sfrench/samba-autobuild/.git
9 years agocompression: added a simple lzxpress test
Andrew Tridgell [Thu, 4 Nov 2010 22:01:28 +0000 (09:01 +1100)]
compression: added a simple lzxpress test

9 years agocompression: fixed an uninitialised data but in lzxpress
Andrew Tridgell [Thu, 4 Nov 2010 22:00:57 +0000 (09:00 +1100)]
compression: fixed an uninitialised data but in lzxpress

9 years agos3: Remove the use of cli_send_trans from cli_get_fs_full_size_info
Volker Lendecke [Thu, 4 Nov 2010 18:23:06 +0000 (19:23 +0100)]
s3: Remove the use of cli_send_trans from cli_get_fs_full_size_info

Autobuild-User: Volker Lendecke <vlendec@samba.org>
Autobuild-Date: Thu Nov  4 22:08:53 UTC 2010 on sn-devel-104

9 years agos4:ldap.py/passwords.py - just use objectclass "user" for users
Matthias Dieter Wallnöfer [Mon, 1 Nov 2010 16:23:34 +0000 (17:23 +0100)]
s4:ldap.py/passwords.py - just use objectclass "user" for users

Autobuild-User: Matthias Dieter Wallnöfer <mdw@samba.org>
Autobuild-Date: Thu Nov  4 21:09:00 UTC 2010 on sn-devel-104

9 years agos3: Remove an unused prototype
Volker Lendecke [Thu, 4 Nov 2010 17:49:55 +0000 (18:49 +0100)]
s3: Remove an unused prototype

Autobuild-User: Volker Lendecke <vlendec@samba.org>
Autobuild-Date: Thu Nov  4 17:44:09 UTC 2010 on sn-devel-104

9 years agos4:dsdb/objectclass_attrs: not all objects have delete protected attributes as must...
Stefan Metzmacher [Thu, 4 Nov 2010 11:37:29 +0000 (12:37 +0100)]
s4:dsdb/objectclass_attrs: not all objects have delete protected attributes as must contain

Before we got the following error, while starting samba after a
'samba-tool vampire':

Failed to store repsFrom - objectclass_attrs: delete protected attribute
'objectSid' on entry 'DC=ForestDnsZones,DC=alpha,DC=sz,DC=salzgitter-ag,DC=lab'
missing!

metze

Autobuild-User: Stefan Metzmacher <metze@samba.org>
Autobuild-Date: Thu Nov  4 17:01:59 UTC 2010 on sn-devel-104

9 years agos4:dsdb/samdb: optimize samldb_prim_group_change()
Stefan Metzmacher [Wed, 3 Nov 2010 11:33:54 +0000 (12:33 +0100)]
s4:dsdb/samdb: optimize samldb_prim_group_change()

We should only do searches when we have to.

metze

9 years agos4:dsdb/common: fix memory leak in samdb_ntds_settings_dn()
Stefan Metzmacher [Thu, 4 Nov 2010 13:29:12 +0000 (14:29 +0100)]
s4:dsdb/common: fix memory leak in samdb_ntds_settings_dn()

fetch and set should use the same name!

metze

9 years agos4:dsdb/kcc: fix memory leak in kcctpl_copy_output_edges()
Stefan Metzmacher [Thu, 4 Nov 2010 12:32:06 +0000 (13:32 +0100)]
s4:dsdb/kcc: fix memory leak in kcctpl_copy_output_edges()

metze

9 years agos3-waf: add check for httpConnect and httpConnectEncrypt.
Günther Deschner [Wed, 3 Nov 2010 16:16:06 +0000 (17:16 +0100)]
s3-waf: add check for httpConnect and httpConnectEncrypt.

Guenther

Autobuild-User: Günther Deschner <gd@samba.org>
Autobuild-Date: Thu Nov  4 14:10:40 UTC 2010 on sn-devel-104

9 years agos4-ldb_module.h: use LDB error code for ldb_error() macro rather than LDB_DEBUG_FATAL
Kamen Mazdrashki [Thu, 4 Nov 2010 01:27:10 +0000 (03:27 +0200)]
s4-ldb_module.h: use LDB error code for ldb_error() macro rather than LDB_DEBUG_FATAL

We end up calling ldb_error_at() which expects an LDB error,
but LDB_DEBUG_FATAL is not such code. It is actually equal to LDB_SUCCESS.

Thus the effect is that we report a *fatal* error, but return
LDB_SUCCESS in many places.

