| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
|
|
|
|
The talloc hirarchy here was a bit odd - we would both steal the
parsed ldif onto 'NULL', then reference it onto a python talloc
wrapper.
Now we just leave the reference, after we complete building the object.
Andrew Bartlett
|
|
The problem here was that some parts of the ldb_message were still
attached to the ldb_ldif structure, and when only the message was
taken (and the ldif free'ed to reclaim memory) we refereced free'ed
memory.
Andrew Bartlett
|
|
It seems quite reasonable to allow modules to re-initialise the set of
cached DNs on the ldb context.
Andrew Bartlett
|
|
When copying an attribute list, ensure the list itself is not NULL before
attempting to access elements of the list.
|
|
This uses groupType as the example, but this actually applies to all
integer types in AD.
Andrew Bartlett
|
|
We have made a lot of useful changes to LDB since the last realese,
that Samba4 now relies on. This ensures that a build against a system
LDB will only succeed against the right version.
Andrew Bartlett
|
|
While tdb has not changed ABI in a way that requires this, we don't
want Samba4 somehow built against the old version with
performance problems on large, growing databases.
Andrew Bartlett
|
|
Patch for bug #6389
|
|
Patch for bug #6388
|
|
Patch for bug #6269
When linking against tevent you also need to link against talloc. This patch
fixes external/libevent.m4 to do so.
|
|
This corrects the issues reaised in bug #6129, and some others that were not
originally identified. It also accounts for some code that was in the original
bug report but appears to have since been made common between S3 and S4.
Thanks to Erik Hovland <erik@hovland.org> for the original bug report.
|
|
|
|
|
|
|
|
breaking in gdb.
|
|
|
|
|
|
|
|
|
|
Found by: Андрей Григорьев <andrew@ei-grad.ru>
|
|
when sorting a record that does not have the requested sort attribute
then put those records last in the sort result
|
|
The sort module uses ldb_comparison_fold() as the comparison function
for case-insensitive attributes. In other places the function is being
used to produce a boolean, but for sorting we care about ordering.
The n1 - n2 return was sorting by length, not value
|
|
|
|
With unique indexes, any rename of a record that has an attribute that
is uniquely indexed needs to be done as a delete followed by an add,
otherwse you'll get an error that the attribute value already exists.
|
|
|
|
|
|
to each other.
|
|
When a attribute is marked unique we know that if we find a match
it will be the only possible match. This means that in a list of
subtrees connected by an &, it is best to first load the index values
for the unique entries, as if they find something then we know we
won't have to look any further.
This helps with searches like this:
(&(objectclass=user)(samaccountname=tridge))
the old code would first have loaded the very large index for the
objectclass=user attribute, and then loaded the single entry for
samaccountname=tridge. Now we load the samaccountname=tridge entry
first, notice that it gives us a single result, and stop, thereby
skipping the load of the objectclass=user index record completely.
|
|
When a attribute is marked as LDB_ATTR_FLAG_UNIQUE_INDEX then attempts
to add a 2nd record that has the same attribute value for this
attribute as another record will fail.
This provides a much more efficient mechanism for ensuring that
attributes like objectGUID are unique
|
|
This fixes a bug where we would look at an uninitialised
dn->linearized
|
|
|
|
When looking at performance problems with ldb it can be useful to see
which searches causes unindexed full searches. This makes it easy to
enable that.
|
|
one-level indexing was not always effective due to some broken logic
in the indexing code. This change means that if normal indexing fails,
we can still fall back on one-level indexing.
This reduces the number of full unindexed searches in s4 quite a lot
|
|
|
|
|
|
of LDB
(including the one in Samba 3).
|
|
The parentGUID test didn't clean up after itself correctly.
Andrew Bartlett
|
|
|
|
|
|
This is all working towards supporting the full WSPP schema without a
major performance penalty.
We now use binary searches when looking up classes and attributes. We
also avoid the loop loading the attributes into ldb, by adding a hook
to override the ldb attribute search function in a module. The
attributes can thus be loaded once, and then saved as part of the
global schema.
Also added support for a few more key attribute syntaxes, as needed
for the full schema.
|
|
|
|
This op will be used by the partition module to give us good
transaction semantics across the 4 partitions that sam.ldb uses.
|
|
it implicitly.
|
|
|
|
|
|
Fix for the problem was substitute talloc_strndup() with
talloc_memdup(), allocate 1 more character and put null character
('\0') in the extra place so data copied is null terminated.
Signed-off-by: Jelmer Vernooij <jelmer@samba.org>
|
|
|
|
I want to mix this with the server-side sort in particular.
Andrew Bartlett
|
