ira/wip.git
16 years agor5998: I was wrong with the highwater mark...
Stefan Metzmacher [Wed, 23 Mar 2005 18:54:06 +0000 (18:54 +0000)]
r5998: I was wrong with the highwater mark...

I think I now understand how it works:-)

metze

16 years agor5992: Rename schannel.c -> schannel_sign.c. The rest of the schannel code
Andrew Bartlett [Wed, 23 Mar 2005 09:05:40 +0000 (09:05 +0000)]
r5992: Rename schannel.c -> schannel_sign.c.  The rest of the schannel code
(from librpc) will be moved into schannel.c soon.

Andrew Bartlett

16 years agor5989: Display authentication information (list of available auth protocols
Jelmer Vernooij [Wed, 23 Mar 2005 01:42:29 +0000 (01:42 +0000)]
r5989: Display authentication information (list of available auth protocols
+ principal names per endpoint) to gepdump. Still need to fix memory management
in the GTK+ utilities...

16 years agor5988: Fix the -P option (use machine account credentials) to use the Samba4
Andrew Bartlett [Wed, 23 Mar 2005 01:30:43 +0000 (01:30 +0000)]
r5988: Fix the -P option (use machine account credentials) to use the Samba4
secrets system, and not the old system from Samba3.

This allowed the code from auth_domain to be shared - we now only
lookup the secrets.ldb in lib/credentials.c.

In order to link the resultant binary, samdb_search() has been moved
from deep inside rpc_server into lib/gendb.c, along with the existing
gendb_search_v().  The vast majority of this patch is the simple
rename that followed,

(Depending on the whole SAMDB for just this function seemed pointless,
and brought in futher dependencies, such as smbencrypt.c).

Andrew Bartlett

16 years agor5987: Add credentials callback for gtk+. The gtk+ apps now no longer
Jelmer Vernooij [Wed, 23 Mar 2005 01:02:29 +0000 (01:02 +0000)]
r5987: Add credentials callback for gtk+. The gtk+ apps now no longer
ask for a password when kerberos is being used.

16 years agor5986: Fix the build. Metze, could you please verify that this fix is correct?
Jelmer Vernooij [Wed, 23 Mar 2005 01:00:23 +0000 (01:00 +0000)]
r5986: Fix the build. Metze, could you please verify that this fix is correct?

16 years agor5985: Actually adding auth_domain.c in -r 5983 would probably have been a
Andrew Bartlett [Wed, 23 Mar 2005 00:15:41 +0000 (00:15 +0000)]
r5985: Actually adding auth_domain.c in -r 5983 would probably have been a
good idea....

Andrew Bartlett

16 years agor5984: Add index and attributes to default ldif for secrets.ldb
Andrew Bartlett [Wed, 23 Mar 2005 00:07:21 +0000 (00:07 +0000)]
r5984: Add index and attributes to default ldif for secrets.ldb

Andrew Bartlett

16 years agor5983: Start support for being a domain member in Samba4.
Andrew Bartlett [Wed, 23 Mar 2005 00:05:44 +0000 (00:05 +0000)]
r5983: Start support for being a domain member in Samba4.

This adds the auth_domain module to the auth subsystem, and cleans up
some small details around the join process (ensuring all the right
info is in the DB).

Andrew Bartlett

16 years agor5980: Fix double free after unexpected disconnect.
Jelmer Vernooij [Tue, 22 Mar 2005 23:20:41 +0000 (23:20 +0000)]
r5980: Fix double free after unexpected disconnect.

16 years agor5977: Fix uninitialised memory bug in ndr_pull_ref_ptr(). This fixes the
Jelmer Vernooij [Tue, 22 Mar 2005 23:00:12 +0000 (23:00 +0000)]
r5977: Fix uninitialised memory bug in ndr_pull_ref_ptr(). This fixes the
Test_DoublePointer test failure.

