Jelmer Vernooij [Wed, 28 Sep 2005 13:47:39 +0000 (13:47 +0000)]
r10576: Add testprog for named pipes. Also add GNUmakefile's for cross-compilation
using mingw32
Stefan Metzmacher [Wed, 28 Sep 2005 09:58:58 +0000 (09:58 +0000)]
r10569: add all info that we have to wrepl_name, as we need it for replication :-)
metze
Andrew Bartlett [Wed, 28 Sep 2005 05:38:20 +0000 (05:38 +0000)]
r10566: Clean up error messages to provide more accurate info.
Andrew Bartlett
Andrew Bartlett [Wed, 28 Sep 2005 04:50:02 +0000 (04:50 +0000)]
r10565: Try to make Kerberos authentication a bit more friendly.
This disables it for 'localhost' as well as for any host our KDC does
not recognise.
Andrew Bartlett
Andrew Bartlett [Wed, 28 Sep 2005 02:58:53 +0000 (02:58 +0000)]
r10564: Make the RPC-SCHANNEL test use the libnet_join code via torture_join_domain
Handle error cases in torture_create_testuser, where we can't connect
to the target server (we were segfaulting due to an untested error
path).
Andrew Bartlett
Andrew Bartlett [Wed, 28 Sep 2005 02:37:03 +0000 (02:37 +0000)]
r10563: a null 'join' is a no-op.
Andrew Bartlett
Andrew Bartlett [Wed, 28 Sep 2005 02:22:31 +0000 (02:22 +0000)]
r10562: Ensure we initalise the error table with hdb errors. This ensures we
get good text error strings.
Andrew Bartlett
Andrew Bartlett [Wed, 28 Sep 2005 01:09:10 +0000 (01:09 +0000)]
r10561: This patch takes over KDC socket routines in Heimdal, and directs them
at the Samba4 socket layer.
The intention here is to ensure that other events may be processed while
heimdal is waiting on the KDC. The interface is designed to be
sufficiently flexible, so that the plugin may choose how to time
communication with the KDC (ie multiple outstanding requests, looking
for a functional KDC).
I've hacked the socket layer out of cldap.c to handle this very
specific case of one udp packet and reply. Likewise I also handle
TCP, stolen from the winbind code.
This same plugin system might also be useful for a self-contained
testing mode in Heimdal, in conjunction with libkdc. I would suggest
using socket-wrapper instead however.
Andrew Bartlett
Stefan Metzmacher [Tue, 27 Sep 2005 16:53:08 +0000 (16:53 +0000)]
r10547: - add wrepl_request timeout handling
- when we got an unexpected READ event, we need to do a socket_recv() to find connection errors
and we need to mark the socket as dead (and remove the fde_event) to prevent,
endless loops on broken connections
tridge: we should look carefull at other protocol, to handle broken connections without spinning
metze
Stefan Metzmacher [Tue, 27 Sep 2005 16:20:17 +0000 (16:20 +0000)]
r10545: map ECONNRESET to NT_STATUS_CONNECTION_RESET
metze
Stefan Metzmacher [Tue, 27 Sep 2005 13:31:17 +0000 (13:31 +0000)]
r10542: if the transport is dead we need to return
tridge: I think this is correct, comments?
