ira/wip.git
13 years agor16131: fix RedHat 9 spec files
Gerald Carter [Fri, 9 Jun 2006 22:06:01 +0000 (22:06 +0000)]
r16131: fix RedHat 9 spec files
(This used to be commit 1ffb0f06adeb1658fa1f5eed747841c6ff1e5665)

13 years agor16130: fix a shell portability issue for Solaris sh in installman.sh
Gerald Carter [Fri, 9 Jun 2006 21:57:42 +0000 (21:57 +0000)]
r16130: fix a shell portability issue for Solaris sh in installman.sh
(This used to be commit 9be3e07884e7f8121bc64766e7d036495cd94a15)

13 years agor16126: Janitor for Volker..... This looks correct to me.
Jeremy Allison [Fri, 9 Jun 2006 17:19:59 +0000 (17:19 +0000)]
r16126: Janitor for Volker..... This looks correct to me.
Jeremy.

Fix a parsing error that became apparent in 'make test': If we have no
group
rids in the info3 we set the array buffer pointer to "1" but fail to
actually
ship the array.

Volker.
(This used to be commit ee1b9207d3119c2b3e7c1c4e59250dbd323eae6a)

13 years agor16122: As we use 'inetOrgPerson' as structural objectclass for new accounts for
Günther Deschner [Fri, 9 Jun 2006 13:04:47 +0000 (13:04 +0000)]
r16122: As we use 'inetOrgPerson' as structural objectclass for new accounts for
eDir, we already add 'sn' as required attribute on LDAP add
operations.

When we modify an entry, we need to request 'sn' as well in our
attribute lists, so that we don't try to add it a second time.

Guenther
(This used to be commit e018ea3d1d70e68e6802843f88b3245dc09244c9)

13 years agor16121: Fix a eDir related memory leak.
Günther Deschner [Fri, 9 Jun 2006 12:55:07 +0000 (12:55 +0000)]
r16121: Fix a eDir related memory leak.

Guenther
(This used to be commit 322f1664df553d95fcdfc24f19bd7f34ce9b834b)

13 years agor16117: Make winbindd work again in security=ads.
Günther Deschner [Fri, 9 Jun 2006 11:02:52 +0000 (11:02 +0000)]
r16117: Make winbindd work again in security=ads.

We still used the old HOST/* UPN to get e.g. users, now we need
samaccountname$@REA.LM.

Guenther
(This used to be commit f6516a799aec2db819f79b9a1e641637422a9b4c)

13 years agor16116: Hoist the slow CLOCK_REALTIME message inside the branch so we never
James Peach [Fri, 9 Jun 2006 10:50:45 +0000 (10:50 +0000)]
r16116: Hoist the slow CLOCK_REALTIME message inside the branch so we never
confuse an uninitialised __profile_clock with CLOCK_REALTIME. Flip the
condition argument to SMB_WARN around so that it's correct (though
completely non-intuitive).
(This used to be commit 60b5f9618b77afb397f96c1ff406c8cd3a4648f3)

13 years agor16115: Make "net ads changetrustpw" work again.
Günther Deschner [Fri, 9 Jun 2006 10:50:28 +0000 (10:50 +0000)]
r16115: Make "net ads changetrustpw" work again.

(adapt to the new UPN/SPN scheme).

Guenther
(This used to be commit 8fc70d0df0c93c29b49f924bac9ff5d9857cfd9d)

13 years agor16114: Make winbindd's group enumeration (set|get|endgrent) work again (when
Günther Deschner [Fri, 9 Jun 2006 10:33:29 +0000 (10:33 +0000)]
r16114: Make winbindd's group enumeration (set|get|endgrent) work again (when
enabled).

Do not bail out when a group just has 0 members.

Jeremy, please check, this has been removed with r13915.

Guenther
(This used to be commit 3a738a855d335e44e167351e6396bf3fe81a03af)

13 years agor16112: Added Volker's fix for extra odd byte at the end
Jeremy Allison [Fri, 9 Jun 2006 01:36:58 +0000 (01:36 +0000)]
r16112: Added Volker's fix for extra odd byte at the end
of UNIX_BASIC infolevel. Checked client implementations
(cifsfs and libsmb) and they ignore it. Thanks Volker.
Jeremy.
(This used to be commit 6cd1cb5c9deba2aa0d7a3251c8142834cbaba6d1)

13 years agor16111: Patch from Björn JACKE <samba@j3e.de>.
James Peach [Fri, 9 Jun 2006 01:02:54 +0000 (01:02 +0000)]
r16111: Patch from Björn JACKE <samba@j3e.de>.

This fixes a problem where the clock definition for clock_gettime() is
present at compile time, but is not available on the running system. In
this case, we fall back to less-preferred clocks until we find one that
we can use.
(This used to be commit fc6ed6a1aa2225ccde04c4ecaf0777dc0de4f1cb)

13 years agor16103: Properly rebuild timelimit on systems with executable extensions.
Paul Green [Thu, 8 Jun 2006 18:41:15 +0000 (18:41 +0000)]
r16103: Properly rebuild timelimit on systems with executable extensions.

Paul
(This used to be commit f5be3e12e2dc42a2092b71c9e4cab6f5d0ef8860)

13 years agor16080: Re-add accidentially excluded in-forest domain trusts (fixes bug #3823).
Günther Deschner [Wed, 7 Jun 2006 13:06:49 +0000 (13:06 +0000)]
r16080: Re-add accidentially excluded in-forest domain trusts (fixes bug #3823).