Autobuild-User: Kamen Mazdrashki <kamenim@samba.org>
Autobuild-Date: Thu Nov  4 12:57:33 UTC 2010 on sn-devel-104

9 years agos4-devel: Verify prefixMap is saved correctly in LDB on disk
Kamen Mazdrashki [Wed, 3 Nov 2010 21:43:21 +0000 (23:43 +0200)]
s4-devel: Verify prefixMap is saved correctly in LDB on disk

9 years agourgent_replication: forbid empty subunit output.
Jelmer Vernooij [Thu, 4 Nov 2010 11:29:42 +0000 (12:29 +0100)]
urgent_replication: forbid empty subunit output.

Autobuild-User: Jelmer Vernooij <jelmer@samba.org>
Autobuild-Date: Thu Nov  4 12:13:47 UTC 2010 on sn-devel-104

9 years agoselftest: Allow some tests to output empty subunit data.
Jelmer Vernooij [Thu, 4 Nov 2010 01:07:57 +0000 (02:07 +0100)]
selftest: Allow some tests to output empty subunit data.

9 years agos4-dsdb Fix urgent_replication test not to set an invalid userAccountControl
Andrew Bartlett [Thu, 4 Nov 2010 02:16:11 +0000 (13:16 +1100)]
s4-dsdb Fix urgent_replication test not to set an invalid userAccountControl

A userAccountControl value of 0 or 1 is invalid, you must include one of
the account types.

Andrew Bartlett

9 years agourgent_replication: Fix syntax error.
Jelmer Vernooij [Thu, 4 Nov 2010 00:36:57 +0000 (01:36 +0100)]
urgent_replication: Fix syntax error.

9 years agos4-dsdb: use LDB_FLAG_MOD_REPLACE for isDeleted
Andrew Tridgell [Thu, 4 Nov 2010 09:34:37 +0000 (20:34 +1100)]
s4-dsdb: use LDB_FLAG_MOD_REPLACE for isDeleted

isDeleted could have been set to FALSE previously

Autobuild-User: Andrew Tridgell <tridge@samba.org>
Autobuild-Date: Thu Nov  4 10:18:10 UTC 2010 on sn-devel-104

9 years agos4-dsdb: use LDB_FLAG_INTERNAL_DISABLE_SINGLE_VALUE_CHECK in dsdb
Andrew Tridgell [Thu, 4 Nov 2010 09:33:31 +0000 (20:33 +1100)]
s4-dsdb: use LDB_FLAG_INTERNAL_DISABLE_SINGLE_VALUE_CHECK in dsdb

when we are creating linked attributes with multiple values (some
deleted), use LDB_FLAG_INTERNAL_DISABLE_SINGLE_VALUE_CHECK to disable
that checking.

9 years agos4-ldb: implement LDB_FLAG_INTERNAL_DISABLE_SINGLE_VALUE_CHECK
Andrew Tridgell [Thu, 4 Nov 2010 09:27:43 +0000 (20:27 +1100)]
s4-ldb: implement LDB_FLAG_INTERNAL_DISABLE_SINGLE_VALUE_CHECK

this disables the single value checking for one attribute. It is much
more specific than a general RELAX control, and also more efficient. I
think we should try to have more precise overrides like this, rather
than using RELAX as a general purpose override

9 years agos4-ldb: don't load modules from the same directory twice
Andrew Tridgell [Thu, 4 Nov 2010 09:13:17 +0000 (20:13 +1100)]
s4-ldb: don't load modules from the same directory twice

this prevents samba dying if you do a 'make install' while it is
running. Otherwise the make install changes the inode numbers of the
modules in the modules directory, causing them to reload, which causes
multiple modules of the same name to try and load

9 years agos4-ldb: honor controls on search in ldbedit
Andrew Tridgell [Thu, 4 Nov 2010 09:05:59 +0000 (20:05 +1100)]
s4-ldb: honor controls on search in ldbedit

9 years agoMove the checking of single valued attributes back into the tdb backend.
Andrew Tridgell [Thu, 4 Nov 2010 09:02:16 +0000 (20:02 +1100)]
Move the checking of single valued attributes back into the tdb backend.

The backend is the only place that can do this properly. It makes no
sense to do it anywhere else. As a result of it moving out of the
backend we ended up with some bugs causing multiple values in single
valued attributes (eg. isDeleted), which can really damage the
inregrity of the database.

For the override of single valued values needed for deleted linked
attributes we should use attribute flags.

This reverts commit 1949864417f3d10fb8996df7db259649eb777271.