16 years agor5976: SIDs can't have more then 5 subauths (caught by [validate] and
Jelmer Vernooij [Tue, 22 Mar 2005 22:11:50 +0000 (22:11 +0000)]
r5976: SIDs can't have more then 5 subauths (caught by [validate] and
range())

16 years agor5963: Fix parameter passing for gentest and locktest
Jelmer Vernooij [Tue, 22 Mar 2005 19:30:59 +0000 (19:30 +0000)]
r5963: Fix parameter passing for gentest and locktest

16 years agor5949: give things more meaning, and reuse structs where it is possible
Stefan Metzmacher [Tue, 22 Mar 2005 14:49:11 +0000 (14:49 +0000)]
r5949: give things more meaning, and reuse structs where it is possible
to make things more clear

metze

16 years agor5947: print out the password hashes when -d 100 is in use,
Stefan Metzmacher [Tue, 22 Mar 2005 14:45:43 +0000 (14:45 +0000)]
r5947: print out the password hashes when -d 100 is in use,
very usefull for creating a keytab file with

metze

16 years agor5942: A couple of small changes to fix things up with the new credentials
Andrew Bartlett [Tue, 22 Mar 2005 10:33:53 +0000 (10:33 +0000)]
r5942: A couple of small changes to fix things up with the new credentials
infrustructure.

Andrew Bartlett

16 years agor5941: Commit this patch much earlier than I would normally prefer, but metze needs...
Andrew Bartlett [Tue, 22 Mar 2005 08:00:45 +0000 (08:00 +0000)]
r5941: Commit this patch much earlier than I would normally prefer, but metze needs a working tree...

The main volume of this patch was what I started working on today:
 - Cleans up memory handling around DCE/RPC pipes, to have a parent talloc context.
 - Uses sepereate inner loops for some of the DCE/RPC tests

The other and more important part of this patch fixes issues
surrounding the new credentials framwork:

This makes the struct cli_credentials always a talloc() structure,
rather than on the stack.  Parts of the cli_credentials code already
assumed this.

There were other issues, particularly in the DCERPC over SMB handling,
as well as little things that had to be tidied up before test_w2k3.sh
would start to pass.

Andrew Bartlett

16 years agor5940: fix schannel against w2k, it skips the confounder in the signature (24 bytes...
Stefan Metzmacher [Tue, 22 Mar 2005 06:58:27 +0000 (06:58 +0000)]
r5940: fix schannel against w2k, it skips the confounder in the signature (24 bytes) for singed packets
but it accepts 32 bytes from the client.

(w2k3 accept it the otherway arround too)

metze

16 years agor5939: improve talloc_realloc() docs after feedback from lifeless
Andrew Tridgell [Tue, 22 Mar 2005 06:00:51 +0000 (06:00 +0000)]
r5939: improve talloc_realloc() docs after feedback from lifeless

16 years agor5938: - allow NULL string argument to talloc_vasprintf_append()
Andrew Tridgell [Tue, 22 Mar 2005 05:51:41 +0000 (05:51 +0000)]
r5938: - allow NULL string argument to talloc_vasprintf_append()

- default to using va_copy(), thus assuming a modern libc

16 years agor5937: - performance improvement to talloc_asprintf_append()
Andrew Tridgell [Tue, 22 Mar 2005 04:22:39 +0000 (04:22 +0000)]
r5937: - performance improvement to talloc_asprintf_append()

- allow standalone talloc to use gcc printf attributes

16 years agor5932: Use cli_credentials somewhat more in the Gtk+ code
Jelmer Vernooij [Tue, 22 Mar 2005 01:35:12 +0000 (01:35 +0000)]
r5932: Use cli_credentials somewhat more in the Gtk+ code
Support ncacn_spx in DCE/RPC bindings.

