Jelmer Vernooij [Mon, 16 Oct 2006 13:06:41 +0000 (13:06 +0000)]
r19339: Merge my 4.0-unittest branch. This adds an API for more fine-grained
output in the testsuite rather than just True or False for a
set of tests.
The aim is to use this for:
* known failure lists (run all tests and detect tests that
started working or started failing). This
would allow us to get rid of the RPC-SAMBA3-* tests
* nicer torture output
* simplification of the testsuite system
* compatibility with other unit testing systems
* easier usage of smbtorture (being able to run one test
and automatically set up the environment for that)
Simo Sorce [Mon, 16 Oct 2006 12:30:02 +0000 (12:30 +0000)]
r19337: never alloc on module unless you mean to attach a context to
it to keep the data around as long as the module lives
(This used to be commit d2073c1f7e1bc674358df5da0dc09e183b4b8712)
Simo Sorce [Mon, 16 Oct 2006 03:15:41 +0000 (03:15 +0000)]
r19314: Commit tridge's fixes for a big mem leak in ltdb I introduced
when the code has been changed to be async.
With the other committed fixes now this works.
(This used to be commit 49fc640b5c0398516ac3a9e3f7c55205cd60b1de)
Simo Sorce [Mon, 16 Oct 2006 02:45:18 +0000 (02:45 +0000)]
r19312: This should fix a nasty bug with values and names being freed before the results where used.
Seem that el is hanging out of a request, and the target is put on the final results.
Simo Sorce [Mon, 16 Oct 2006 01:00:04 +0000 (01:00 +0000)]
r19303: It seem ok to have around this amount of blocks hanging out of an ldb context for sam.ldb
(This used to be commit 5861ff69cc929c553e2c828ff42761cec6647d79)
Andrew Tridgell [Sun, 15 Oct 2006 21:31:58 +0000 (21:31 +0000)]
r19296: added a leak detector to ldb_wrap_connect()
this makes 'make test' _very_ noisy, apologies for that, but nearly
all the noise is real memory leaks, so if you don't like it then jump
in and help fix the leaks :)
(This used to be commit d791fd7b7b7a0a2674a70afe6cbbf0e3bd014544)
Andrew Tridgell [Sun, 15 Oct 2006 20:53:20 +0000 (20:53 +0000)]
r19294: a little speed tester for ldb/tdb
I wrote this to see if using a ldb for idmap store (the equivalent of
winbindd_idmap.tdb) makes sense. It's a long lived database that
administrators may wish to edit, so ldb makes sense, but at the moment
ldb is too slow. This test will help us speed it up.
(This used to be commit 10f8260455a55530f5701a136af8fa8b05ca8532)
Brad Henry [Sat, 14 Oct 2006 20:27:33 +0000 (20:27 +0000)]
r19281: Generalize what we do for the RAW-QFILEINFO test call so that the same process
is used for other tests that operate on a share when run against windows.
(This used to be commit 3d0cadb303fe9db49289d300a1c9fca44eee9d33)
Brad Henry [Sat, 14 Oct 2006 20:18:53 +0000 (20:18 +0000)]
r19279: These setup and cleanup functions are general enough for any test which needs
a share to operate on, not just the qfileinfo test, so now they are named
more appropriately.
(This used to be commit eb16d5324d6e1f4e4b33993e23a6446c3289a85f)
Brad Henry [Sat, 14 Oct 2006 19:25:14 +0000 (19:25 +0000)]
r19278: Change the delete_directory() expect function so that rmdir will
automatically remove any subdirs and files of the directory to be deleted.
(This used to be commit a41eefaaa738177f55d93b227f341f903a902939)
Brad Henry [Sat, 14 Oct 2006 14:50:30 +0000 (14:50 +0000)]
r19277: This commit adds support to run the smbtorture RPC-ASYNCBIND RPC-ATSVC
RPC-DSSETUP RPC-EPMAPPER RPC-INITSHUTDOWN RPC-LSA-GETUSER RPC-MULTIBIND RPC-ROT
RPC-SECRETS RPC-SRVSVC RPC-SVCCTL RPC-WKSSVC tests against windows, either
by 'make wintest' or in the build farm.
Andrew Tridgell [Sat, 14 Oct 2006 04:43:51 +0000 (04:43 +0000)]
r19273: - fixed error handling with the ldap backend
- propogate errors to the ldbadd command line tool
- use the rdn_name module when testing the tdb backend to allow the
same test code to correctly test the ldap and non-ldap backends
(This used to be commit dd82c474a123d90329bda653a4cb73c93e087b15)
Andrew Bartlett [Fri, 13 Oct 2006 13:01:48 +0000 (13:01 +0000)]
r19266: Add a target_hostname element to the binding struct. This allows us
to perform a lookup once, resolve the name to an IP, while still
communicating the full name to the lower layers, for kerberos etc.
This fixes 'net samdump', which was failing due to the schannel target
name being *smbserver.
Andrew Bartlett [Fri, 13 Oct 2006 07:25:51 +0000 (07:25 +0000)]
r19261: Fix use of unitialised variables. (The binding string is used, if not
NULL).
This showed up in a manual pre-TP3 test of the 'net samdump' code, and
shows the critical need for the windows testing infrustructure on the
build farm.
Derrell Lipman [Fri, 13 Oct 2006 02:49:16 +0000 (02:49 +0000)]
r19259: An ldb browser.
This has the layout complete so anyone interested can see what a qooxdoo
application might look like. It doesn't yet issue any JSON-RPC calls to do
anything useful (that's next), and it still has the appearance of a qooxdoo
test rather than something unique to Samba. I'll adjust the appearance later.
This layout is loosely (or not so loosely) based on 'gq'. A few things in
this layout won't apply to an ldb browser, but they're there for the time
being anyway until I decide exactly what to do with them.
Andrew Bartlett [Fri, 13 Oct 2006 01:35:52 +0000 (01:35 +0000)]
r19258: Don't delete the contents of the partitions twice, and in particular
don't delete their contents until we have specified the new partition
locations.
However, preserve the important part of tridge's change, that is to
ensure that no database index is present when the mass delete occours.
In my testing, it is best to leave the index until the provision is
compleated.
Andrew Tridgell [Thu, 12 Oct 2006 08:32:58 +0000 (08:32 +0000)]
r19252: - fixed 'erase' argument to setup_ldb()
- when wiping a ldb, wipe within each naming context first. By not
wiping the naming contexts we didn't wipe the partitions, which
caused a massive slowdown in re-provisioning due to re-indexing of
the schema.
(This used to be commit b62437214cf7c98c81598c4f37c91ab284928dbb)
Andrew Tridgell [Mon, 9 Oct 2006 11:13:49 +0000 (11:13 +0000)]
r19199: split out the xattr NTACL code into a separate part of the posix
backend, allowing other ACL backends to be added. The xattr backend is
still the default backend
(This used to be commit 90f044e63b12d32228310c7529382198bd7e6dfe)
Derrell Lipman [Sat, 7 Oct 2006 20:35:59 +0000 (20:35 +0000)]
r19167: - Various JSON-RPC facilities will desire to keep resources open in a
session. Provide a common method of doing this, which allows limiting,
somewhat (at least on a per-session basis) the possibility of DOS attacks.
Derrell Lipman [Fri, 6 Oct 2006 15:50:26 +0000 (15:50 +0000)]
r19142: ensure no race conditions during installation by having same name in source and swat directory; install new apps and services
(This used to be commit a2b996317f81aa61d7d5bf427003399560e64f77)