9 years agos4-test: fixed quoting in deletetest.py
Andrew Tridgell [Thu, 4 Nov 2010 02:23:47 +0000 (13:23 +1100)]
s4-test: fixed quoting in deletetest.py

9 years agos4-auth: unconditionally set previous_ev
Andrew Tridgell [Thu, 4 Nov 2010 01:52:08 +0000 (12:52 +1100)]
s4-auth: unconditionally set previous_ev

we need the caller to know when the previous_ev was NULL

Pair-Programmed-With: Andrew Bartlett <abartlet@samba.org>

9 years agodebug: fixed default debug settings
Andrew Tridgell [Wed, 3 Nov 2010 23:25:30 +0000 (10:25 +1100)]
debug: fixed default debug settings

this fixes -d in our command line tools (eg. samba-tool)

Pair-Programmed-With: Andrew Bartlett <abartlet@samba.org>

Autobuild-User: Andrew Tridgell <tridge@samba.org>
Autobuild-Date: Thu Nov  4 01:48:15 UTC 2010 on sn-devel-104

9 years agoSpeed test with user objects
Zahari Zahariev [Fri, 29 Oct 2010 16:34:13 +0000 (19:34 +0300)]
Speed test with user objects

This is a modest speed test that aims to show performance difference
between Samba4 vs MS Active Directory. It should be used with Samba-style
credentials and no arguments like every other python unittest.

Signed-off-by: Anatoliy Atanasov <anatoliy.atanasov@postpath.com>
Autobuild-User: Anatoliy Atanasov <anatoliy@samba.org>
Autobuild-Date: Thu Nov  4 00:11:20 UTC 2010 on sn-devel-104

9 years agolibrpc/ndr: use new strlen_m_ext_term() in ndr_charset_length(): fix bug #7594
Michael Adam [Sun, 31 Oct 2010 00:04:25 +0000 (02:04 +0200)]
librpc/ndr: use new strlen_m_ext_term() in ndr_charset_length(): fix bug #7594

This fixes the calculation of needed space for destination unicode charset.

Autobuild-User: Michael Adam <obnox@samba.org>
Autobuild-Date: Wed Nov  3 23:28:07 UTC 2010 on sn-devel-104

9 years agolib/util/charset/charconv: clarify comments in next_codepoint_convenience_ext()
Michael Adam [Sun, 31 Oct 2010 07:21:41 +0000 (08:21 +0100)]
lib/util/charset/charconv: clarify comments in next_codepoint_convenience_ext()

Give the unicod U+<hexnumber> notation of the codepoints
referred to in the comments. Also reformat the comments some.

9 years agolib/util/charset/util_unistr: clarify the comment header for strlen_m().
Michael Adam [Sun, 31 Oct 2010 07:02:17 +0000 (08:02 +0100)]
lib/util/charset/util_unistr: clarify the comment header for strlen_m().

9 years agolib/util/charset/util_unistr: add strlen_m_ext_term() - variant of strlen_m_ext(...
Michael Adam [Sun, 31 Oct 2010 00:02:16 +0000 (02:02 +0200)]
lib/util/charset/util_unistr: add strlen_m_ext_term() - variant of strlen_m_ext() counting terminator

9 years agolib/util/charset/util_unistr: add strlen_m_ext that takes input and output charset
Michael Adam [Sat, 30 Oct 2010 00:03:02 +0000 (02:03 +0200)]
lib/util/charset/util_unistr: add strlen_m_ext that takes input and output charset

The function calculates the number of units (8 or 16-bit, depending
on the destination charset), that would be needed to convert the
input string which is expected to be in in src_charset encoding
to the dst_charset (which should be a unicode charset).

9 years agolib/charcnv/util_unistr: add next_codepoint_ext() that accepts input charset.
Michael Adam [Sun, 31 Oct 2010 01:18:46 +0000 (02:18 +0100)]
lib/charcnv/util_unistr: add next_codepoint_ext() that accepts input charset.

next_codepoint() takes as string in CH_UNIX encoding and returns the
unicode codepoint of the next (possibly multibyte) character of the
input string.

The new next_codepoint_ext() function adds the encoding of the input
string as a parameter. next_codepoint() now only calls next_codepoint_ext()
with CH_UNIX als src_charset argument.

9 years agolib/charset/charcnv: rename a parameter of next_codepoint_convenience_ext() for clarity
Michael Adam [Mon, 1 Nov 2010 14:53:43 +0000 (15:53 +0100)]
lib/charset/charcnv: rename a parameter of next_codepoint_convenience_ext() for clarity

9 years agolib/charset/charcnv: add next_codepoint_convenience_ext() that accepts input charset.
Michael Adam [Fri, 29 Oct 2010 20:06:05 +0000 (22:06 +0200)]
lib/charset/charcnv: add next_codepoint_convenience_ext() that accepts input charset.

next_codepoint_convenience() takes as string in CH_UNIX encoding and returns the
unicode codepoint of the next (possibly multibyte) character of the
input string.

The new next_codepoint_convenience_ext() function adds the encoding of the input
string as a parameter. next_codepoint_convenience() now only calls
next_codepoint_convenience_ext() with CH_UNIX als src_charset argument.