16 years agor5930: Fix initialisation of dcerpc_binding->authservice
Jelmer Vernooij [Tue, 22 Mar 2005 00:26:27 +0000 (00:26 +0000)]
r5930: Fix initialisation of dcerpc_binding->authservice

16 years agor5929: Use cli_credentials for the SMB functions as well.
Jelmer Vernooij [Mon, 21 Mar 2005 23:35:58 +0000 (23:35 +0000)]
r5929: Use cli_credentials for the SMB functions as well.
Fix a couple of bugs in the new cli_credentials code

16 years agor5928: Use cli_credentials in:
Jelmer Vernooij [Mon, 21 Mar 2005 21:22:07 +0000 (21:22 +0000)]
r5928: Use cli_credentials in:
 - gtk+ (returned by GtkHostBindingDialog as well now)
 - torture/
 - librpc/
 - lib/com/dcom/

16 years agor5924: Use cli_credentials in libnet/.
Jelmer Vernooij [Mon, 21 Mar 2005 18:42:32 +0000 (18:42 +0000)]
r5924: Use cli_credentials in libnet/.

16 years agor5917: First step in using the new cli_credentials structure. This patch
Jelmer Vernooij [Mon, 21 Mar 2005 02:08:38 +0000 (02:08 +0000)]
r5917: First step in using the new cli_credentials structure. This patch
puts support for it into popt_common, adds a few utility functions
(in lib/credentials.c) and the callback functions for the command-line
(lib/cmdline/credentials.c). Comments are welcome :-)

16 years agor5906: Fix the usage of the internal popt (make proto should ignore it)
Jelmer Vernooij [Sat, 19 Mar 2005 19:31:25 +0000 (19:31 +0000)]
r5906: Fix the usage of the internal popt (make proto should ignore it)
Updated included popt to 1.7.

16 years agor5903: While I can't test IPv6, metze asked me to commit a matching change
Andrew Bartlett [Sat, 19 Mar 2005 10:28:31 +0000 (10:28 +0000)]
r5903: While I can't test IPv6, metze asked me to commit a matching change
for unknown hosts that I just did for IPv4.

Andrew Bartlett

16 years agor5902: A rather large change...
Andrew Bartlett [Sat, 19 Mar 2005 08:34:43 +0000 (08:34 +0000)]
r5902: A rather large change...

I wanted to add a simple 'workstation' argument to the DCERPC
authenticated binding calls, but this patch kind of grew from there.

With SCHANNEL, the 'workstation' name (the netbios name of the client)
matters, as this is what ties the session between the NETLOGON ops and
the SCHANNEL bind.  This changes a lot of files, and these will again
be changed when jelmer does the credentials work.

I also correct some schannel IDL to distinguish between workstation
names and account names.  The distinction matters for domain trust
accounts.

Issues in handling this (issues with lifetime of talloc pointers)
caused me to change the 'creds_CredentialsState' and 'struct
dcerpc_binding' pointers to always be talloc()ed pointers.

In the schannel DB, we now store both the domain and computername, and
query on both.  This should ensure we fault correctly when the domain
is specified incorrectly in the SCHANNEL bind.

In the RPC-SCHANNEL test, I finally fixed a bug that vl pointed out,
where the comment claimed we re-used a connection, but in fact we made
a new connection.

This was achived by breaking apart some of the
dcerpc_secondary_connection() logic.

The addition of workstation handling was also propogated to NTLMSSP
and GENSEC, for completeness.

The RPC-SAMSYNC test has been cleaned up a little, using a loop over
usernames/passwords rather than manually expanded tests.  This will be
expanded further (the code in #if 0 in this patch) to use a newly
created user account for testing.

In making this test pass test_rpc.sh, I found a bug in the RPC-ECHO
server, caused by the removal of [ref] and the assoicated pointer from
the IDL.  This has been re-added, until the underlying pidl issues are
solved.

16 years agor5901: Add another option to the test script - the realm, which must match
Andrew Bartlett [Sat, 19 Mar 2005 08:18:24 +0000 (08:18 +0000)]
r5901: Add another option to the test script - the realm, which must match
the real ream, not just the short domain name.

Andrew Bartlett

16 years agor5900: Use flatname to specify the netbios domain name (matches what win2k3
Andrew Bartlett [Sat, 19 Mar 2005 08:11:49 +0000 (08:11 +0000)]
r5900: Use flatname to specify the netbios domain name (matches what win2k3
uses for trusted domain records) in the secrets join records.