Guenther
(This used to be commit 8759a00fedfe5d8d789c8b707c924d8116da1102)

13 years agor16079: How embarassing. :-( Broke the build.
Gerald Carter [Wed, 7 Jun 2006 12:35:16 +0000 (12:35 +0000)]
r16079: How embarassing. :-(  Broke the build.

I will *always* compile before checking in...
I will *always* compile before checking in...
I will *always* compile before checking in...
I will *always* compile before checking in...
I will *always* compile before checking in...
(This used to be commit f3bd1c8c7de93917787e7341f759699c8399f006)

13 years agor16078: Patch from Aleksey Fedoseev <fedoseev@ru.ibm.com> to correctly
Gerald Carter [Wed, 7 Jun 2006 12:01:54 +0000 (12:01 +0000)]
r16078: Patch from Aleksey Fedoseev <fedoseev@ru.ibm.com> to correctly
set the type of 'acl compatibility from string to integer
after the changes in r14207.
(This used to be commit b01f289f1a788fe8f62a213b67a73a9bfee01bb1)

13 years agor16076: Fix for machine password timeout overflow from Shlomi Yaakobovich
Jeremy Allison [Wed, 7 Jun 2006 04:45:50 +0000 (04:45 +0000)]
r16076: Fix for machine password timeout overflow from Shlomi Yaakobovich
<Shlomi@exanet.com>.
Jeremy.
(This used to be commit 5cd234a1fff1e9d025eea6600649e56c997eafc2)

13 years agor16065: Re-add a strlower_m(account) in samr_create_user that was dropped for no
Volker Lendecke [Tue, 6 Jun 2006 20:34:26 +0000 (20:34 +0000)]
r16065: Re-add a strlower_m(account) in samr_create_user that was dropped for no
reason but to increase fidelity with W2k3. Tom Bork has raised valid concerns
that Unix scripts might rely on the account names being lower-case, so keep
that. We might later decide to only lower-case the unix name passed to
'add [user|group] script' but keep the passdb entry upper-case. But there are
enough user-visible changes in 3_0 already so that we should push this off to
a later date.

Tom, waiting for more bug reports from you ;-))

Thanks for insisting!

Volker
(This used to be commit bc78cca290559c5ca7623b9f6d9933e32668b9c4)

13 years agor16064: Bug fix for another one Tom Bork has reported:
Volker Lendecke [Tue, 6 Jun 2006 20:18:11 +0000 (20:18 +0000)]
r16064: Bug fix for another one Tom Bork has reported:

'valid users = +unixgroup' failed with smbpasswd if 'unixgroup' has a
(non-algorithmic) group mapping.

Thanks a lot!

People out there listening, please test current code, this release is

   **BIG**

:-)

Volker
(This used to be commit 8f9ba5f96c9b506623ef97b7ed3d84f39d914a3c)

13 years agor16060: This is one of the more dirty patches I've put in lately. Parse enough of
Volker Lendecke [Tue, 6 Jun 2006 14:18:12 +0000 (14:18 +0000)]
r16060: This is one of the more dirty patches I've put in lately. Parse enough of
SetUserInfo level 25 to survive the join method XP uses if the user did not
exist before. For good taste this contains way too much cut&paste, but for a
real fix there is just not enough time.

Up to 3.0.22 we completely ignored that a full level 21 is being sent together
with level 25, but we got away with that because on creation we did not set
the "disabled" flag on the workstation account. Now we correctly follow W2k3
in this regard, and we end up with a disabled workstation after join.

Man, I hate rpc_parse/. The correct fix would be to import PIDL generated samr
parsing, but this is would probably be a bit too much for .23...

Thanks to Tom Bork for finding this one.

Volker
(This used to be commit 5a37aba10551456042266443cc0a92f28f8c3d0d)

13 years agor16057: Coalesce the DMAPI configure tests into a single macro. Add
James Peach [Tue, 6 Jun 2006 07:43:17 +0000 (07:43 +0000)]
r16057: Coalesce the DMAPI configure tests into a single macro. Add
a more specific probe to try and eliminate old, incompatible
DMAPI implementations provided by IRIX 6.4 and AIX 4.3.
(This used to be commit aafd4db457ce8a60c628d54a3ace3b97c8885dca)

13 years agor16054: Janitor for Volker ? Volker - some reason you didn't
Jeremy Allison [Tue, 6 Jun 2006 00:34:26 +0000 (00:34 +0000)]
r16054: Janitor for Volker ? Volker - some reason you didn't
fix this in 3.0 ?
Jeremy.

We had no way to return NT_STATUS_OK from the netlogon serverpwset,
although
we successfully set the machine password...

One thing the samba3 join test found.

Volker
(This used to be commit e5b7acc9b5cb6e8cf3d03c9d392fad06e0d282d9)

13 years agor16047: Remove unnecessary line, as this value is set in either branch of the
Jim McDonough [Mon, 5 Jun 2006 16:59:10 +0000 (16:59 +0000)]
r16047: Remove unnecessary line, as this value is set in either branch of the
'if' below.  Spotted by Aleksey Fedoseev.
(This used to be commit 00eb42791abf3e34dc42a2067f07b8549ea31abb)

13 years agor16046: Hmmm. I don't fully understand popt, but this is necessary for smbclient to
Volker Lendecke [Mon, 5 Jun 2006 13:00:24 +0000 (13:00 +0000)]
r16046: Hmmm. I don't fully understand popt, but this is necessary for smbclient to
show a prompt again :-)

James, could you check that this is ok? I think this was part of your Revision
15848.