9 years agos3:util_str: add strlen_m_ext_term() - variant of strlen_m_ext() counting terminator
Michael Adam [Sun, 31 Oct 2010 00:02:16 +0000 (02:02 +0200)]
s3:util_str: add strlen_m_ext_term() - variant of strlen_m_ext() counting terminator

9 years agos3:lib/util_str: add strlen_m_ext() that takes input and output charset
Michael Adam [Sat, 30 Oct 2010 00:03:02 +0000 (02:03 +0200)]
s3:lib/util_str: add strlen_m_ext() that takes input and output charset

The function calculates the number of units (8 or 16-bit, depending
on the destination charset), that would be needed to convert the
input string which is expected to be in in src_charset encoding
to the dst_charset (which should be a unicode charset).

9 years agos3:lib/util_str: clarify the comment header for strlen_m().
Michael Adam [Fri, 29 Oct 2010 20:21:47 +0000 (22:21 +0200)]
s3:lib/util_str: clarify the comment header for strlen_m().

9 years agos3:lib/charcnv: clarify comments in next_codepoint_ext()
Michael Adam [Fri, 29 Oct 2010 20:11:30 +0000 (22:11 +0200)]
s3:lib/charcnv: clarify comments in next_codepoint_ext()

(giving the unicod U+<hexnumber> notation of the codepoints
 referred to in the comments)

9 years agos3:lib/charcnv: rename a parameter for clarity in next_codepoint_ext()
Michael Adam [Mon, 1 Nov 2010 14:42:21 +0000 (15:42 +0100)]
s3:lib/charcnv: rename a parameter for clarity in next_codepoint_ext()

9 years agos3:lib/charcnv: reformat comments in next_codepoint_ext()
Michael Adam [Mon, 1 Nov 2010 14:42:21 +0000 (15:42 +0100)]
s3:lib/charcnv: reformat comments in next_codepoint_ext()

9 years agos3:lib/charcnv: add next_codepoint_ext() that accepts input charset.
Michael Adam [Fri, 29 Oct 2010 20:06:05 +0000 (22:06 +0200)]
s3:lib/charcnv: add next_codepoint_ext() that accepts input charset.

next_codepoint() takes as string in CH_UNIX encoding and returns the
unicode codepoint of the next (possibly multibyte) character of the
input string.

The new next_codepoint_ext() function adds the encoding of the input
string as a parameter. next_codepoint() now only calls next_codepoint_ext()
with CH_UNIX als src_charset argument.

9 years agoutil/charset: remove a duplicate comment.
Michael Adam [Fri, 29 Oct 2010 18:50:28 +0000 (20:50 +0200)]
util/charset: remove a duplicate comment.

This seems to have been copied twice from source3/ code.

9 years agobuild: a more portable way of finding waf in makefiles
Andrew Tridgell [Wed, 3 Nov 2010 21:29:11 +0000 (08:29 +1100)]
build: a more portable way of finding waf in makefiles

this avoids using the non-portable shell command in makefiles

Autobuild-User: Andrew Tridgell <tridge@samba.org>
Autobuild-Date: Wed Nov  3 22:44:59 UTC 2010 on sn-devel-104

9 years agoselftest/tests.py: Fix typos.
Jelmer Vernooij [Wed, 3 Nov 2010 18:14:02 +0000 (19:14 +0100)]
selftest/tests.py: Fix typos.

Autobuild-User: Jelmer Vernooij <jelmer@samba.org>
Autobuild-Date: Wed Nov  3 19:55:55 UTC 2010 on sn-devel-104

9 years agoprovision: Look for in source setup data first, in case an older version
Jelmer Vernooij [Wed, 3 Nov 2010 18:13:03 +0000 (19:13 +0100)]
provision: Look for in source setup data first, in case an older version
of samba 4 is installed in the system.

9 years agosamba: Make in_source_tree() public.
Jelmer Vernooij [Wed, 3 Nov 2010 18:12:17 +0000 (19:12 +0100)]
samba: Make in_source_tree() public.

9 years agourgent_replication: Fix formatting
Jelmer Vernooij [Wed, 3 Nov 2010 18:11:26 +0000 (19:11 +0100)]
urgent_replication: Fix formatting

9 years agopossibleinferiors: Report subunit.
Jelmer Vernooij [Wed, 3 Nov 2010 17:43:38 +0000 (18:43 +0100)]
possibleinferiors: Report subunit.

9 years agopossibleinferiors: Fix formatting.
Jelmer Vernooij [Wed, 3 Nov 2010 17:41:49 +0000 (18:41 +0100)]
possibleinferiors: Fix formatting.