Andrew Bartlett

16 years agor5899: Fix spelling.
Andrew Bartlett [Sat, 19 Mar 2005 06:49:03 +0000 (06:49 +0000)]
r5899: Fix spelling.

Andrew Bartlett

16 years agor5898: Handle errors in the 'sync' name and IP address handling code.
Andrew Bartlett [Sat, 19 Mar 2005 06:07:33 +0000 (06:07 +0000)]
r5898: Handle errors in the 'sync' name and IP address handling code.

Andrew Bartlett

16 years agor5895: Remove old auth_domain code - to be replaced with entirely new implementation.
Andrew Bartlett [Sat, 19 Mar 2005 03:15:23 +0000 (03:15 +0000)]
r5895: Remove old auth_domain code - to be replaced with entirely new implementation.

Andrew Bartlett

16 years agor5879: Rename SAMR_FIELD_WORKSTATION to SAMR_FIELD_WORKSTATIONS - it is a list.
Andrew Bartlett [Fri, 18 Mar 2005 04:25:10 +0000 (04:25 +0000)]
r5879: Rename SAMR_FIELD_WORKSTATION to SAMR_FIELD_WORKSTATIONS - it is a list.

Andrew Bartlett

16 years agor5878: Be clear which machine name (We have one worksation, and one BDC) we
Andrew Bartlett [Fri, 18 Mar 2005 04:09:52 +0000 (04:09 +0000)]
r5878: Be clear which machine name (We have one worksation, and one BDC) we
are doing logins with.

Andrew Bartlett

16 years agor5877: It is not an error to have a zero-length secret, after decryption.
Andrew Bartlett [Fri, 18 Mar 2005 03:17:30 +0000 (03:17 +0000)]
r5877: It is not an error to have a zero-length secret, after decryption.

Andrew Bartlett

16 years agor5876: Add a test account for the duration of the samsync - to ensure we have
Andrew Bartlett [Fri, 18 Mar 2005 03:16:53 +0000 (03:16 +0000)]
r5876: Add a test account for the duration of the samsync - to ensure we have
a good variety of things to test against.

Add code to testjoin to handle this just like test machine accounts

Soon I'll remove the 'must change password' flag, so we can do logins with it.

Andrew Bartlett

16 years agor5871: Remove file with unused function (that uses fstring)
Jelmer Vernooij [Fri, 18 Mar 2005 00:17:10 +0000 (00:17 +0000)]
r5871: Remove file with unused function (that uses fstring)
Remove fstring usage from version.c

16 years agor5867: winreg depends on initshutdown now (uses initshutdown_String)
Jelmer Vernooij [Thu, 17 Mar 2005 20:29:18 +0000 (20:29 +0000)]
r5867: winreg depends on initshutdown now (uses initshutdown_String)

16 years agor5866: Add InitShutdown IDL and torture test.
Jelmer Vernooij [Thu, 17 Mar 2005 20:28:01 +0000 (20:28 +0000)]
r5866: Add InitShutdown IDL and torture test.
Implement push side of NDR_LEN4|NDR_NOTERM strings (pull side was already present)

16 years agor5853: Move some of the functions not specific to the Samba NDR parser generator
Jelmer Vernooij [Thu, 17 Mar 2005 12:45:10 +0000 (12:45 +0000)]
r5853: Move some of the functions not specific to the Samba NDR parser generator
to a new ndr.pm.
Add function that can generate a "OrderTable" describing the order
in which the NDR data will be pushed/pulled.

16 years agor5852: Rename ndr.pm to ndr_parser.pm
Jelmer Vernooij [Thu, 17 Mar 2005 12:12:57 +0000 (12:12 +0000)]
r5852: Rename ndr.pm to ndr_parser.pm
I'm going to add a ndr.pm later on that'll generate a
tree with necessary information for the two NDR backends
(eparser, ndr_parser) containing alignment info, etc.