Thanks,

Volker
(This used to be commit 91534146b4718206f0e2a7f456154a0497277dfd)

13 years agor16025: make sure the samba4's smbtorture can be executed
Stefan Metzmacher [Sat, 3 Jun 2006 09:51:59 +0000 (09:51 +0000)]
r16025: make sure the samba4's smbtorture can be executed

metze
(This used to be commit bed220afb816138a3e806c5d20e812667f6994ae)

13 years agor16019: This should not be a level zero message - it's harmless
Jeremy Allison [Sat, 3 Jun 2006 00:36:00 +0000 (00:36 +0000)]
r16019: This should not be a level zero message - it's harmless
and can happen though misconfiguration.
Jeremy.
(This used to be commit 4b9cf399a691ba4a7392caca558d0e98b4d19104)

13 years agor16017: Add Samba4 replacement for timegm to work
Jeremy Allison [Fri, 2 Jun 2006 23:58:56 +0000 (23:58 +0000)]
r16017: Add Samba4 replacement for timegm to work
on Solaris.
Jeremy.
(This used to be commit b65c6468651282879a39fa8029c190f2d1a91a28)

13 years agor16016: Add debug to be symetrical with reading from
Jeremy Allison [Fri, 2 Jun 2006 23:58:02 +0000 (23:58 +0000)]
r16016: Add debug to be symetrical with reading from
cache.
Jeremy.
(This used to be commit da26565a2e85dc36b283f6b81378a706f3ae5f26)

13 years agor16014: Correctly set the group RID in init_sam_from_buffer.
Volker Lendecke [Fri, 2 Jun 2006 21:53:09 +0000 (21:53 +0000)]
r16014: Correctly set the group RID in init_sam_from_buffer.

BIG THANKS to Tom Bork for reporting that Bug!

Volker
(This used to be commit 40339fdcced67d62e449ba6f19329d89c808e139)

13 years agor15997: Fix bug in OS/2 Warp - it doesn't set the ff_last
Jeremy Allison [Fri, 2 Jun 2006 00:52:11 +0000 (00:52 +0000)]
r15997: Fix bug in OS/2 Warp - it doesn't set the ff_last
offset correctly when doing info level 1 directory
scans. Thanks to Guenter Kukkukk <Guenter.Kukkukk@kukkukk.com>
for reporting this problem and testing the fix.
Jeremy.
(This used to be commit 65d4dfbd6045a4e3f9eaf520c70ef29ff7ddee82)

13 years agor15995: Make smbldap obey config tests. Patch from "John E. Malmberg"
Jeremy Allison [Thu, 1 Jun 2006 19:25:21 +0000 (19:25 +0000)]
r15995: Make smbldap obey config tests. Patch from "John E. Malmberg"
<wb8tyw@qsl.net>.
Jeremy.
(This used to be commit 93dd476a9108adde42de422e1886bbaee75b952b)

13 years agor15989: try to make timelimit.c more portable
Stefan Metzmacher [Thu, 1 Jun 2006 06:11:31 +0000 (06:11 +0000)]
r15989: try to make timelimit.c more portable

Patch from Bjoern Jacke (Thanks!)

metze
(This used to be commit b9d1440f46f443851607300e31b123d6afa81c06)

13 years agor15985: Adding "own-domain" switch to wbinfo which is handy from time to time.
Günther Deschner [Wed, 31 May 2006 17:15:33 +0000 (17:15 +0000)]
r15985: Adding "own-domain" switch to wbinfo which is handy from time to time.

Guenther
(This used to be commit 3c9416c2bedeec7f075e94d45d08f37ae6dd41d1)

13 years agor15984: Correctly handle the case when there is no configuration file for
Günther Deschner [Wed, 31 May 2006 15:45:19 +0000 (15:45 +0000)]
r15984: Correctly handle the case when there is no configuration file for
pam_winbind.

Guenther
(This used to be commit 29758ea1c4e1b9b57d27765d539306058299fcd1)

13 years agor15983: Honour the krb5 principal name change (of the new ads join code) in the
Günther Deschner [Wed, 31 May 2006 15:41:54 +0000 (15:41 +0000)]
r15983: Honour the krb5 principal name change (of the new ads join code) in the
kerberized winbind pam_auth.

Guenther
(This used to be commit 216125fe132fa6b886b99139e38988725beb88f0)

13 years agor15982: Fix confusing order of DEBUG statements in winbindds pam_auth.
Günther Deschner [Wed, 31 May 2006 15:39:12 +0000 (15:39 +0000)]
r15982: Fix confusing order of DEBUG statements in winbindds pam_auth.

Guenther
(This used to be commit 3f5a2e49c108bfe8f8b875af9e69d5ad3b0567ee)

13 years agor15980: Correctly destroy talloc_ctx when the LDAP posix attribute query has
Günther Deschner [Wed, 31 May 2006 10:32:12 +0000 (10:32 +0000)]
r15980: Correctly destroy talloc_ctx when the LDAP posix attribute query has
failed. Noticed by Bob Gautier.

Guenther
(This used to be commit 7327f94546a90df25c688dcafd42e0993133057a)

13 years agor15977: Fillup the password_policy method in winbindd for winbindd_passdb. This
Günther Deschner [Wed, 31 May 2006 10:09:31 +0000 (10:09 +0000)]
r15977: Fillup the password_policy method in winbindd for winbindd_passdb. This
should make pam_winbind work again on a Samba PDC (and fix Bug #3800).