9 years agoreplace/wscript: add size checks for stdint.h types
Stefan Metzmacher [Wed, 3 Nov 2010 16:40:39 +0000 (17:40 +0100)]
replace/wscript: add size checks for stdint.h types

metze

Autobuild-User: Stefan Metzmacher <metze@samba.org>
Autobuild-Date: Wed Nov  3 19:12:39 UTC 2010 on sn-devel-104

9 years agoreplace/wscript: do the size checks directly after the type checks
Stefan Metzmacher [Wed, 3 Nov 2010 16:40:12 +0000 (17:40 +0100)]
replace/wscript: do the size checks directly after the type checks

metze

9 years agoreplace/wscript: define bool to int instead of off_t
Stefan Metzmacher [Wed, 3 Nov 2010 16:35:47 +0000 (17:35 +0100)]
replace/wscript: define bool to int instead of off_t

metze

9 years agoreplace/wscript: check for uint8_t
Stefan Metzmacher [Wed, 3 Nov 2010 16:34:35 +0000 (17:34 +0100)]
replace/wscript: check for uint8_t

metze

9 years agos4:dsdb/samldb: avoid nested unindexed searches in samldb_member_check()
Stefan Metzmacher [Wed, 3 Nov 2010 11:34:47 +0000 (12:34 +0100)]
s4:dsdb/samldb: avoid nested unindexed searches in samldb_member_check()

With 20000 objects in the database it's no fun to add members...

metze

9 years agos4:dsdb/tests/python/sam.py: test with member: <SID=...>
Stefan Metzmacher [Wed, 3 Nov 2010 16:41:10 +0000 (17:41 +0100)]
s4:dsdb/tests/python/sam.py: test with member: <SID=...>

metze

9 years agos4:descriptor LDB module - a bit cleanup
Matthias Dieter Wallnöfer [Mon, 1 Nov 2010 16:50:23 +0000 (17:50 +0100)]
s4:descriptor LDB module - a bit cleanup

- add more OOM checks where needed
- remove message of an error which cannot happen anymore (since now the
  structural objectclass is always checked by the objectclass LDB module)

Autobuild-User: Matthias Dieter Wallnöfer <mdw@samba.org>
Autobuild-Date: Wed Nov  3 18:30:52 UTC 2010 on sn-devel-104

9 years agos4:sam.py - assign valid values when performing the special-attributes constraint...
Matthias Dieter Wallnöfer [Wed, 3 Nov 2010 17:02:28 +0000 (18:02 +0100)]
s4:sam.py - assign valid values when performing the special-attributes constraint checks

The problem is that s4 per construction does the checks in a different order. It
first checks for validity (pre-operation trigger in samldb LDB module) and then
for the schema (post-operation trigger in objectclass_attrs LDB module).
constraints (post-operation trigger

9 years agos4:samldb LDB module - improve the "get_single_valued_attr" call and move it into...
Matthias Dieter Wallnöfer [Wed, 3 Nov 2010 16:50:30 +0000 (17:50 +0100)]
s4:samldb LDB module - improve the "get_single_valued_attr" call and move it into "ldb_modules/util.c"

It will be used by other LDB modules as well.

9 years agos4:objectclass_attrs LDB module - introduce the hardcoded delete-protected attributes...
Matthias Dieter Wallnöfer [Wed, 3 Nov 2010 16:29:58 +0000 (17:29 +0100)]
s4:objectclass_attrs LDB module - introduce the hardcoded delete-protected attributes list

This is a hardcoded list in AD of attributes, which can never be deleted.

9 years agopydsdb: Import testtools before subunit for those that don't have
Jelmer Vernooij [Wed, 3 Nov 2010 16:53:19 +0000 (17:53 +0100)]
pydsdb: Import testtools before subunit for those that don't have
testtools installed.

Also, cleanup some imports.

Autobuild-User: Jelmer Vernooij <jelmer@samba.org>
Autobuild-Date: Wed Nov  3 17:47:55 UTC 2010 on sn-devel-104

9 years agofilter-subunit: Add --fail-on-empty argument.
Jelmer Vernooij [Wed, 3 Nov 2010 15:30:40 +0000 (16:30 +0100)]
filter-subunit: Add --fail-on-empty argument.

9 years agos3: Align nttrans replies the same way Windows does it
Volker Lendecke [Fri, 3 Jul 2009 11:28:01 +0000 (13:28 +0200)]
s3: Align nttrans replies the same way Windows does it

If you look at traces of "smbcacls" reading a secdesc from Samba and from W2k3
shows this difference in the querysecdesc nttrans reply. This patch is
necessary to enable viewing previous versions via the shadow copy modules from
Windows 2008. The other client versions seem not to care, but W2k8 does.