16 years agor5850: enable parsing of revision 4 security acl's
Stefan Metzmacher [Thu, 17 Mar 2005 04:24:35 +0000 (04:24 +0000)]
r5850: enable parsing of revision 4 security acl's

metze

16 years agor5839: add LDAP DirSync control idl
Stefan Metzmacher [Thu, 17 Mar 2005 00:33:56 +0000 (00:33 +0000)]
r5839: add LDAP DirSync control idl

metze

16 years agor5830: start to analyse the attribute values, depending on the attribute type
Stefan Metzmacher [Wed, 16 Mar 2005 15:47:19 +0000 (15:47 +0000)]
r5830: start to analyse the attribute values, depending on the attribute type

metze

16 years agor5828: add some idl for DsAddEntry()
Stefan Metzmacher [Wed, 16 Mar 2005 09:25:52 +0000 (09:25 +0000)]
r5828: add some idl for DsAddEntry()

metze

16 years agor5827: Make ndrdump accept a uuid as well as a pipe name to specify
Tim Potter [Wed, 16 Mar 2005 06:18:20 +0000 (06:18 +0000)]
r5827: Make ndrdump accept a uuid as well as a pipe name to specify
which rpc interface to use.

16 years agor5800: fix recursiv printing in ndr_print_DsGetNCChangesInfo1()
Stefan Metzmacher [Tue, 15 Mar 2005 16:05:43 +0000 (16:05 +0000)]
r5800: fix recursiv printing in ndr_print_DsGetNCChangesInfo1()

metze

16 years agor5799: more DsGetNCChanges updates, I'm starting to understand it...
Stefan Metzmacher [Tue, 15 Mar 2005 14:42:09 +0000 (14:42 +0000)]
r5799: more DsGetNCChanges updates, I'm starting to understand it...

also add a really simple torture test for DsGetNCChanges

metze

16 years agor5798: limit the size of an sid, 28 bytes complete is the biggest SID
Stefan Metzmacher [Tue, 15 Mar 2005 14:37:02 +0000 (14:37 +0000)]
r5798: limit the size of an sid, 28 bytes complete is the biggest SID
that can be handled.

tridge: do you think it would make sense to change the sub_auth[num_auths] to sub_auth[5],
        so we can copy the struct by sid1 = sid2;
comments please

metze

16 years agor5797: - add idl property [subcontext_size()]
Stefan Metzmacher [Tue, 15 Mar 2005 14:33:38 +0000 (14:33 +0000)]
r5797: - add idl property [subcontext_size()]

  this can be used like this
  [subcontext_size(28),subcontext(0)] dom_sid sid;

  this descripes a fixed 28 byte buffer which contains a dom_sid,
  and the rest of the buffer is padded with zero bytes if the dom_sid doesn't
  need all 28 byte in it's ndr encoding.

- only push and pull the subcontext when we are in the NDR_SCALARS section
  (tridge, jelmer: I hope this is correct for all cases...!?:-)

metze

16 years agor5796: add ndr_* function for int8 and int16
Stefan Metzmacher [Tue, 15 Mar 2005 14:25:59 +0000 (14:25 +0000)]
r5796: add ndr_* function for int8 and int16

metze

16 years agor5785: Fix subcontext pushes
Jelmer Vernooij [Mon, 14 Mar 2005 01:11:13 +0000 (01:11 +0000)]
r5785: Fix subcontext pushes

16 years agor5783: Test renaming of accounts in the RPC-SAMR test, and add support into
Andrew Bartlett [Sun, 13 Mar 2005 06:43:34 +0000 (06:43 +0000)]
r5783: Test renaming of accounts in the RPC-SAMR test, and add support into
the SAMR server.

Andrew Bartlett

16 years agor5782: Use standard input for reading packet data if filename not specified.
Tim Potter [Sun, 13 Mar 2005 03:16:07 +0000 (03:16 +0000)]
r5782: Use standard input for reading packet data if filename not specified.

16 years agor5781: Fix build.
Tim Potter [Sun, 13 Mar 2005 02:48:16 +0000 (02:48 +0000)]
r5781: Fix build.

16 years agor5780: Remove unused arguments.
Tim Potter [Sun, 13 Mar 2005 02:20:51 +0000 (02:20 +0000)]
r5780: Remove unused arguments.