Guenther
(This used to be commit 4addabd054a2627133d3fff71234db18cf2c822c)

13 years agor15976: Set our internal domains to "online" by default in winbindd.
Günther Deschner [Wed, 31 May 2006 09:25:44 +0000 (09:25 +0000)]
r15976: Set our internal domains to "online" by default in winbindd.

Guenther
(This used to be commit 2678582c6cc7fb100cb3bfd867816878461ae7b4)

13 years agor15975: Only call the printer publishing calls if 'security = ads'
Gerald Carter [Wed, 31 May 2006 01:31:01 +0000 (01:31 +0000)]
r15975: Only call the printer publishing calls if 'security = ads'
(prevent a segv)
(This used to be commit a2ef525d9e3b4f050cb4e02fad67808d3e916373)

13 years agor15971: Obey the manpage description and make changesecretpw accept a password via...
Simo Sorce [Tue, 30 May 2006 22:10:09 +0000 (22:10 +0000)]
r15971: Obey the manpage description and make changesecretpw accept a password via stdin
(This used to be commit 60d4aabc3205aa80f8d49c3c2db95927c61a81a5)

13 years agor15967: enable RAW-OPLOCK in samba3 make test
Stefan Metzmacher [Tue, 30 May 2006 20:11:19 +0000 (20:11 +0000)]
r15967: enable RAW-OPLOCK in samba3 make test

metze
(This used to be commit 6d35268639a7796583f5cc7bee12cf884a207af2)

13 years agor15965: Pam modules install fix, thanks to Ryan Novosielski for spotting this one
Simo Sorce [Tue, 30 May 2006 19:58:16 +0000 (19:58 +0000)]
r15965: Pam modules install fix, thanks to Ryan Novosielski for spotting this one
(This used to be commit 8f5b198acaca5ce14ab1098d86f34df9a67619d4)

13 years agor15960: Don't double-free (ensure we always free file_fsp in
Jeremy Allison [Tue, 30 May 2006 18:38:10 +0000 (18:38 +0000)]
r15960: Don't double-free (ensure we always free file_fsp in
the same place on error - I missed the extra free in
delay_for_oplocks).
Jeremy.
(This used to be commit 146a2648156563b41642d42c46aff401e435a42b)

13 years agor15959: Ooops. Use the right file_free call...
Jeremy Allison [Tue, 30 May 2006 18:26:49 +0000 (18:26 +0000)]
r15959: Ooops. Use the right file_free call...
Jeremy.
(This used to be commit 1117a327bd4fb425d25ea63df782095701fb83bb)

13 years agor15958: Make us pass RAW-OPLOCK with kernel oplocks off.
Jeremy Allison [Tue, 30 May 2006 18:17:37 +0000 (18:17 +0000)]
r15958: Make us pass RAW-OPLOCK with kernel oplocks off.
This allows a requestor to set FORCE_OPLOCK_BREAK_TO_NONE
to ensure we don't break to level 2. Fixed a couple
of resource leaks in error paths in open_file_ntcreatex.
Jeremy.
(This used to be commit c7c9adcce7f13d01445f31b07fb28a76f0a1d6df)

13 years agor15951: oplock_request must be an int, not a BOOL. We
Jeremy Allison [Tue, 30 May 2006 05:24:21 +0000 (05:24 +0000)]
r15951: oplock_request must be an int, not a BOOL. We
were getting away with mixing types. Not cool.
Jeremy.
(This used to be commit ad3bc112a21afd3a04449ec3f604e017d753c224)

13 years agor15949: Patch for bug #3308 to stop us returning duplicate
Jeremy Allison [Tue, 30 May 2006 05:08:15 +0000 (05:08 +0000)]
r15949: Patch for bug #3308 to stop us returning duplicate
mid replies on path based set-eof trans2 calls.
Needs modification for HEAD (as in head open_file_ntcreateX
properly returns NTSTATUS - I'll fix this tomorrow my
time). Secondly it still fails the Samba4 RAW-OPLOCK
smbtorture because of an interesting case. Our oplock
code always returns "break to level 2" if it can.
In this case (path-based set-eof or set-allocation size
on an exclusive oplocked file) W2K3 always sends a
break-to-none. We send the break to none (from level2)
after we've done the write for eof or allocation size.
I need to work out some way of telling our break code
to always break to none (might need to extend the message
field).
Jeremy.
(This used to be commit ad9895c654f400e242adcd4099f7cd004521ee92)

13 years agor15943: Update misleading DEBUG statements and comments that refer to
James Peach [Mon, 29 May 2006 23:54:53 +0000 (23:54 +0000)]
r15943: Update misleading DEBUG statements and comments that refer to
open_file_shared.
(This used to be commit 784126edff942d1c2f79f7c9feb59980426f331e)

13 years agor15936: Fix memory leak on error exit path. Trying to
Jeremy Allison [Mon, 29 May 2006 21:44:19 +0000 (21:44 +0000)]
r15936: Fix memory leak on error exit path. Trying to
track down #3308.
Jeremy.
(This used to be commit e39c11c91153c1ed547bc635d02769a1cdf3ecc0)

13 years agor15928: When we search for clock_gettime, we might find it in librt, but librt
James Peach [Mon, 29 May 2006 02:25:36 +0000 (02:25 +0000)]
r15928: When we search for clock_gettime, we might find it in librt, but librt
might pull in libpthread. This is quite bad, firstly because it can
cause oplock signals on Linux to go wonky, and secondly because merely
linking with pthreads can cause performance degradations due to implicit
locking requirements.