Signed-off-by: Michael Adam <obnox@samba.org>
Autobuild-User: Volker Lendecke <vlendec@samba.org>
Autobuild-Date: Wed Nov  3 17:04:28 UTC 2010 on sn-devel-104

9 years agos4:objectclass LDB module - the structural objectclass has always to be specified
Matthias Dieter Wallnöfer [Mon, 1 Nov 2010 14:23:48 +0000 (15:23 +0100)]
s4:objectclass LDB module - the structural objectclass has always to be specified

Autobuild-User: Matthias Dieter Wallnöfer <mdw@samba.org>
Autobuild-Date: Wed Nov  3 16:20:55 UTC 2010 on sn-devel-104

9 years agos4:objectclass_attrs LDB module - refactor the "dSHeuristics" checker
Matthias Dieter Wallnöfer [Wed, 3 Nov 2010 15:25:17 +0000 (16:25 +0100)]
s4:objectclass_attrs LDB module - refactor the "dSHeuristics" checker

The checks are done when there are more than 0 values. The other checks should
be performed by the other parts of the module.

9 years agos4:objectclass_attrs LDB module - simplify the invoke of the "dSHeuristics" checker
Matthias Dieter Wallnöfer [Wed, 3 Nov 2010 15:21:37 +0000 (16:21 +0100)]
s4:objectclass_attrs LDB module - simplify the invoke of the "dSHeuristics" checker

It's always invoked on add and modify operations.

9 years agos4-tests: Tests for the dSHeuristics attribute value restrictions
Nadezhda Ivanova [Wed, 3 Nov 2010 13:15:02 +0000 (15:15 +0200)]
s4-tests: Tests for the dSHeuristics attribute value restrictions

Autobuild-User: Nadezhda Ivanova <nivanova@samba.org>
Autobuild-Date: Wed Nov  3 13:58:42 UTC 2010 on sn-devel-104

9 years agos4-dsdb: Implemented value restrictions for the dSHeuristics attribute
Nadezhda Ivanova [Wed, 3 Nov 2010 13:14:06 +0000 (15:14 +0200)]
s4-dsdb: Implemented value restrictions for the dSHeuristics attribute

9 years agos3: Fix a getgrent crash with many groups
Volker Lendecke [Wed, 3 Nov 2010 12:08:37 +0000 (13:08 +0100)]
s3: Fix a getgrent crash with many groups

Autobuild-User: Volker Lendecke <vlendec@samba.org>
Autobuild-Date: Wed Nov  3 12:58:49 UTC 2010 on sn-devel-104

9 years agos4-server: cleanup allocation of process models
Andrew Tridgell [Wed, 3 Nov 2010 09:57:38 +0000 (20:57 +1100)]
s4-server: cleanup allocation of process models

Autobuild-User: Andrew Tridgell <tridge@samba.org>
Autobuild-Date: Wed Nov  3 10:40:52 UTC 2010 on sn-devel-104

9 years agos4-ldb: fixed sorting of module names
Andrew Tridgell [Wed, 3 Nov 2010 09:37:08 +0000 (20:37 +1100)]
s4-ldb: fixed sorting of module names

this should also fix a valgrind error on freebsd

thanks to Matthieu for spotting this

9 years agos4:objectguid LDB module - remove "objectguid_find_attribute"
Matthias Dieter Wallnöfer [Wed, 3 Nov 2010 08:39:15 +0000 (09:39 +0100)]
s4:objectguid LDB module - remove "objectguid_find_attribute"

It's exactly the same as "ldb_msg_find_element".

In addition remove a comment which points out a "fixme" for a semi-async call
since we started to permit them again.

Autobuild-User: Matthias Dieter Wallnöfer <mdw@samba.org>
Autobuild-Date: Wed Nov  3 09:29:46 UTC 2010 on sn-devel-104

9 years agos4:RPC server - always set the response pad data in base of the request one
Matthias Dieter Wallnöfer [Wed, 3 Nov 2010 08:23:33 +0000 (09:23 +0100)]
s4:RPC server - always set the response pad data in base of the request one

Otherwise it could remain uninitialised - should fix bug #7769.

9 years agos4:RPC server - cosmetic fixes - indentation, comments
Matthias Dieter Wallnöfer [Wed, 3 Nov 2010 08:23:00 +0000 (09:23 +0100)]
s4:RPC server - cosmetic fixes - indentation, comments

9 years agos4:waf - fix the build on Gentoo platforms
Philip M. White [Wed, 3 Nov 2010 07:21:34 +0000 (08:21 +0100)]
s4:waf - fix the build on Gentoo platforms

Signed-off-by: Matthias Dieter Wallnöfer <mdw@samba.org>
9 years agos4-test: if using a system version of ldb, override the ldb modules path
Andrew Tridgell [Wed, 3 Nov 2010 03:26:04 +0000 (14:26 +1100)]
s4-test: if using a system version of ldb, override the ldb modules path