16 years agor5779: Remove signal and timeout gubbage from tdb.
Tim Potter [Sun, 13 Mar 2005 01:40:45 +0000 (01:40 +0000)]
r5779: Remove signal and timeout gubbage from tdb.

16 years agor5778: Remove tdb_set_lock_alarm() function from wrappers at tridge's request.
Tim Potter [Sun, 13 Mar 2005 01:01:24 +0000 (01:01 +0000)]
r5778: Remove tdb_set_lock_alarm() function from wrappers at tridge's request.

16 years agor5777: At some stage SIG_ATOMIC_T was renamed to sig_atomic_t. This should
Tim Potter [Sat, 12 Mar 2005 22:24:46 +0000 (22:24 +0000)]
r5777: At some stage SIG_ATOMIC_T was renamed to sig_atomic_t.  This should
fix abartlet's compile problem.

16 years agor5775: Remove some unused functions (unions are no longer as special as they used...
Jelmer Vernooij [Sat, 12 Mar 2005 17:12:52 +0000 (17:12 +0000)]
r5775: Remove some unused functions (unions are no longer as special as they used to be)
Add oxid mapping table support in DCOM

16 years agor5769: Fix unused variable warning.
Tim Potter [Sat, 12 Mar 2005 09:55:08 +0000 (09:55 +0000)]
r5769: Fix unused variable warning.

16 years agor5766: get some more DsGetNCChanges stuff parsing correct
Stefan Metzmacher [Sat, 12 Mar 2005 09:25:39 +0000 (09:25 +0000)]
r5766: get some more DsGetNCChanges stuff parsing correct

metze

16 years agor5765: remove unused var
Stefan Metzmacher [Sat, 12 Mar 2005 08:51:37 +0000 (08:51 +0000)]
r5765: remove unused var

metze

16 years agor5764: add support for embedded ref pointers
Stefan Metzmacher [Sat, 12 Mar 2005 08:29:54 +0000 (08:29 +0000)]
r5764: add support for embedded ref pointers

metze

16 years agor5742: - add torture test and idl for DsReplicaUpdateRefs()
Stefan Metzmacher [Fri, 11 Mar 2005 12:15:50 +0000 (12:15 +0000)]
r5742: - add torture test and idl for DsReplicaUpdateRefs()
  (the torture test currently only tests if the idl is correct)

- add start for idl for DsGetNCChanges()
  (if someone didn't noticed the current ethereal trunk code can
   successful decrypt DCERPC and LDAP gsskrb5 encrypted blobs,
   when you provide a keytab and have compiled against heimdal :-)

- add a view bitmaps and enum's for better debugging

metze

16 years agor5737: add some error codes
Stefan Metzmacher [Fri, 11 Mar 2005 10:33:01 +0000 (10:33 +0000)]
r5737: add some error codes

metze

16 years agor5736: fix to avoid endless recursion in ndr_size_*() calculation
Stefan Metzmacher [Fri, 11 Mar 2005 10:09:16 +0000 (10:09 +0000)]
r5736: fix to avoid endless recursion in ndr_size_*() calculation

metze

16 years agor5730: More fun with delayed write semantics with multiple file handles
Jeremy Allison [Thu, 10 Mar 2005 21:42:20 +0000 (21:42 +0000)]
r5730: More fun with delayed write semantics with multiple file handles
open on the same file on the same connection.
Jeremy.

16 years agor5716: Add new Unix extensions defines for infolevels
Steve French [Wed, 9 Mar 2005 21:56:00 +0000 (21:56 +0000)]
r5716: Add new Unix extensions defines for infolevels

16 years agor5709: Leave out the pull/push call with NDR_BUFFERS for not just
Jelmer Vernooij [Wed, 9 Mar 2005 15:39:55 +0000 (15:39 +0000)]
r5709: Leave out the pull/push call with NDR_BUFFERS for not just
scalars but also for structs/unions containing only scalars.