The solution is to only search for clock_gettime if --with-profiling-data
was specified. If we do end up searching for it, then we test whether
linking with librt pulled in libpthread, and we only allow the definition
for clock_gettime to succeed if libpthread was NOT linked in.

Problem reported by Thomas Bork and diagnosed by Volker Lendecke.
(This used to be commit 5712644fcc00939b9101b2e3143674d2cbdf1e12)

13 years agor15924: Ok, the Solaris fix broke AIX. Next try :-)
Volker Lendecke [Sun, 28 May 2006 16:06:56 +0000 (16:06 +0000)]
r15924: Ok, the Solaris fix broke AIX. Next try :-)

Volker
(This used to be commit 53bd3fa30ad1f8bada4a49e4c79ea7ff70188174)

13 years agor15918: Attempt to bring Solaris into make test. The Solaris shell does not like the
Volker Lendecke [Sun, 28 May 2006 09:59:16 +0000 (09:59 +0000)]
r15918: Attempt to bring Solaris into make test. The Solaris shell does not like the
isolated !.

Metze, can you check this?

Thanks,

Volker
(This used to be commit c9542c64ce2a21671b8f4e30780e8a6e07ab6632)

13 years agor15917: Activate RPC-AUTHCONTEXT in "make test"
Volker Lendecke [Sun, 28 May 2006 08:30:35 +0000 (08:30 +0000)]
r15917: Activate RPC-AUTHCONTEXT in "make test"
(This used to be commit 17f701ab4f46502ca7a957d3e79b005a0acebdc1)

13 years agor15911: Make us survive rpc-authcontext committed next
Volker Lendecke [Sat, 27 May 2006 21:38:54 +0000 (21:38 +0000)]
r15911: Make us survive rpc-authcontext committed next
(This used to be commit c24bfdce625782637b5f4d11a5117ef795ddfc2f)

13 years agor15910: vfs_full_audit does not need current_user
Volker Lendecke [Sat, 27 May 2006 16:58:36 +0000 (16:58 +0000)]
r15910: vfs_full_audit does not need current_user
(This used to be commit 09f3c7a86f931371dc6f886edccc15d3f5437d34)

13 years agor15909: Implement recycle:subdir_mode
Volker Lendecke [Sat, 27 May 2006 16:55:30 +0000 (16:55 +0000)]
r15909: Implement recycle:subdir_mode
(This used to be commit 4dd8694a250e4d064a790fe8f422c965ab533880)

13 years agor15906: smbpasswd help text for -W option (patch from Aruna Prabakar <aruna.prabakar...
Gerald Carter [Fri, 26 May 2006 23:06:56 +0000 (23:06 +0000)]
r15906: smbpasswd help text for -W option (patch from Aruna Prabakar <aruna.prabakar@hp.com>
(This used to be commit 0a81af4fef5b794fea257d9e3e11b16c3a4de12f)

13 years agor15904: This does two things:
Volker Lendecke [Fri, 26 May 2006 12:28:55 +0000 (12:28 +0000)]
r15904: This does two things:

Fix more potential segfaults when something on our way to a DC connection
fails.

We can not continue if dcip_to_name() fails. With

192.168.234.100 nt4pdc
192.168.234.100 windows#1c
192.168.234.100 windows#1b

in the lmhosts file when nt4pdc is rebooted, we do find the DC's IP address,
we can connect to TCP 139 while it is booting but anything else fails. So we
fall back to put the IP address into domain->dcname. When the DC is fully up
later on we try to do the auth2 against \\192.168.234.100 which gives
INVALID_COMPUTER_NAME. And we never get out of this loop again.

Fix this.

Jerry, maybe you can take a look.

Thanks,

Volker
(This used to be commit b1244e79068af9e287252b2dfbb8d612e717674a)

13 years agor15895: Ensure all new rid allocation goes through
Jeremy Allison [Fri, 26 May 2006 00:13:06 +0000 (00:13 +0000)]
r15895: Ensure all new rid allocation goes through
the same function (deals with races).
Jeremy.
(This used to be commit 4962548dfe8ec2854e209217066556f339d3186e)

13 years agor15890: Use correct enum type (bug #3722) from Jason Mader <jason@ncac.gwu.edu>.
Jeremy Allison [Thu, 25 May 2006 15:47:36 +0000 (15:47 +0000)]
r15890: Use correct enum type (bug #3722) from Jason Mader <jason@ncac.gwu.edu>.
Jeremy.
(This used to be commit a8eb1186a12b44178a28b013373f6f8a4356d9ec)

13 years agor15888: Fix bug #3804 from jason@ncac.gwu.edu
Jeremy Allison [Thu, 25 May 2006 15:44:41 +0000 (15:44 +0000)]
r15888: Fix bug #3804 from jason@ncac.gwu.edu
Invalid comparisons.
Jeremy.
(This used to be commit 9890a31c5f4a8911b0f56eee67cfbcc46f15ee43)

13 years agor15887: Ensure we use sys_write so we're not interrupted.
Jeremy Allison [Thu, 25 May 2006 15:43:12 +0000 (15:43 +0000)]
r15887: Ensure we use sys_write so we're not interrupted.
Jeremy.
(This used to be commit c66620770d2154543a6ec99d369771b339df5463)

