linear search version of idl_search()

Add bdb_bind, clean up init, s/Backend/BackendDB/

Add dn2entry.c

Add compare routine

rough in an add/delete and deadlock avoidance

Add bdb_add() and supporting routines

Add dn2id code
Add idl_insert

First round of imports from HEAD

Fix pointers

Fix NO_THREADS issues

Sync with HEAD, including
	- subclass indexing
	- bdb locker id fixes

mostly new logging

Suck in changes from HEAD
    Fix slapd schema error alignment bug
    Update slapd SASL proxy policy code
    Update slapd require "strong" check
    Update back-bdb entry locking
    Add slapd disallow bind_simple_unprotected directive
    Add back-monitor added monitorContext support
    Add slapd IPv4/v6 only options
    Ass slapd SASL setpass support
    Remove lint
    Build Environment
        Add monitor backend test
        Update <ac/socket.h> SHUT_RDWR
        Fix SASL-less build issues
        Fix back-perl build issues

Suck in latest changes from HEAD

Suck in HEAD changes since 2.1alpha

Import ITS#3688 fix from HEAD

ITS#3647 additional hdb dn2idl fixes

Import crash fixes for ITS#3638, 3647 from HEAD

ITS#3559 - import fix from HEAD

Happy new year!

Import HDB signed char fix from HEAD

Changes from HEAD, namely back-bdb updates
but also SASL, DN, matchedValues, and misc other fixes

Changes from HEAD, including
+       Fixed back-bdb ignore deadlock bug (ITS#3188)
+       Fixed back-bdb pagedResults no end cookie bug (ITS#3161)
+       Fixed back-bdb pagedResults clear controls bug (ITS#3182)
+       Fixed back-bdb pagedResults ignore control bug
+       Fixed slapd internal search limit bugs (ITS#3183)
+       Added slapd -l USER/DAEMON options (ITS#3187)
+       Build environments
+               Fixed back-sql trace build

Sync with HEAD

Happy new year

Sync with HEAD

Sync with HEAD

Merge in latest changes from HEAD

More for 2.2beta

Changes from HEAD for beta

Fix bei_state access

Happy new year! (belated)

More for ITS#5262, -q fixes from HEAD

ITS#5262 fixes from HEAD

happy new year

Happy new year!

Changes from HEAD suggested by Howard (1st round, less tests)

ITS#4223: don't store negative results in IDL

Sync with HEAD

Sync with HEAD

Sync with HEAD
Changes needs updating for release...

Sync with HEAD

Sync with HEAD

Sync with HEAD

sync with HEAD

Initial round of changes for 2.3 beta

Sync with HEAD

Sync with HEAD

Sync with HEAD

ITS#6758

Happy New Year!

ITS#6577, revert dn2id lock put in temporarily for 2.4.23 release

ITS#6577

ITS#6474

happy belated new year

dn2id_lock is no longer needed

ITS#6177

ITS#6095

Don't underflow malloc'd memory in hdb_fix_dn()

ITS#5859

Update copyright for next release

make messages more esplicative; fix style

silence warnings

ITS#5729

Use read-only txn's instead of read lockers. Support BDB 4.4-4.7

Fix bei_state access

ITS#5411

Happy new year! (belated)

Remove unused variables

Sync with HEAD

Sync with HEAD

Sync with HEAD

ITS#5262, sync with HEAD

Sync with HEAD for OL 2.4.5

Quick merge: everything from HEAD