1
0
mirror of https://github.com/samba-team/samba.git synced 2024-12-24 21:34:56 +03:00
Commit Graph

11837 Commits

Author SHA1 Message Date
Volker Lendecke
9606397ab1 Fix a memory leak in pdb_tdb.c.
Simo, I remember you complaining about a memleak there, could it be
this one, or did you resolve it at that time?

Volker
(This used to be commit c660595ded)
2003-06-17 12:03:48 +00:00
Jelmer Vernooij
c0e24984f0 Make static (patch from metze)
(This used to be commit 908b16cc2a)
2003-06-17 10:48:06 +00:00
Jelmer Vernooij
b9bf3252bf Fix building idmap_winbind as shared module
(This used to be commit 2556037796)
2003-06-17 10:41:08 +00:00
Jelmer Vernooij
cafb5c512d The return value of init_module functions is NTSTATUS, not int
(This used to be commit f09df852ac)
2003-06-17 10:38:22 +00:00
Volker Lendecke
7dae7d3fb4 Fix memory leak in idmap. Valgrind is soooo cool.
Volker
(This used to be commit dcdb6683a7)
2003-06-17 10:16:20 +00:00
Volker Lendecke
ea1cec68bb Const fixes by metze
Volker
(This used to be commit c0e35f3be8)
2003-06-17 09:40:35 +00:00
Jelmer Vernooij
ac44100674 Update link to DTD
(This used to be commit 2737518b25)
2003-06-16 21:34:59 +00:00
Richard Sharpe
5dd2d09a37 Squelch one small compiler warning
(This used to be commit d9ac2540dd)
2003-06-16 18:37:54 +00:00
Richard Sharpe
2740d35a7b Replace all use of bzero with memset ...
(This used to be commit e21aab516b)
2003-06-16 17:49:01 +00:00
Volker Lendecke
979c447060 This glosses over John's problem at SambaXP 2003. When we want to join
a NT4 domain as a BDC with an existing workstation account (existing
bdc is fine), we fail. Print a friendly error message in this case.

The correct solution would probably be to delete the account and try
again. But even this makes us better than NT: NT4 fails in this
situation with an empty warning message box and an unusable BDC. It
has unsuccessfully tried to suck down the domain database, and thus
has no administrator account to log in after reboot....

Volker
(This used to be commit 1ddeea2179)
2003-06-16 15:59:05 +00:00
Volker Lendecke
e2cda4a24e Fix misleading debug message.
Volker
(This used to be commit a4f76f2520)
2003-06-16 10:10:37 +00:00
Tim Potter
6a2b43db9c Make net rpc vampire return an error if the sam sync RPC returns an error.
E.g if we are pointing at a win2k native mode domain we are returned
an NT_STATUS_NOT_SUPPORTED error.
(This used to be commit 6053c30f26)
2003-06-16 05:39:26 +00:00
Tim Potter
3d65c033da Quieten another debug message.
(This used to be commit 2c47893c75)
2003-06-16 05:22:53 +00:00
Tim Potter
94d1dd96d6 Quieten debug message.
(This used to be commit 02d12cc4b9)
2003-06-16 05:20:37 +00:00
Tim Potter
7c16f64e38 Update nsstest to cope with wins NSS module as well as winbind NSS
module.  Use "wins" as the nss name to invoke this behaviour.

Also, fixed nsstest so it doesn't segfault when a nss function can't
be dlopened().  Log an error and abort the test gracefully instead.
(This used to be commit 66bafbe371)
2003-06-16 03:30:53 +00:00
Tim Potter
35d0f987db Build libnss_wins.so as part of nsswitch target. Fix for bug #160
found by andreas@conectiva.com.br
(This used to be commit fa24fd6f5b)
2003-06-16 03:28:27 +00:00
Andrew Tridgell
ec0303820f we need to call ads_first_entry() before using a ldap result,
otherwise we can segv or return garbage
(This used to be commit d1316656b0)
2003-06-16 02:42:00 +00:00
Andrew Tridgell
8cd67d7668 reverted locale patch put in by jht (originally from vorlon).
There are lots of things wrong with this patch, including:

1) it overrides a user chosen configuration option

2) it adds lots of complexity inside a loop when a tiny piece of code
   outside the loop would do the same thing

3) it does no error checking, and is sure to crash on some systems

If you want this functionality then try something like this at the end
of charset_name():