13 years agor15884: Fix bug #3803 from jason@ncac.gwu.edu - bad comparison.
Jeremy Allison [Thu, 25 May 2006 15:40:40 +0000 (15:40 +0000)]
r15884: Fix bug #3803 from jason@ncac.gwu.edu - bad comparison.
Jeremy.
(This used to be commit 1d9dbe3b662ea757224ca4d8c991ae73bb92ae8d)

13 years agor15873: Use short signal names to placate ksh trap. bash seems to accept either
Jim McDonough [Wed, 24 May 2006 20:06:06 +0000 (20:06 +0000)]
r15873: Use short signal names to placate ksh trap.  bash seems to accept either
and even uses them on the manpage.  this should now enable make test on
AIX.
(This used to be commit de836d2077e53c07a50c51d1580f8e93700eb5bd)

13 years agor15872: Take one step toward getting this working on AIX. the [-z ] stuff
Jim McDonough [Wed, 24 May 2006 19:07:21 +0000 (19:07 +0000)]
r15872: Take one step toward getting this working on AIX.  the [-z ] stuff
doesn't work there.
(This used to be commit 0ee78d96d9bd039084a270172316d80df4c00248)

13 years agor15864: build fixes for IRIX 6.4 in the build farm; only enable the C++ reserved...
Gerald Carter [Wed, 24 May 2006 14:38:11 +0000 (14:38 +0000)]
r15864: build fixes for IRIX 6.4 in the build farm; only enable the C++ reserved word check when we selecte --enable-developer
(This used to be commit ece5fe3e78da460fb1c0eca3185da7f80c5cfdc4)

13 years agor15863: trying to fix a compile issue on HP-UX 11.x caused by conflicts of int16...
Gerald Carter [Wed, 24 May 2006 14:26:34 +0000 (14:26 +0000)]
r15863: trying to fix a compile issue on HP-UX 11.x caused by conflicts of int16 and int32 definitions in internal and system headers
(This used to be commit 094443ed6c0fdd06f459211dbca50feed58463ef)

13 years agor15848: Introduce commandline options to set the remainder of the parameters in
James Peach [Wed, 24 May 2006 04:02:14 +0000 (04:02 +0000)]
r15848: Introduce commandline options to set the remainder of the parameters in
dynconfig.c. This is mainly useful for test harness scripts, hence the
lack of short options.
(This used to be commit bf3b71c84595608d71e0f15a6158adacb295518e)

13 years agor15845: Ok. This was a tough one. If for some reason the tconX fails towards a domain
Volker Lendecke [Tue, 23 May 2006 22:35:08 +0000 (22:35 +0000)]
r15845: Ok. This was a tough one. If for some reason the tconX fails towards a domain
controller the next time we connect this child ran into a segfault because it
tried to reference a half-baked connection.

Volker
(This used to be commit c8a8204c744cf7aa1a1a6992a3433d99b6bb73a1)

13 years agor15842: patch from volker to instruct winbindd to find a trusted DC on its own when...
Gerald Carter [Tue, 23 May 2006 18:51:03 +0000 (18:51 +0000)]
r15842: patch from volker to instruct winbindd to find a trusted DC on its own when runing on a Samba DC (since we don't implement the getdcname() call that well
(This used to be commit 39f7ff75a7a21b85b54cba954f1c5552e562be5c)

13 years agor15838: Back-port tridge's talloc fixes (r15824, r15828) from Samba4.
Jeremy Allison [Tue, 23 May 2006 15:57:26 +0000 (15:57 +0000)]
r15838: Back-port tridge's talloc fixes (r15824, r15828) from Samba4.
Jeremy.
(This used to be commit f6c110ddb8cfaa1a57dea52818e7611134c2dcfe)

13 years agor15822: Add suggestion made by Ralf Haferkamp.
Lars Müller [Mon, 22 May 2006 20:35:55 +0000 (20:35 +0000)]
r15822: Add suggestion made by Ralf Haferkamp.
(This used to be commit 7c375fd540fa54ac8ae71c42ed07e01c593044b3)

13 years agor15817: Remove some unused code
Volker Lendecke [Mon, 22 May 2006 18:52:54 +0000 (18:52 +0000)]
r15817: Remove some unused code
(This used to be commit 72f103708d17aa86e09fa7a02699f969f7ee9190)

13 years agor15755: Fix Coverity bug # 294. Apparently password can be NULL, but cli_session_setup
Volker Lendecke [Sat, 20 May 2006 18:30:09 +0000 (18:30 +0000)]
r15755: Fix Coverity bug # 294. Apparently password can be NULL, but cli_session_setup
derefences it.