Autobuild-User: Andrew Tridgell <tridge@samba.org>
Autobuild-Date: Wed Nov  3 04:13:22 UTC 2010 on sn-devel-104

9 years agowaf: more agressively expand subsystem syslib deps
Andrew Tridgell [Wed, 3 Nov 2010 01:23:43 +0000 (12:23 +1100)]
waf: more agressively expand subsystem syslib deps

this solves an openchange build problem with an indirect dependency on
talloc when talloc is a syslib

Autobuild-User: Andrew Tridgell <tridge@samba.org>
Autobuild-Date: Wed Nov  3 02:17:30 UTC 2010 on sn-devel-104

9 years agos4-ldb: make ldbtest depend on ldb
Andrew Tridgell [Wed, 3 Nov 2010 01:22:21 +0000 (12:22 +1100)]
s4-ldb: make ldbtest depend on ldb

9 years agowaf: added reconfigure targets to our libraries
Andrew Tridgell [Wed, 3 Nov 2010 01:09:23 +0000 (12:09 +1100)]
waf: added reconfigure targets to our libraries

This allows you to do "make reconfigure" to re-run configure only if
needed

9 years agowaf: ensure pkgconfig files depend on the prefix
Andrew Tridgell [Wed, 3 Nov 2010 00:30:23 +0000 (11:30 +1100)]
waf: ensure pkgconfig files depend on the prefix

9 years agowaf: split pkgconfig logic into a separate module
Andrew Tridgell [Wed, 3 Nov 2010 00:14:40 +0000 (11:14 +1100)]
waf: split pkgconfig logic into a separate module

9 years agos4-build: use -Wl,--as-needed if supported
Andrew Tridgell [Tue, 2 Nov 2010 23:50:16 +0000 (10:50 +1100)]
s4-build: use -Wl,--as-needed if supported

this makes our runtime load times smaller, and also means we find bugs
that this may cause on gentoo. gentoo uses this by default

Autobuild-User: Andrew Tridgell <tridge@samba.org>
Autobuild-Date: Wed Nov  3 00:35:56 UTC 2010 on sn-devel-104

9 years agowaf: fixed wildcard build for full paths
Andrew Tridgell [Tue, 2 Nov 2010 23:49:08 +0000 (10:49 +1100)]
waf: fixed wildcard build for full paths

9 years agowaf: rerun deps calculation on LDFLAGS or CFLAGS change
Andrew Tridgell [Tue, 2 Nov 2010 23:25:38 +0000 (10:25 +1100)]
waf: rerun deps calculation on LDFLAGS or CFLAGS change

9 years agowaf: fixed wildcard build
Andrew Tridgell [Tue, 2 Nov 2010 21:18:05 +0000 (08:18 +1100)]
waf: fixed wildcard build

this fixes minimal includes for subsystems that use autoproto

9 years agos4-ldb: give the user a hint as to what may be wrong
Andrew Tridgell [Mon, 1 Nov 2010 23:58:49 +0000 (10:58 +1100)]
s4-ldb: give the user a hint as to what may be wrong

when a backend or module can't be found, give a hint about
the LDB_MODULES_PATH environment variable

Autobuild-User: Andrew Tridgell <tridge@samba.org>
Autobuild-Date: Tue Nov  2 21:28:22 UTC 2010 on sn-devel-104

9 years agos4-dsdb: removed the use of ldb_private.h from s4
Andrew Tridgell [Mon, 1 Nov 2010 23:41:28 +0000 (10:41 +1100)]
s4-dsdb: removed the use of ldb_private.h from s4

this will allow s4 to use a system version of ldb

9 years agos4-ldb: expose some more ldb module functions
Andrew Tridgell [Mon, 1 Nov 2010 23:40:52 +0000 (10:40 +1100)]
s4-ldb: expose some more ldb module functions

this exposes enough functions for s4 to not need ldb_private.h any
more.

9 years agos3: Add "net registry getvaluesraw"
Volker Lendecke [Tue, 15 Dec 2009 19:17:01 +0000 (20:17 +0100)]
s3: Add "net registry getvaluesraw"

Autobuild-User: Volker Lendecke <vlendec@samba.org>
Autobuild-Date: Tue Nov  2 15:42:22 UTC 2010 on sn-devel-104

9 years agos3: Do not connect to ctdb if it is blocked for some reason
Volker Lendecke [Mon, 16 Nov 2009 11:03:24 +0000 (12:03 +0100)]
s3: Do not connect to ctdb if it is blocked for some reason

9 years agocluster_fatal() exit code should not indicate success.
Martin Schwenke [Thu, 22 Oct 2009 11:03:20 +0000 (13:03 +0200)]
cluster_fatal() exit code should not indicate success.

cluster_fatal() logs a fatal event and then exits with 0.  This seems
wrong.  Sometimes command like "net" use this code and return
incorrect empty output but then exit with 0.