16 years agor5697: Give an error message when a ref pointer is NULL (requested by tpot)
Jelmer Vernooij [Tue, 8 Mar 2005 22:24:47 +0000 (22:24 +0000)]
r5697: Give an error message when a ref pointer is NULL (requested by tpot)

16 years agor5682: fix the build
Stefan Metzmacher [Mon, 7 Mar 2005 12:45:19 +0000 (12:45 +0000)]
r5682: fix the build

metze

16 years agor5681: fix the build and compiler wanings
Stefan Metzmacher [Mon, 7 Mar 2005 12:27:03 +0000 (12:27 +0000)]
r5681: fix the build and compiler wanings

metze

16 years agor5680: Don't crash if none of the predefined keys is available (reported by Alexander)
Jelmer Vernooij [Mon, 7 Mar 2005 12:02:48 +0000 (12:02 +0000)]
r5680: Don't crash if none of the predefined keys is available (reported by Alexander)

16 years agor5679: Complete basic test of rpc_composite_userinfo function (which
Rafal Szczesniak [Mon, 7 Mar 2005 05:12:33 +0000 (05:12 +0000)]
r5679: Complete basic test of rpc_composite_userinfo function (which
btw proves the function is completely broken but that's what
tests are for)

rafal

16 years agor5678: Torture tests require also libnet in order to run NET-* tests.
Rafal Szczesniak [Mon, 7 Mar 2005 05:10:44 +0000 (05:10 +0000)]
r5678: Torture tests require also libnet in order to run NET-* tests.

rafal

16 years agor5677: Split structure definitions from implementation.
Rafal Szczesniak [Mon, 7 Mar 2005 05:09:37 +0000 (05:09 +0000)]
r5677: Split structure definitions from implementation.

rafal

16 years agor5676: Fix some alignment issues and IRemoteActivation
Jelmer Vernooij [Mon, 7 Mar 2005 02:14:38 +0000 (02:14 +0000)]
r5676: Fix some alignment issues and IRemoteActivation

16 years agor5675: - More DCOM bug fixes
Jelmer Vernooij [Sun, 6 Mar 2005 22:37:31 +0000 (22:37 +0000)]
r5675: - More DCOM bug fixes
- Keep COM and DCOM more seperated

16 years agor5674: - Re-enable DCOM support.
Jelmer Vernooij [Sun, 6 Mar 2005 21:43:54 +0000 (21:43 +0000)]
r5674: - Re-enable DCOM support.
- Always put IID in vtables (useful for asserts)
- Add table to keep track of DCOM proxy classes
- Bunch of smaller bug fixes

16 years agor5672: Use switch_type() and the token storage mechanism for unions:
Jelmer Vernooij [Sun, 6 Mar 2005 17:02:14 +0000 (17:02 +0000)]
r5672: Use switch_type() and the token storage mechanism for unions:
 - Makes union handling less special
 - Allows unions in arrays, etc
 - Compatible with midl
 - Pidl will warn about switch_type() and the type of the switch_is() variable being different

16 years agor5671: Bunch of fixes related to arrays and pointers to arrays.
Jelmer Vernooij [Sun, 6 Mar 2005 15:57:08 +0000 (15:57 +0000)]
r5671: Bunch of fixes related to arrays and pointers to arrays.

16 years agor5670: simplify and clarify ldb_modules.c code
Simo Sorce [Sun, 6 Mar 2005 15:33:40 +0000 (15:33 +0000)]
r5670: simplify and clarify ldb_modules.c code
rectify the test schema
correct a glitch in schema module

16 years agor5669: Couple of minor clearifications, simplifications.
Jelmer Vernooij [Sun, 6 Mar 2005 12:33:44 +0000 (12:33 +0000)]
r5669: Couple of minor clearifications, simplifications.

16 years agor5668: Add tests to RPC-SAMLOGON to test for user@REALM style logins. These
Andrew Bartlett [Sun, 6 Mar 2005 08:24:34 +0000 (08:24 +0000)]
r5668: Add tests to RPC-SAMLOGON to test for user@REALM style logins.  These
need a NULL domain (or a "" domain, except this breaks NTLMv2, and I
need to look into it a bit more).

