Age | Commit message (Collapse) | Author | Files | Lines |
|
This avoids pulling the record and doing an allocation when we just
want to know if it exists.
Andrew Bartlett
|
|
This is all in a transaction, but when we are handling rename errors in the repl_meta_data
module, we key off the error, and do not close the transaction. We found that the
old record was gone and so could not try renaming it again to a conflict DN.
Andrew Bartlett
|
|
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
|
|
Otherwise the following surgery will break the SAMBA build and testsuite.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
|
|
This is essential for fixing up wrong ordered "objectClass" attributes.
Signed-off-by: Andrew Tridgell <tridge@samba.org>
|
|
Achieve this by introducing a "disallowDNFilter" flag.
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
|
|
We should ignore the LDB result.
Signed-off-by: Stefan Metzmacher <metze@samba.org>
|
|
|
|
This fixes a race when two processes initialise the same ldb database
at the same time. One of them could fail due to the other creating the
@BASEINFO record first.
Pair-Programmed-With: Andrew Bartlett <abartlet@samba.org>
Autobuild-User: Andrew Tridgell <tridge@samba.org>
Autobuild-Date: Tue Oct 18 03:54:42 CEST 2011 on sn-devel-104
|
|
Add the ecode arg to all the log functions, and log it.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
|
|
the spy code in ldb_tdb was added a while ago to overcome a memory
hierarchy problem with async ldb errors. Recently we started to get
valgrind errors related to the order of free in the spy code. This
patch ensures that we don't try to use a freed spy pointer. This
prevents the valgrind errors, although I suspect that the memory
hierarchy we have here is more complex than it needs to be
Pair-Programmed-With: Andrew Bartlett <abartlet@samba.org>
Autobuild-User: Andrew Tridgell <tridge@samba.org>
Autobuild-Date: Thu Sep 1 08:54:23 CEST 2011 on sn-devel-104
|
|
Autobuild-User: Andrew Bartlett <abartlet@samba.org>
Autobuild-Date: Sat Aug 13 13:49:44 CEST 2011 on sn-devel-104
|
|
value test should be performed or not
The function ltdb_add_internal is called either from ltdb_add or
ltdb_rename. In case of add we enforce the unique test (unless it has
been relaxed by a upper module through the
LDB_FLAG_INTERNAL_DISABLE_SINGLE_VALUE_CHECK flag), but for rename as it
is translated by a delete + a add we relax the test as we can have one
or more attribute which are supposed to be single valued but that are
not (ie. when we have a couple of deleted value on a single valued
attribute), we have already done the tests on insert so make the
assumption that the values are OK.
Without this patch deleting a subnet that has been affected to more than
one site fails as the delete is in fact a rename to GUID\0DEL ... with
an attribute siteObject that has 1 active link value and 1 inactive link
value
Autobuild-User: Matthieu Patou <mat@samba.org>
Autobuild-Date: Wed Jul 13 02:29:20 CEST 2011 on sn-devel-104
|
|
Signed-off-by: Andrew Tridgell <tridge@samba.org>
|