metze
Stefan Metzmacher [Tue, 27 Sep 2005 13:04:07 +0000 (13:04 +0000)]
r10540: fix compiler warning
metze
Stefan Metzmacher [Tue, 27 Sep 2005 12:54:08 +0000 (12:54 +0000)]
r10537: - we now use a much nicer way to handle talloc_free(timed_event)
the events code replaces a destructor to one that returns allways -1
while it's calling the event handler
- we don't need the composite and winsrepl specific fixes any more
- this also fixes the problem with smbcli, dcerpc, cldap, ldap and nbt
request timeouts
metze
Andrew Tridgell [Tue, 27 Sep 2005 11:59:39 +0000 (11:59 +0000)]
r10535: fixed the pidfile code (it didn't survive the recent pstring changes)
Jelmer Vernooij [Tue, 27 Sep 2005 11:10:57 +0000 (11:10 +0000)]
r10533: Eliminate another use of next_token()
Jelmer Vernooij [Tue, 27 Sep 2005 11:02:06 +0000 (11:02 +0000)]
r10532: Replace next_token() with str_list_make()
Jelmer Vernooij [Tue, 27 Sep 2005 10:32:49 +0000 (10:32 +0000)]
r10531: (hopefully) improve detection of socket-related functions in external libraries
Stefan Metzmacher [Tue, 27 Sep 2005 10:31:57 +0000 (10:31 +0000)]
r10530: - fix some crash bugs when we lost the connection...
metze
Stefan Metzmacher [Tue, 27 Sep 2005 10:29:13 +0000 (10:29 +0000)]
r10529: fix a crash bug in full async code the uses the composite_trigger_done() code
the event subsystem wants to free timed_events!
metze
Jelmer Vernooij [Tue, 27 Sep 2005 10:00:27 +0000 (10:00 +0000)]
r10528: Add credentials.h back into includes.h as some compilers don't
seem to be able to handle incomplete enum types.
Andrew Tridgell [Tue, 27 Sep 2005 07:11:33 +0000 (07:11 +0000)]
r10527: don't attempt self gdb attach if running under valgrind. This was
causing fort to get rather unhappy
Andrew Tridgell [Tue, 27 Sep 2005 05:11:14 +0000 (05:11 +0000)]
r10526: BASEDIR must be set or we end up installing most of the binaries into lib/
Andrew Tridgell [Tue, 27 Sep 2005 03:11:08 +0000 (03:11 +0000)]
r10525: change from AC_CHECK_TYPES() to AC_CHECK_TYPE() for intptr_t, so the
type is always available, which means we need less #ifdefs
Andrew Tridgell [Tue, 27 Sep 2005 03:09:38 +0000 (03:09 +0000)]
r10524: SAFE_FREE() in tdb does not need the discard_const_p()
the discard_const_p() was causing problems on openbsd where intptr_t is not defined
Andrew Tridgell [Tue, 27 Sep 2005 02:36:56 +0000 (02:36 +0000)]
r10523: fixed timegm() to not depend on get_time_zone(), so it works in lib/replace/
the old timegm() replacement was also broken (it returned the wrong value)
Andrew Tridgell [Tue, 27 Sep 2005 01:26:34 +0000 (01:26 +0000)]
r10522: finally got the locking working on solaris10. This adds a read lock on
the transaction lock in tdb_traverse_read(). This prevents a pattern
of locks which triggers the deadlock detection code in solaris10. I
suspect solaris10 is trying to prevent lock starvation by granting
locks in the order they were requested, which makes it much easier to
produce deadlocks.
Jelmer Vernooij [Tue, 27 Sep 2005 00:11:21 +0000 (00:11 +0000)]
r10521: Also check sys/socket.h for definition of socklen_t (needed for AIX)
Andrew Bartlett [Mon, 26 Sep 2005 22:27:44 +0000 (22:27 +0000)]
r10520: The join is a nice quick RPC test.
Andrew Bartlett
Jelmer Vernooij [Mon, 26 Sep 2005 18:16:38 +0000 (18:16 +0000)]
r10517: Get rid of use of next_token() in lib/samba3/
Jelmer Vernooij [Mon, 26 Sep 2005 18:16:23 +0000 (18:16 +0000)]
r10516: Add seperator argument to str_list_{make,join}_shell()
Jelmer Vernooij [Mon, 26 Sep 2005 18:15:24 +0000 (18:15 +0000)]
r10515: Handle replacement of "domain logons" and "domain master" by "server role"
Jelmer Vernooij [Mon, 26 Sep 2005 17:42:12 +0000 (17:42 +0000)]
r10514: Add str_list_make_shell() and str_list_join_shell()
Jelmer Vernooij [Mon, 26 Sep 2005 16:57:08 +0000 (16:57 +0000)]
r10513: Reduce some use of pstring. The main reason some parts of the code still
use pstring is next_token() now.