Add support to the Samba4 server for these logins.  This will need
extension when we handle trusted domains as a DC, as it is a principal
name, not just another format for the username.

Andrew Bartlett

16 years agor5667: Move schannel state into libcli/auth (as it belongs with schannel,
Andrew Bartlett [Sun, 6 Mar 2005 06:37:38 +0000 (06:37 +0000)]
r5667: Move schannel state into libcli/auth (as it belongs with schannel,
which will move in with the rest of GENSEC shortly).

Add the RID as another element in the schannel state.

Andrew Bartlett

16 years agor5666: winxp will use a NTTIME of -1 to mean "don't change" in setfileinfo
Andrew Tridgell [Sat, 5 Mar 2005 22:50:13 +0000 (22:50 +0000)]
r5666: winxp will use a NTTIME of -1 to mean "don't change" in setfileinfo
basic_info. Add null_nttime() as the equivalent of the existing
null_time() call for cheecking for valid NTTIME values

16 years agor5665: the data within el2->values can still be used at this point, so don't free
Andrew Tridgell [Sat, 5 Mar 2005 21:59:48 +0000 (21:59 +0000)]
r5665: the data within el2->values can still be used at this point, so don't free

16 years agor5664: simo, please look into this. It is possible for the number of elements
Andrew Tridgell [Sat, 5 Mar 2005 21:59:02 +0000 (21:59 +0000)]
r5664: simo, please look into this. It is possible for the number of elements
in msg[0] to be 0, in which case we crash. This is a workaround.

Also, if you could please split up this function into readable
pieces. It's a bit of a mess at the moment.

16 years agor5663: Revert some changes I didn't mean to commit...
Jelmer Vernooij [Sat, 5 Mar 2005 19:26:45 +0000 (19:26 +0000)]
r5663: Revert some changes I didn't mean to commit...

16 years agor5662: eventlog uses dom_sid so it depends on security
Jelmer Vernooij [Sat, 5 Mar 2005 19:20:12 +0000 (19:20 +0000)]
r5662: eventlog uses dom_sid so it depends on security

16 years agor5661: Be a little stricter on syntax regarding arrays. A pointer to an
Jelmer Vernooij [Sat, 5 Mar 2005 18:34:18 +0000 (18:34 +0000)]
r5661: Be a little stricter on syntax regarding arrays. A pointer to an
array can now only be :

 type *name[];

rather then :

 type *name;

which was supported in the past. Warnings will be given when the first
syntax is used. Reasons for this change in behaviour include improved
readability and the fact that the second format makes dealing with multiple
levels of pointers harder.

16 years agor5660: Reduce number of generated "{}"'s
Jelmer Vernooij [Sat, 5 Mar 2005 16:50:02 +0000 (16:50 +0000)]
r5660: Reduce number of generated "{}"'s
No longer complain about pointers to arrays - we support those.

16 years agor5652: Add libnet tests section to torture test.
Rafal Szczesniak [Fri, 4 Mar 2005 00:25:16 +0000 (00:25 +0000)]
r5652: Add libnet tests section to torture test.

rafal

16 years agor5651: A bit more code in userinfo test and hook up test function to
Rafal Szczesniak [Fri, 4 Mar 2005 00:24:21 +0000 (00:24 +0000)]
r5651: A bit more code in userinfo test and hook up test function to
main torture binary.

rafal

16 years agor5646: state->loadfile might be NULL after allocation so this is really
Alexander Bokovoy [Thu, 3 Mar 2005 09:26:08 +0000 (09:26 +0000)]
r5646: state->loadfile might be NULL after allocation so this is really
should be NT_STATUS_HAVE_NO_MEMORY(state->loadfile) instead of
NT_STATUS_NOT_OK_RETURN(status).

16 years agor5611: Add EnumPrinterDataEx-test (for now just for PrinterDriverData, we need
Günther Deschner [Wed, 2 Mar 2005 00:15:06 +0000 (00:15 +0000)]
r5611: Add EnumPrinterDataEx-test (for now just for PrinterDriverData, we need
to enumerate (recursively!) over all keys with EnumPrinterKey later on.

Guenther