#ifdef HAVE_NL_LANGINFO
	if (strcasecmp(ret, "LOCALE") == 0) {
		const char *ln = nl_langinfo(CODESET);
		if (ln) {
			DEBUG(5,("Substituting charset '%s' for LOCALE\n", ln));
			return ln;
		}
	}
#endif

then users can set 'display charset = LOCALE' to get the locale based
charset. You could even make that the default for systems that have
nl_langinfo().
(This used to be commit 382b9b806b)
2003-06-16 02:22:52 +00:00
Tim Potter
145b01a496 Fix some compiler warnings.
(This used to be commit eec1758e04)
2003-06-16 02:16:58 +00:00
Andrew Tridgell
1209615504 another improved debug statement
(This used to be commit ac69b9c83c)
2003-06-16 01:13:17 +00:00
Andrew Tridgell
a2039bc4c6 made a debug statement more useful
(This used to be commit 3f7a24d183)
2003-06-16 01:12:42 +00:00
John Terpstra
489956c823 Patch from vorlon@debian.org, see bugzilal #122
Samba should preferentially use the locale information from the native system,
and only fall back on 'display charset' if this is unavailable or unsupported.
(This used to be commit 1e445fb422)
2003-06-15 06:07:53 +00:00
Volker Lendecke
bd1333ea74 Add 'net idmap restore'. This restores a broken idmap file
from the output of 'net idmap dump'.

'net idmap dump' now also prints the USER/GROUP HWM.

Volker
(This used to be commit c0575be936)
2003-06-14 17:51:09 +00:00
Andrew Bartlett
b85664047c This patch modifies 'net rpc vampire' to add new and existing users to both
the idmap and the SAM.

The basic idea is this:  Lookup the user with GetPwnam(), and if they
exist then use that uid.  This is what people expect.  If the user does
not exist, try and run the right script.

This is also what people expect from previous Samba 3.0 behaviour, where
the Get_Pwnam() was at runtime.

If the idmap entry for this SID isn't valid, or isn't the right value,
modify the idmap to account for this mapping.

Also, the same logic is applied to the primary gid - if it has changed,
update the user's primary unix group.

This patch allows users to be added without a mapping - this is fine for
machine accounts, for example.  I've given it a quick test against my
Win2k DC, and I *think* it's sane.

Andrew Bartlett
(This used to be commit d2a70bfff1)
2003-06-14 00:49:02 +00:00
Jeremy Allison
292a51eda1 Forward port the app-head changes for dc name cache into 3.0.
Jeremy.
(This used to be commit 8bcc3116a2)
2003-06-13 21:03:15 +00:00
Richard Sharpe
3d0d8f609d Fix an spelling mistake.
(This used to be commit 8e37c1d579)
2003-06-13 17:49:03 +00:00
Volker Lendecke
8ad364dd2e Trivial extension to 'net' to dump current local idmap.
(This used to be commit 18f3a5efea)
2003-06-13 16:19:02 +00:00
Tim Potter
0a9396dcca Rename some uuid functions so as not to conflict with system
versions.  Fixes bug #154.
(This used to be commit 986eae40f7)
2003-06-13 04:35:53 +00:00
Tim Potter
61742d1117 Fix shadow variable warning.
(This used to be commit c22a4074bd)
2003-06-13 04:29:20 +00:00
Tim Potter
88a5cb83a5 Delete outdated file.
(This used to be commit 406956645f)
2003-06-13 00:07:59 +00:00
Volker Lendecke
e1f4d66b73 Fix for bug#3. Show comments when doing 'net group -l'.
Volker
(This used to be commit e5664adc07)
2003-06-12 16:21:22 +00:00
Volker Lendecke
c8667b03f8 Working on bug#3. We want all of the aliases, so start with 0.
Volker
(This used to be commit ec1a58d09e)
2003-06-12 15:37:06 +00:00
Volker Lendecke
511789b855 Andrew's change to make 'security = domain' work again. Leave the user and group
that has been authenticated unmapped.

We need to make sure that every caller of idmap handles failure gracefully.

Volker
(This used to be commit 902d607b66)
2003-06-12 14:24:15 +00:00
Volker Lendecke
2cd38cd8e8 Fix some misleading debug messages.
(This used to be commit 9c003ae4ff)
2003-06-12 08:22:55 +00:00
Volker Lendecke
c899e4c874 Get rid of a const warning.
Volker
(This used to be commit 6684f5b11f)
2003-06-12 07:54:13 +00:00
Volker Lendecke
3117b4b6fc Fix typo found by Vance
(This used to be commit 678e68a295)
2003-06-12 07:32:44 +00:00
Volker Lendecke
0d8307fac3 Andrew is right, this is probably the wrong approach.
Take away the automatic mapping of users and groups again.

