cvs2svn Import User [Thu, 13 Apr 2000 03:40:09 +0000 (03:40 +0000)]
This commit was manufactured by cvs2svn to create tag
'release-alpha-2-4'.
Samba Release Account [Thu, 13 Apr 2000 03:40:08 +0000 (03:40 +0000)]
preparing for release of alpha.2.4
Luke Leighton [Thu, 13 Apr 2000 03:27:51 +0000 (03:27 +0000)]
#ifdef AUTOCONF moved to avoid compile warning.
Luke Leighton [Thu, 13 Apr 2000 02:58:29 +0000 (02:58 +0000)]
- got fed up of vuser_db spewing 150 lines of trash, made it possible
to increase log levels for parse info. *whew*, now i can actually
_read_ log.smb.
- codepage koi8-r not KOI-R
- fixup of util_sock.c from yesterday, all very painful. took out
"min read length" from read_pipe because NT5 expects it to be
ignored. stupid. i mean, what's the damn point of putting
something in a protocol if it's never used, man?
- disabled SChannel because it's still not quite complete (the
sequence counfounders are wrong, except i'd never seen anything
other than sequence 1-request or sequence 2-request).
- all the rest is getting UDP 138 SAMLOGON and GETDC working. again.
Tim Potter [Thu, 13 Apr 2000 02:16:00 +0000 (02:16 +0000)]
Decoupled domain related information from uid and gid ranges.
make proto.
Tim Potter [Thu, 13 Apr 2000 02:14:37 +0000 (02:14 +0000)]
Removed some annoying debug messages.
Decoupled domain related information from uid and gid ranges.
Don't try to process machine accounts when enumerating users.
Jeremy Allison [Wed, 12 Apr 2000 23:01:09 +0000 (23:01 +0000)]
Removed "ole locking compat" parameter (no longer used).
We now get/set/check POSIX locks, but I still need to code up the
close fd braindamage...
Jeremy.
Jeremy Allison [Wed, 12 Apr 2000 21:47:29 +0000 (21:47 +0000)]
Implmented mapping of lock offset/count from 64 bit MS ranges
to either 63 or 31 bit POSIX ranges. Code to get these locks
not yet added.
Jeremy.
Jeremy Allison [Wed, 12 Apr 2000 18:04:48 +0000 (18:04 +0000)]
Code to map tdb locks onto POSIX. Mainly placeholder code at the moment,
but the structure is done enough so that Andrew can look it over and give
a yea/nay decision.
Jeremy.
Luke Leighton [Wed, 12 Apr 2000 08:18:12 +0000 (08:18 +0000)]
uhh, guys: i spent my entire day updating from yesterday's cvs main
commits. i mean, nothing else.
i wasn't able to do any debugging of the issues from yesterday.
so now, there are two possible sets of bugs, compounded. this
commit compiles but hasn't been tested, i haven't had a chance
to.
i'd really appreciate it if this stopped happening. once is enough.
now i am going to get something to eat and if i am lucky, i will be
able to hold a fork. this is serious.
now i've updated TNG to cvs main, don't make me maintain it, please
do it yourselves.
thanks.
p.s i'm never, ever, working in a branch again. if that's not ok,
i won't do the work. it's not worth it.
Tim Potter [Wed, 12 Apr 2000 06:22:26 +0000 (06:22 +0000)]
winbindd prototype Mk II
Makefile.in:
- trashed winbindd_proto.h generation
- made NTDOM_OBJS look nice and pretty
- testsuite directory now a toplevel one
- added WINBINDD_OBJ to PROTO_OBJ
- added nsswitch target to make ntdom.so and winbindd
Tim Potter [Wed, 12 Apr 2000 05:59:25 +0000 (05:59 +0000)]
Missed a 0x02000000 to SEC_RIGHTS_MAXIMUM_ALLOWED conversion.
Tim Potter [Wed, 12 Apr 2000 05:59:00 +0000 (05:59 +0000)]
winbindd prototype Mk II
script/mkproto.awk:
- added "enum winbindd_result" as a function return value
Tim Potter [Wed, 12 Apr 2000 05:53:26 +0000 (05:53 +0000)]
winbindd prototype Mk II
include/smb.h:
- #include winbindd.h
include/proto.h:
- make proto
Tim Potter [Wed, 12 Apr 2000 05:51:33 +0000 (05:51 +0000)]
Added *.lo
Tim Potter [Wed, 12 Apr 2000 05:45:35 +0000 (05:45 +0000)]
winbindd prototype Mk II
nsswitch/winbindd.c:
- converted lots of 0x02000000 to SEC_RIGHTS_MAXIMUM_ALLOWED
- removed debugging DEBUG statements
- removed global vars for domain controller sid and domain name
- multiple domain support
- fixed logging
nsswitch/winbindd_user.c:
nsswitch/winbindd_group.c:
- mega rewrite of nearly everything - works much better now!