Jelmer Vernooij [Mon, 26 Sep 2005 15:59:43 +0000 (15:59 +0000)]
r10510: Decrease the amount of data included by includes.h a bit
Jelmer Vernooij [Mon, 26 Sep 2005 15:15:50 +0000 (15:15 +0000)]
r10509: Some more sconscript fixes. Now getting to link stage for smbclient
Stefan Metzmacher [Mon, 26 Sep 2005 13:42:42 +0000 (13:42 +0000)]
r10508: - implement WINBINDD_NETBIOS_NAME, WINBINDD_DOMAIN_NAME and WINBINDD_INFO
is there a way to test this calls?
vl: please review this
metze
Jelmer Vernooij [Mon, 26 Sep 2005 13:12:02 +0000 (13:12 +0000)]
r10507: Let lib/replace take care of defining socklen_t for heimdal
Stefan Metzmacher [Mon, 26 Sep 2005 13:02:47 +0000 (13:02 +0000)]
r10506: set return data correctly
metze
Stefan Metzmacher [Mon, 26 Sep 2005 11:47:55 +0000 (11:47 +0000)]
r10504: - seperate implementation specific stuff, from the generic composite
stuff.
- don't use SMBCLI_REQUEST_* state's in the genreic composite stuff
- move monitor_fn to libnet.
NOTE: I have maybe found some bugs, in code that is dirrectly in DONE or ERROR
state in the _send() function. I haven't fixed this bugs in this
commit! We may need some composite_trigger_*() functions or so.
And maybe some other generic helper functions...
metze
Tim Potter [Mon, 26 Sep 2005 07:36:26 +0000 (07:36 +0000)]
r10500: More progress to getting ldb tools building. Create a list of ldb modules
to link in. There's still a whole lot of random other junk (dcerpc, gensec,
ndr stuff, lp_foo - what a mess).
Tim Potter [Mon, 26 Sep 2005 07:34:09 +0000 (07:34 +0000)]
r10499: Export cli_ldap library.
Tim Potter [Mon, 26 Sep 2005 07:33:55 +0000 (07:33 +0000)]
r10498: Detect some more networking headers.
Hardcode value of RETSIGTYPE for now.
Diddle with order of including SConscript files to make ldb happier.
Tim Potter [Mon, 26 Sep 2005 07:32:14 +0000 (07:32 +0000)]
r10497: Export the list of dsdb ldb modules.
Andrew Tridgell [Mon, 26 Sep 2005 02:03:06 +0000 (02:03 +0000)]
r10496: - added configure test for sys/time.h
- print length of failed locks
Andrew Tridgell [Mon, 26 Sep 2005 01:46:02 +0000 (01:46 +0000)]
r10495: older redhat boxes need sys/time.h for select()
Andrew Tridgell [Mon, 26 Sep 2005 01:37:46 +0000 (01:37 +0000)]
r10494: - don't generate a tdb log message for any type of failed lock probe
- if the lock upgrade loop ever fails then log a warning
Andrew Tridgell [Mon, 26 Sep 2005 01:28:49 +0000 (01:28 +0000)]
r10493: we need sys/select.h to enable select() in the solaris workaround
Andrew Tridgell [Mon, 26 Sep 2005 01:12:12 +0000 (01:12 +0000)]
r10492: work around a bug in solaris which cases lock upgrades to fail with
EDEADLK even when progress can be made. This is not a good solution,
but I can't find anything better.
Volker Lendecke [Sun, 25 Sep 2005 21:01:56 +0000 (21:01 +0000)]
r10491: First step towards wbinfo -t: This issues a name request for the primary
domain and gets the DC's name via a mailslot call.