Volker
(This used to be commit 74510369d4)
2003-06-12 07:21:26 +00:00
Volker Lendecke
98c9504583 Set the user's primary unix group from usrmgr.exe.
This part of a fix to bug#45.

Volker
(This used to be commit 43d306011f)
2003-06-11 20:42:10 +00:00
Volker Lendecke
36a9033de8 Sorry idra for messing around with your stuff, but this was obviously not
tested.

Incrementing HWW_USER while allocating a GROUPID looked somewhat wrong.

Volker
(This used to be commit d1eac2c758)
2003-06-11 20:06:11 +00:00
Simo Sorce
6d360a15d9 avoid races in getting high watermark
(This used to be commit df0df941d8)
2003-06-11 18:14:34 +00:00
Volker Lendecke
a7e1bbbd06 Fix 'security = domain' without winbind. This stores the sid we got
from the PDC as a mapping to the uid we got from getpwnam in the
local idmap.

This should not be worse than the current state, so I decided to
commit it. It is different from abartlet's preliminary patch, but I
believe this is the better solution. Feel free to comment and/or
revert it.

Volker
(This used to be commit 0c16965e6f)
2003-06-11 16:36:04 +00:00
Andrew Tridgell
1d03ec88f1 several places in client.c rely in commands[i].name == NULL being a
termination condition for checking for commands. Make sure the last
element is indeed NULL.
(This used to be commit 294f79884d)
2003-06-11 06:23:53 +00:00
Andrew Tridgell
98f448b1f4 i'm getting rather sick of this
**********************************************************
**** PLEASE TEST YOUR CHANGES BEFORE CHECKING THEM IN ****
**********************************************************

Jelmer, your change did not fix the -b option, it just make it not
segfault on debian. It did however add a warning and make it not
compile on some systems. It also gave the io_bufsize the value of a
pointer, not the value the user specified.

(this doesn't apply just to Jelmer, I'm just seeing rather too many
checkins from everyone lately where the code could never have possibly
worked, so was obviously checked in without testing)
(This used to be commit 21931ad7a0)
2003-06-11 06:10:10 +00:00
Andrew Tridgell
d368845418 on AIX FD_ZERO() is defined in terms of bzero(), so we can't have
the "don't use bzero" macros.

In general I think it would be better to have a separate script that
checks for deprecated functions like these using grep rather than
using these cpp tricks. They just get us into trouble.
(This used to be commit 2a227c880d)
2003-06-11 05:34:14 +00:00
Jeremy Allison
3ef930a8e9 Add in rety loop for query_user_list (from APP_HEAD). Deals with a bug
using MSRPC backend and should be safe with ldap backend.
Jeremy.
(This used to be commit 67535329a2)
2003-06-10 22:11:30 +00:00
Jelmer Vernooij
33a7f1bba0 Don't segfault when -b is specified (fixes debian bug #196833)
(This used to be commit 4af9074397)
2003-06-10 19:27:32 +00:00
Jeremy Allison
5cee22714c Ok, I've tried being Mr. Nice Guy and people (you know who you are) still
keep putting bzero BSD'ism's into our source code. Make this an error like
bcopy and others to prevent it in future.
Jeremy.
(This used to be commit 80d0432316)
2003-06-10 17:30:28 +00:00
Richard Sharpe
09530bfa60 More updates on editreg.c to bring it better in line with the Samba source.
(This used to be commit dc69a638b9)
2003-06-10 17:18:07 +00:00
Andrew Tridgell
50585d4c04 removed editreg from standard build until it is portable. Right now it
fails to build on a ton of platforms as it completely bypasses all of
our portability code.

if you want it then use 'make bin/editreg'. If some distros want to
add that to their spec files then thats up to them, but we really
can't have non-portable code unconditionally built in our main tree.
(This used to be commit 3c66111f32)
2003-06-10 14:09:20 +00:00
Andrew Tridgell
d286e44209 fixed libsmb code to set correct timeout in cli_state when waiting for
a blocking lock
(This used to be commit 3515476fe4)
2003-06-10 14:02:46 +00:00