| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
|
|
Using DLIST_ADD_END() to construct a long list is very inefficient (it
is O(n^2). These lists are not ordered, so using DLIST_ADD() is much
better.
|
|
This helps ensure we don't load the schema too often in the provision
(allowing a reference in of the schema before the modules load).
Andrew Bartlett
|
|
This allows us to load the schema against one ldb context, but apply
it to another. This will be useful in the provision script, as we
need the schema before we start the LDAP server backend.
Adnrew Bartlett
|
|
Before this change, the first opener of the sam ldb context would
become the owner of the global schema, then the autofree context got a
reference to the schema. Any subsequent opens of the sam ldb also got
a reference. This meant that the talloc hierarchy was inconsistent
between the first sam ldb open and subsequent opens. With this change
the autofree context becomes the owner of the global schema, and all
ldb contexts get a reference.
|
|
|
|
|
|
|
|
|
|
Sadly it still segfaults at this stage
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 removes a level of indirection via external binaries in the
provision-backend code, and also makes better use of our internal code
for loading schema from an LDIF file.
Remaining to do: Sort the output again, as the load from LDIF is
unsorted (also needed because the normal LDB load from sorted input is too slow
anyway, and is only needed here).
Andrew Bartlett
|
|
do not reference it from ldb.h
|
|
The only 2 modules escaping the rule so far are rootdse and partitions
|
|
3.
|
|
|
|
The previous ldb_search() interface made it way too easy to leak results,
and being able to use a printf-like expression turns to be really useful.
|
|
This also tries to simplify the logic in the schema -> @ATTRIBUTES and
@INDEXES code.
Andrew Bartlett
(This used to be commit a383b8bf88a5681f9c9c6839ba645c872a735051)
|
|
I think it is just too complex and error prone to init and cancel
transactions during the module init code. Instead, this isn't prone
to races as it will always achieve a steady state (eventually), and
most cases will never do the write.
Andrew Bartlett
(This used to be commit d60977cc7f89f89f34187f310c91d1ab7db6ccf2)
|
|
Something in the search stack adds a distinguisedName record, which
isn't in the message we generate. So we compare, fail and rewrite the
record - every time ldb starts up...
Andrew Bartlett
(This used to be commit 44775d1ed4a4b8edc66a06e2b3710aba6a0dd019)
|
|
This ensures that a rudementary schema is always present (for
bootstrapping), and that the indexes are maintained equal to the
schema (rather than hard-coded).
Andrew Bartlett
(This used to be commit 747d683b0d92c3b1cde67245d514977a2c87dc44)
|
|
This changes the @ATTRIBUTES record to be for bootstrapping only,
before we find the schema.
Andrew Bartlett
(This used to be commit 358477fcc041d5fb2e6ac5641c2f899cc49cfb69)
|
|
This should make schema manipulation a little easier to follow.
Andrew Bartlett
(This used to be commit 300ed83526e75d834bd23ddd1c1c26ebe2555e0f)
|