Metze, I renamed wbsrv_queue_reply to wbsrv_send_reply in accordance with
irpc_send_reply. Having _queue_ here and _send_ there is a bit confusing. And
as everything is async anyway, the semantics should not be too much of a
problem.
Volker
Andrew Tridgell [Sun, 25 Sep 2005 13:17:03 +0000 (13:17 +0000)]
r10490: - allow deferred irpc replies to set the status
- add an example of deferred reply for echodata in LOCAL-IRPC
Andrew Tridgell [Sun, 25 Sep 2005 13:01:26 +0000 (13:01 +0000)]
r10489: added the ability for irpc server to defer replies instead of replying
immediately. They set m->defer_reply = True;
Andrew Bartlett [Sun, 25 Sep 2005 12:38:07 +0000 (12:38 +0000)]
r10488: (Missing file from previous commit, adding a new RPC-JOIN test)
This is a merge of Brad Henry's 'net join' rework, to better perform
an ADS join, particularly as a DC. This represents the bulk of his
Google SOC work, and I'm very pleased to intergrate it into the tree.
(Metze will intergrate the DRSUAPI work later).
Both metze and myself have also put a lot of time into this patch, and
in mentoring Brad in general. In return, Brad has been a very good
student, and has taken the comments well.
Since it's last appearance on samba-technical@, I have made
correctness and valgrind fixups, as well as adding a new 'BINDING'
mode to the libnet_rpc routines. This allows the exact binding string
to be passed down from the torture code, including options and exact
target host.
Jelmer Vernooij [Sun, 25 Sep 2005 12:27:36 +0000 (12:27 +0000)]
r10487: Some small updates to the manpage.
Andrew Bartlett [Sun, 25 Sep 2005 12:26:07 +0000 (12:26 +0000)]
r10486: This is a merge of Brad Henry's 'net join' rework, to better perform
an ADS join, particularly as a DC. This represents the bulk of his
Google SOC work, and I'm very pleased to intergrate it into the tree.
(Metze will intergrate the DRSUAPI work later).
Both metze and myself have also put a lot of time into this patch, and
in mentoring Brad in general. In return, Brad has been a very good
student, and has taken the comments well.
Since it's last appearance on samba-technical@, I have made
correctness and valgrind fixups, as well as adding a new 'BINDING'
mode to the libnet_rpc routines. This allows the exact binding string
to be passed down from the torture code, including options and exact
target host.
Andrew Bartlett
Andrew Tridgell [Sun, 25 Sep 2005 00:14:45 +0000 (00:14 +0000)]
r10485: run autoheader before autoconf
Andrew Tridgell [Sun, 25 Sep 2005 00:13:25 +0000 (00:13 +0000)]
r10484: try to fix the pread/pwrite declaration problems
Andrew Tridgell [Sat, 24 Sep 2005 23:52:30 +0000 (23:52 +0000)]
r10483: fixed some uninitialised variables warnings
Tim Potter [Sat, 24 Sep 2005 23:35:50 +0000 (23:35 +0000)]
r10482: Add files to proto_files enough to build libcli and librpc directories.
Tim Potter [Sat, 24 Sep 2005 22:58:18 +0000 (22:58 +0000)]
r10481: #define _GNU_SOURCE for comparison_fn_t test.
Jelmer Vernooij [Sat, 24 Sep 2005 22:37:54 +0000 (22:37 +0000)]
r10480: Fix build for people that don't have pidl installed
Volker Lendecke [Sat, 24 Sep 2005 17:49:33 +0000 (17:49 +0000)]
r10479: Apply some const
Jelmer Vernooij [Sat, 24 Sep 2005 16:23:41 +0000 (16:23 +0000)]
r10478: More work on proto headers; we now generate a couple of smaller ones
that are then included by include/proto.h
Simo Sorce [Sat, 24 Sep 2005 15:42:15 +0000 (15:42 +0000)]
r10477: expose transactions outside ldb and change the API once more
do not autostart transactions on ldb operations if a transaction is already in place
test transactions on winsdb
all my tests passes so far
tridge please confirm this is ok for you
Jelmer Vernooij [Sat, 24 Sep 2005 14:58:18 +0000 (14:58 +0000)]
r10476: Move some more types to libreplace. Fix missing strndup errors
for heimdal
Andrew Tridgell [Sat, 24 Sep 2005 13:10:19 +0000 (13:10 +0000)]
r10475: make sure we report failures in tdbtorture (ie. get the exit status right)
Volker Lendecke [Sat, 24 Sep 2005 08:24:55 +0000 (08:24 +0000)]
r10472: Check for strndup
Andrew Tridgell [Sat, 24 Sep 2005 07:30:20 +0000 (07:30 +0000)]
r10471: stratos doesn't have getpagesize(), so guess 8k on systems that don't
have it. Overestimating is harmless.