Volker
(This used to be commit b013b6908d22cfd38fcc56a9cb2ca675d75996d1)

13 years agor15735: also export PREFIX as testit from test_functions.sh needs it
Stefan Metzmacher [Sat, 20 May 2006 08:55:19 +0000 (08:55 +0000)]
r15735: also export PREFIX as testit from test_functions.sh needs it

metze
(This used to be commit 8709bba24a2359aff14920b1dab6b51f21401149)

13 years agor15732: samba3 now passes BASE-DENY1
Stefan Metzmacher [Sat, 20 May 2006 07:42:54 +0000 (07:42 +0000)]
r15732: samba3 now passes BASE-DENY1

metze
(This used to be commit e5a8553a57e0b1463a421ba69631f234184404b0)

13 years agor15729: Second part of Aleksey Fedoseev <fedoseev@ru.ibm.com> patch.
Jeremy Allison [Sat, 20 May 2006 02:09:35 +0000 (02:09 +0000)]
r15729: Second part of Aleksey Fedoseev <fedoseev@ru.ibm.com> patch.
Jeremy.
(This used to be commit 5137006233a0c2de94ce3da7340ff0032ef7426d)

13 years agor15727: Fix msgtest - noticed by Aleksey Fedoseev <fedoseev@ru.ibm.com>.
Jeremy Allison [Sat, 20 May 2006 02:04:19 +0000 (02:04 +0000)]
r15727: Fix msgtest - noticed by Aleksey Fedoseev <fedoseev@ru.ibm.com>.
Jeremy.
(This used to be commit 9f5c95c41872deac5e029b69daa5004d5c0fb4cc)

13 years agor15717: BUG 3711: patch from Leonid Kabanov <lkabanov@mail.ru> for some solaris...
Gerald Carter [Fri, 19 May 2006 14:44:50 +0000 (14:44 +0000)]
r15717: BUG 3711: patch from    Leonid Kabanov <lkabanov@mail.ru> for some solaris sheel scripting bugs with 'make test'
(This used to be commit 94e0bc50135646039a2ebcf5c922a82cc0c4a2e7)

13 years agor15712: BUG 3435: patch from volker to fix 'msdfs root = yes' in [homes]
Gerald Carter [Fri, 19 May 2006 14:00:12 +0000 (14:00 +0000)]
r15712: BUG 3435: patch from volker to fix 'msdfs root = yes' in [homes]
(This used to be commit 466478f07e6233b89f442660ad42ef7ee870ad48)

13 years agor15705: Fix bug number 3788. Thanks to Jeff Wright.
Volker Lendecke [Fri, 19 May 2006 07:01:53 +0000 (07:01 +0000)]
r15705: Fix bug number 3788. Thanks to Jeff Wright.

Volker
(This used to be commit e4a2cb4b9143394a54ae1de91e59722c11a0b2e4)

13 years agor15704: Prefer LDAP error codes in ads_search_retry_sid().
Günther Deschner [Thu, 18 May 2006 22:34:16 +0000 (22:34 +0000)]
r15704: Prefer LDAP error codes in ads_search_retry_sid().

Guenther
(This used to be commit 6cfc65ea20793a72ff1666759bd4e8e446247071)

13 years agor15703: Fix d_printf call.
Günther Deschner [Thu, 18 May 2006 22:13:03 +0000 (22:13 +0000)]
r15703: Fix d_printf call.

Guenther
(This used to be commit 741602e03ad2404d4e38e55b9e5fd20b85fd205d)

13 years agor15701: change 'net ads leave' to disable the machine account in the domain (since...
Gerald Carter [Thu, 18 May 2006 20:12:45 +0000 (20:12 +0000)]
r15701: change 'net ads leave' to disable the machine account in the domain (since removal implies greater permissions that Windows clients require)
(This used to be commit ad1f947625612ef16adb69fc2cfeffc68a9a2e02)

13 years agor15700: Make nmbd udp sockets non-blocking to prevent problem
Jeremy Allison [Thu, 18 May 2006 19:49:44 +0000 (19:49 +0000)]
r15700: Make nmbd udp sockets non-blocking to prevent problem
with select returning true but no data being available.
Fix for bug #3779.
Jeremy.
(This used to be commit e5787cf75b2e7d50f551f34f28d280c27b0aa134)

13 years agor15698: An attempt to make the winbind lookup_usergroups() call in security=ads
Günther Deschner [Thu, 18 May 2006 19:34:25 +0000 (19:34 +0000)]
r15698: An attempt to make the winbind lookup_usergroups() call in security=ads
more scalable:

The most efficient way is to use the "tokenGroups" attribute which gives
the nested group membership. As this attribute can not always be
retrieved when binding with the machine account (the only garanteed way
to get the tokenGroups I could find is when the machine account is a
member of the "Pre Win2k Access" builtin group).

Our current fallback when "tokenGroups" failed is looking for all groups
where the userdn was in the "member" attribute. This behaves not very
well in very large AD domains.

The patch first tries the "memberOf" attribute on the user's dn in that
case and directly retrieves the group's sids by using the LDAP Extended
DN control from the user's object.

The way to pass down the control to the ldap search call is rather
painfull and probably will be rearranged later on.

Successfully tested on win2k sp0, win2k sp4, wink3 sp1 and win2k3 r2.

Guenther
(This used to be commit 7d766b5505e4099ef7dd4e88bb000ebe38d71bd0)

13 years agor15697: I take no comments as no objections :)
Günther Deschner [Thu, 18 May 2006 16:08:28 +0000 (16:08 +0000)]
r15697: I take no comments as no objections :)

Expand the "winbind nss info" to also take "rfc2307" to support the
plain posix attributes LDAP schema from win2k3-r2.

This work is based on patches from Howard Wilkinson and Bob Gautier
(and closes bug #3345).

Guenther
(This used to be commit 52423e01dc209ba5abde808a446287714ed11567)

13 years agor15696: Free LDAP search result.
Günther Deschner [Thu, 18 May 2006 15:08:09 +0000 (15:08 +0000)]
r15696: Free LDAP search result.