- multiple domain support
- correct operation for multiple simultaneous get{pw,gr}ent() calls
nsswitch/winbindd_sock.c:
- renamed socket timeout constant
nsswitch/winbindd_surs.c
- removed initialised arrays of well-known local and domain objects
- trashed all the dynamically uid/allocation stuff based on tdb
- added sid and domain name lookup functions
- parse "winbind [ug]id" smb.conf parameters
- implemented sid <--> unixid mapping functions
Tim Potter [Wed, 12 Apr 2000 05:27:29 +0000 (05:27 +0000)]
winbindd prototype Mk II
param/loadparam.c:
- added two new parameters "winbind uid" and "winbind gid" giving the
range of user and group ids the winbind daemon maps NT users into.
Usage:
winbind uid = TPOTDOM/3000-10999, BUILTIN/2000-2999
winbind gid = TPOTDOM/6000-60999, BUILTIN/2000-5999
Tim Potter [Wed, 12 Apr 2000 05:21:39 +0000 (05:21 +0000)]
winbindd prototype Mk II
nsswitch/ntdom.c:
- optimistically remove debugging statements
- removal of unused variables
- tidy up of group membership code
- {set,get,end}{pw,gr}ent functions take process pid
- general comment fest
Tim Potter [Wed, 12 Apr 2000 05:15:32 +0000 (05:15 +0000)]
winbindd prototype Mk II
include/winbindd.h:
- include file for winbind daemon
Tim Potter [Wed, 12 Apr 2000 05:05:53 +0000 (05:05 +0000)]
Added *.log and *.sum for DejaGnu test output.
Luke Leighton [Wed, 12 Apr 2000 04:42:46 +0000 (04:42 +0000)]
removing some ptr checks which were causing failures instead of "no members"
reports.
Tim Potter [Wed, 12 Apr 2000 04:03:14 +0000 (04:03 +0000)]
Fixed usage info for samlookuprids.
Luke Leighton [Tue, 11 Apr 2000 09:57:47 +0000 (09:57 +0000)]
removed call to pass_check() in smbd/password.c, it's already called
in the net_login_general() code in netlogond.
Luke Leighton [Tue, 11 Apr 2000 09:38:57 +0000 (09:38 +0000)]
messy messing, eventually discovered a) pass_check.c hadn't been
updated from cvs main b) i needed to compile with --with-pam.
now, encrypt passwords = no, security = user has been tested.
dang me if overloading an info level (decided to create one)
net_login_general, don't actually work :)
Samba Release Account [Tue, 11 Apr 2000 06:48:00 +0000 (06:48 +0000)]
preparing for release of alpha.2.3
Luke Leighton [Tue, 11 Apr 2000 06:42:25 +0000 (06:42 +0000)]
nggh! damn select timeout, FD_ISSET... THING.
Luke Leighton [Tue, 11 Apr 2000 05:37:30 +0000 (05:37 +0000)]
andrew been moving functions about :)
Luke Leighton [Tue, 11 Apr 2000 05:32:38 +0000 (05:32 +0000)]
urr... *wobble*... committing what i _think_ are andrew's smbd changes.
urr...
Luke Leighton [Tue, 11 Apr 2000 05:28:00 +0000 (05:28 +0000)]
change_oem_password() now only takes a UNISTR2 not char*.
Luke Leighton [Tue, 11 Apr 2000 05:27:22 +0000 (05:27 +0000)]
made IPC$ LANMAN functions call client-side DCE/RPC functions.
there isn't an equivalent of NetServerEnum.
Luke Leighton [Tue, 11 Apr 2000 05:25:07 +0000 (05:25 +0000)]
simplifying write_pipe and read_pipe code -- also it looks like there
is a minimum pipe read parameter which i make sure is passed in
to read_pipe, which gets passed in as the minimum value to
read_data_with_timeout()
created two functions read_data_outstanding() and write_data_outstanding()
which do a select with a timeout (0 = instantaneous) to tell if there is
read or write data left.