Andrew Tridgell [Sat, 24 Sep 2005 07:07:22 +0000 (07:07 +0000)]
r10470: solaris8 has a problem with tdbtorture with 3 processes. To see if
this is just a solaris issue this patch changes the default to 3, and
I'll see how many build farm boxes break
Andrew Tridgell [Sat, 24 Sep 2005 06:53:24 +0000 (06:53 +0000)]
r10469: use the older style of structure initialisation for tdb to make it
more portable. tdb is used in more than just Samba4, and I think the
portability/readability balance is a bit different
Andrew Tridgell [Sat, 24 Sep 2005 06:49:28 +0000 (06:49 +0000)]
r10468: - terminate tdbtorture quickly when an error is detected
- more workarounds for aix not handling malloc of size 0
Andrew Tridgell [Sat, 24 Sep 2005 05:02:02 +0000 (05:02 +0000)]
r10467: aix doesn't like zero length malloc :(
Andrew Tridgell [Sat, 24 Sep 2005 04:56:51 +0000 (04:56 +0000)]
r10466: work around missing pread/pwrite declaration on openbsd
Andrew Tridgell [Sat, 24 Sep 2005 03:43:02 +0000 (03:43 +0000)]
r10465: separate out a read_only db from a read-only traversal to ensure we
don't end up doing a mmap read only
Andrew Bartlett [Sat, 24 Sep 2005 03:31:01 +0000 (03:31 +0000)]
r10464: Use more consistant names.
Andrew Bartlett
Andrew Tridgell [Sat, 24 Sep 2005 03:29:50 +0000 (03:29 +0000)]
r10463: consider it an error if tdbtorture produces any log messages
Andrew Tridgell [Sat, 24 Sep 2005 03:28:10 +0000 (03:28 +0000)]
r10462: cope better with compilers that don't put the object file in the same directory
as the source
Andrew Tridgell [Sat, 24 Sep 2005 02:39:09 +0000 (02:39 +0000)]
r10461: fixed tdb build on systems without stdint.h
Andrew Tridgell [Sat, 24 Sep 2005 02:37:22 +0000 (02:37 +0000)]
r10460: fixed portability of transaction code to systems with integer
alignment constraints (like sparc)
Andrew Tridgell [Sat, 24 Sep 2005 02:31:43 +0000 (02:31 +0000)]
r10459: fixed some portability problems
Jelmer Vernooij [Fri, 23 Sep 2005 16:32:52 +0000 (16:32 +0000)]
r10456: More SCons fixes:
- Add framework for fallback generating code
- Move pread / pwrite replacement functions to libreplace
- Support pidl builds correctly
- Support asn1 builds correctly
- Move OS-specific checks to lib/replace/SConscript
Jelmer Vernooij [Fri, 23 Sep 2005 15:08:17 +0000 (15:08 +0000)]
r10453: Fix the build
Jelmer Vernooij [Fri, 23 Sep 2005 14:09:38 +0000 (14:09 +0000)]
r10452: Couple of small scons updates - ignore .sconsign files
Jelmer Vernooij [Fri, 23 Sep 2005 13:06:23 +0000 (13:06 +0000)]
r10450: Use exit code 1 in case of fatal errors
Stefan Metzmacher [Fri, 23 Sep 2005 07:53:58 +0000 (07:53 +0000)]
r10449: remove double copyright header
metze
Tim Potter [Fri, 23 Sep 2005 07:43:43 +0000 (07:43 +0000)]
r10448: Add librpc SConscript file to build.