Guenther
(This used to be commit ec26c355b3ef1d3d809c4fbe911ce6fcef5db955)

13 years agor15691: - samba3 also pass RAW-CHKPATH, RAW-SEEK, RAW-SFILEINFO-BUG
Stefan Metzmacher [Thu, 18 May 2006 12:24:09 +0000 (12:24 +0000)]
r15691: - samba3 also pass RAW-CHKPATH, RAW-SEEK, RAW-SFILEINFO-BUG
- samba3 pass the limited version of RAW-LOCK
  (the lock cancel and error checking tests are skipped)

metze
(This used to be commit b79ceece9550c0fe9f59ae59bad6709351e93906)

13 years agor15686: - start testing the RAW-* tests in make test
Stefan Metzmacher [Thu, 18 May 2006 09:39:31 +0000 (09:39 +0000)]
r15686: - start testing the RAW-* tests in make test
- RAW-READ now passes samba3
- skip the rest for now

metze
(This used to be commit fc343051afe560b1ddb8715dbc574e2ae645571e)

13 years agor15682: fix option passing
Stefan Metzmacher [Thu, 18 May 2006 09:21:23 +0000 (09:21 +0000)]
r15682: fix option passing

metze
(This used to be commit cd2fa1da6cdcaaa8a4d92a38163d54ecc0f9d756)

13 years agor15681: fix segv in 'kinit && net ads join'
Gerald Carter [Thu, 18 May 2006 04:33:43 +0000 (04:33 +0000)]
r15681: fix segv in 'kinit && net ads join'
(This used to be commit d77768cb237461b06119ee19f822b120623d77dd)

13 years agor15680: use the user creds when calling net_set_machine_spn() rather than the machin...
Gerald Carter [Thu, 18 May 2006 04:13:07 +0000 (04:13 +0000)]
r15680: use the user creds when calling net_set_machine_spn()  rather than the machine creds (just like WinXP)
(This used to be commit ae2bf464c47eb52ff24400d1cc362e74e77fbac5)

13 years agor15679: fix some popt warnings
Gerald Carter [Thu, 18 May 2006 04:10:52 +0000 (04:10 +0000)]
r15679: fix some popt warnings
(This used to be commit 126ab8ccec69240e8e526190f4712bd274237985)

13 years agor15676: Fix meaningless debug statement from uninitialized variable.
Jeremy Allison [Thu, 18 May 2006 02:05:53 +0000 (02:05 +0000)]
r15676: Fix meaningless debug statement from uninitialized variable.
Spotted by "John E. Malmberg" <wb8tyw@qsl.net>.
Jeremy.
(This used to be commit ff3fe39b837e0d0de2edaa284c2dd7d1c8161c46)

13 years agor15675: Man pages say never look at the fd_set after a select
Jeremy Allison [Thu, 18 May 2006 01:45:18 +0000 (01:45 +0000)]
r15675: Man pages say never look at the fd_set after a select
if it returned -1 (treat as undefined). Ensure we obey
this.
Jeremy.
(This used to be commit 256ae3a16bcafe70cc1a00496681c709380e4fc3)

13 years agor15672: Fix for bug #3783. winbindd_cm.c calls open_any_socket_out()
Jeremy Allison [Thu, 18 May 2006 01:30:00 +0000 (01:30 +0000)]
r15672: Fix for bug #3783. winbindd_cm.c calls open_any_socket_out()
to make connections to ports 445 or 139 on the DC it's trying
to contact. It calls sys_select() on the non-blocking sockets,
not sys_select_intr(). This is a mistake (I believe) as it allows
a signal to early terminate the connection attempts - whereas
sys_select_intr() will ignore signals until we get back to
the main processing loop where they'll be handled correctly.
This change means winbindd_cm will not early terminate if it
gets a message whilst trying to connect to DC's.
Gunther, Volker and Jerry please review (but I think this
is correct).
Jeremy.
(This used to be commit 24aaa486771f797d35ea6b0711c12cd3e663dd8c)

13 years agor15670: Fix valgrind-spotted issue in BASE-DELETE test.
Jeremy Allison [Thu, 18 May 2006 00:22:07 +0000 (00:22 +0000)]
r15670: Fix valgrind-spotted issue in BASE-DELETE test.
We were forgetting to increment after copying
the primary group gid.
Jeremy
(This used to be commit 31d16c434e1ee94691f013ed0b31d9f26baeb2cb)

13 years agor15668: DOS or FCB opens share one share mode entry from different
Jeremy Allison [Wed, 17 May 2006 23:15:53 +0000 (23:15 +0000)]
r15668: DOS or FCB opens share one share mode entry from different
fsp pointers. Ensure we cope with this to pass Samba4
DENY tests (we used to pass these, there must have been
a regression with newer code). We now pass them.
Jeremy
(This used to be commit fd6fa1d4eaf61783df74ee2da50d331477f06998)

13 years agor15660: Without this when using smbcquotas I get
Volker Lendecke [Wed, 17 May 2006 15:01:57 +0000 (15:01 +0000)]
r15660: Without this when using smbcquotas I get

close fd=-1 fnum=4321 (numopen=1)
close_file: Could not get share mode lock for file $Extend/$Quota:$Q:$INDEX_ALLOCATION
unix_error_packet: error string = Das Argument ist ungültig
error packet at smbd/reply.c(3325) cmd=4 (SMBclose) NT_STATUS_INVALID_HANDLE

so a fake file needs special close handling I think. Jeremy, can you check
this?

Thanks,

Volker
(This used to be commit f66b9701b5c6bb6302fa11889adab6902cbaf2e3)