This simply changes the exit code to 1.

Signed-off-by: Martin Schwenke <martin@meltin.net>
9 years agos3: Fix a typo in a DEBUG msg
Volker Lendecke [Tue, 2 Nov 2010 12:09:50 +0000 (13:09 +0100)]
s3: Fix a typo in a DEBUG msg

Autobuild-User: Volker Lendecke <vlendec@samba.org>
Autobuild-Date: Tue Nov  2 13:13:39 UTC 2010 on sn-devel-104

9 years agoMAINTAINERS: add myself as maintainer for tevent and tsocket
Stefan Metzmacher [Tue, 2 Nov 2010 10:32:30 +0000 (11:32 +0100)]
MAINTAINERS: add myself as maintainer for tevent and tsocket

metze

Autobuild-User: Stefan Metzmacher <metze@samba.org>
Autobuild-Date: Tue Nov  2 12:30:21 UTC 2010 on sn-devel-104

9 years agos4-selftest Run RPC-PAC against all the DC environments.
Andrew Bartlett [Tue, 2 Nov 2010 10:57:13 +0000 (21:57 +1100)]
s4-selftest Run RPC-PAC against all the DC environments.

This is important, as it will validate that the PAC handling is
correct with the different encryption types.

Andrew Bartlett

Autobuild-User: Andrew Bartlett <abartlet@samba.org>
Autobuild-Date: Tue Nov  2 11:45:50 UTC 2010 on sn-devel-104

9 years agos4-process_model Fix valgrind-found use of un-initialised variable
Andrew Bartlett [Tue, 2 Nov 2010 09:55:18 +0000 (20:55 +1100)]
s4-process_model Fix valgrind-found use of un-initialised variable

The relloc didn't zero the structure for the new process modal, and so
m->initialised wasn't initialised.

Andrew Bartlett

9 years agos4-torture Add tests for DES-only accounts PAC behaviour/validation.
Andrew Bartlett [Tue, 2 Nov 2010 06:14:55 +0000 (17:14 +1100)]
s4-torture Add tests for DES-only accounts PAC behaviour/validation.

Previously we didn't support DES-only in a Samba4 domain.  This is
important for some legacy systems that have not yet migrated from this
weak crypto.

Andrew Bartlett

9 years agos4-selftest Allow weak crypto so we can test DES-only behaviour.
Andrew Bartlett [Tue, 2 Nov 2010 06:10:47 +0000 (17:10 +1100)]
s4-selftest Allow weak crypto so we can test DES-only behaviour.

Andrew Bartlett

9 years agoheimdal Add handling for PAC signatures over all encryption types
Andrew Bartlett [Tue, 2 Nov 2010 06:08:34 +0000 (17:08 +1100)]
heimdal Add handling for PAC signatures over all encryption types

There are exceptions from the expected behaviour of 'checksum type
matches key type' that we must deal with here, or else we can't serve
DES-only servers.

Andrew Bartlett

9 years agos3-param Fix up lp_set_cmdline() not to re-store cmdline options on each reload
Andrew Bartlett [Tue, 2 Nov 2010 04:33:42 +0000 (15:33 +1100)]
s3-param Fix up lp_set_cmdline() not to re-store cmdline options on each reload

The previous code was buggy in that it did not honour the 'store'
argument to lp_set_cmdline_helper(), and would use the stored
parameter after freeing it when handling overwritten values.

Andrew Bartlett

Autobuild-User: Andrew Bartlett <abartlet@samba.org>
Autobuild-Date: Tue Nov  2 05:19:17 UTC 2010 on sn-devel-104

9 years agos3-libsmbclient Don't store 'debug_stderr' on the libsmbclient context
Andrew Bartlett [Tue, 2 Nov 2010 04:32:16 +0000 (15:32 +1100)]
s3-libsmbclient Don't store 'debug_stderr' on the libsmbclient context

Following the review of this patch series by Derrell Lipman, remove
the seperate storage of the debug_stderr variable from the
libsmbclient SMBC_internal_data context.

Andrew Bartlett

9 years agos3-libsmbclient Add comments to describe the behaviour of DEBUG()
Andrew Bartlett [Mon, 1 Nov 2010 09:55:04 +0000 (20:55 +1100)]
s3-libsmbclient Add comments to describe the behaviour of DEBUG()

This isn't quite what you would expect from this interface, but actually
avoids some really nasty situations if you ever have more than one
libsmbclient context in a process.

In the real world, if you have asked for DEBUG() to stderr in one part
of the code, you will want it globally, even in a different thread
(which in the past would have rest everything to stdout again, at
least while starting up).

Andrew Bartlett