Tim Potter [Fri, 23 Sep 2005 07:43:06 +0000 (07:43 +0000)]
r10447: Add gensec sources to proto.h
Tim Potter [Fri, 23 Sep 2005 07:42:17 +0000 (07:42 +0000)]
r10446: Add ndr_base and rpc_base sources to proto.h
Generate NDR mashalling code for all idl files (why isn't this automatically by the scanner when it finds librpc/gen_ndr/{misc,dcerpc,orpc}.h in includes.h?)
Tim Potter [Fri, 23 Sep 2005 07:40:15 +0000 (07:40 +0000)]
r10445: Add --header to pidl flags.
Tim Potter [Fri, 23 Sep 2005 07:36:09 +0000 (07:36 +0000)]
r10444: Add LIBBASIC files to proto.h
Tim Potter [Fri, 23 Sep 2005 06:39:34 +0000 (06:39 +0000)]
r10443: Fix emitter for NdrMarshaller tool.
Tim Potter [Fri, 23 Sep 2005 02:28:30 +0000 (02:28 +0000)]
r10442: Build librpc/libndr_misc.a from librpc/idl/misc.idl. Doesn't quite work
yet as scons somehow thinks it needs to put a directory called ndr_idl
in the command line.
Andrew Bartlett [Fri, 23 Sep 2005 02:05:26 +0000 (02:05 +0000)]
r10440: Start passing against Win2k3 SP1 again, with the NTLMv2 changes
described on the list. I probably need to write more specific NTLMv2
sucess and failure mode tests.
Andrew Bartlett
Jelmer Vernooij [Fri, 23 Sep 2005 00:38:22 +0000 (00:38 +0000)]
r10438: Move portability functions to lib/replace/; replace now simply ensures
that a given set of (working) POSIX functions are available (without
prefixes to their names, etc). See lib/replace/README for a list.
Functions that behave different from their POSIX specification
(such as sys_select, sys_read, etc) have kept the sys_ prefix.
Tim Potter [Thu, 22 Sep 2005 23:30:08 +0000 (23:30 +0000)]
r10436: Some thoughts on getting pidl working in scons.
Stefan Metzmacher [Thu, 22 Sep 2005 19:44:08 +0000 (19:44 +0000)]
r10435: fill in the reference to the generic wbsrv_call in the wbsrv_samba3_call,
so that async function can use it.
metze
Stefan Metzmacher [Thu, 22 Sep 2005 19:36:43 +0000 (19:36 +0000)]
r10434: add a short path to the event context that should be used for async replies
metze
Stefan Metzmacher [Thu, 22 Sep 2005 18:35:08 +0000 (18:35 +0000)]
r10426: - restructure the winbind server code a bit
- remove the echo test stuff
- abstract out the used protocol
- we have a seperate handler for the samba3 protocol now
- the backend can easy do async replies
by setting WBSRV_CALL_FLAGS_REPLY_ASYNC in wbsrv_call
and then call wbsrv_queue_reply() later
metze
Andrew Tridgell [Thu, 22 Sep 2005 13:20:28 +0000 (13:20 +0000)]
r10424: for caller convenience, automatically turn a tdb_traverse() into a
tdb_traverse_read() for read only databases
Andrew Tridgell [Thu, 22 Sep 2005 13:16:23 +0000 (13:16 +0000)]
r10423: minor changes to the ldb test suite to allow it to work correctly with
a real ldap backend (such as openldap)
Andrew Tridgell [Thu, 22 Sep 2005 13:14:12 +0000 (13:14 +0000)]
r10422: ldb_search() can now use tdb_traverse_read() to ensure it can run in
parallel with any transaction