Tim Potter [Mon, 10 Apr 2000 23:57:20 +0000 (23:57 +0000)]
Fixed misspelling of samquerysec
Jeremy Allison [Mon, 10 Apr 2000 21:04:35 +0000 (21:04 +0000)]
Fix for soft quotas not being set from
Norbert Püschel <Pueschel.Norbert@Walzbarren-VAW.ne.uunet.de>
Jeremy.
Elrond [Mon, 10 Apr 2000 18:23:19 +0000 (18:23 +0000)]
cli_close(), cli_issue_write(), receive_smb():
Check for NULL-pointers. This isn't the real problem, but
it should help finding it.
How to report errors from cli_issue_write() ?
Elrond [Mon, 10 Apr 2000 17:33:13 +0000 (17:33 +0000)]
prs_grow didn't return False, when you wanted to grow in
read-mode beyond the data_size
fudged smb_io_unistr2 to ignore the rest, when we read a
_long_ string (Citrix has a munged_dial-length of 0x185
unicode-chars)
smb_io_logon_hours didn't truncate to the right length
Luke Leighton [Mon, 10 Apr 2000 08:49:59 +0000 (08:49 +0000)]
comment-formatting
Luke Leighton [Mon, 10 Apr 2000 08:49:07 +0000 (08:49 +0000)]
help from matthew brichacek to track down chgpasswd() usage error.
Luke Leighton [Mon, 10 Apr 2000 06:30:36 +0000 (06:30 +0000)]
another debug-only-used thing (a function, this time!)
Luke Leighton [Mon, 10 Apr 2000 06:26:58 +0000 (06:26 +0000)]
added option to compile with absolutely _no_ debugging macros / info.
found some unused variables, already.
Luke Leighton [Mon, 10 Apr 2000 05:24:01 +0000 (05:24 +0000)]
formatting changes. plus, of course, you don't mash the username
into Unix names when calling msrpc functions, that's the job of
the msrpc functions.
cvs2svn Import User [Mon, 10 Apr 2000 03:09:43 +0000 (03:09 +0000)]
This commit was manufactured by cvs2svn to create branch 'SAMBA_TNG'.
Andrew Tridgell [Mon, 10 Apr 2000 03:09:42 +0000 (03:09 +0000)]
split out the lpq parsing code into a separate file
printing/lpq_parse.c
getting ready for the new printing backend
Andrew Tridgell [Mon, 10 Apr 2000 02:57:54 +0000 (02:57 +0000)]
don't echo warning messages twice
Andrew Tridgell [Mon, 10 Apr 2000 02:44:06 +0000 (02:44 +0000)]
utmp compile fix from peter@cadcamlab.org
Samba Release Account [Mon, 10 Apr 2000 02:35:45 +0000 (02:35 +0000)]
preparing for release of alpha.2.2
Luke Leighton [Mon, 10 Apr 2000 02:07:42 +0000 (02:07 +0000)]
extra parameter.
Luke Leighton [Mon, 10 Apr 2000 01:48:03 +0000 (01:48 +0000)]
fixing up recursion-problems in netlogond (told you so!)
turns out that standalone mode (sec=user, dm=no, dl=no) was thinking
it was a PDC, so was referring out and therefore back to itself. argh!
Luke Leighton [Sun, 9 Apr 2000 13:53:41 +0000 (13:53 +0000)]
groups being a pain, had to copy them.
also, the usual corruption on trust account password change...
Luke Leighton [Sun, 9 Apr 2000 08:30:26 +0000 (08:30 +0000)]
trying to track down sys_setgroups() problem: an unbecome_root() was
being done, i think on uninitialised current_user_saved.
Luke Leighton [Sun, 9 Apr 2000 08:28:18 +0000 (08:28 +0000)]
lost smbd_pid in initialisation.
Luke Leighton [Sun, 9 Apr 2000 08:02:04 +0000 (08:02 +0000)]
retired
Luke Leighton [Sun, 9 Apr 2000 07:24:44 +0000 (07:24 +0000)]
making sure that a local authentication request (specifying hostname
as domain name) works.
Luke Leighton [Sun, 9 Apr 2000 06:50:03 +0000 (06:50 +0000)]
it's a little unclear where the line is between netlogond processing
incoming samlogon requests and passing requests on up (using client-side
code). added some debug messages in that say BEGIN here and END here.
by the way, i really, really hope i don't get this wrong, it's going to
cause absolute chaos if netlogond ends up calling itself.
Samba Release Account [Sun, 9 Apr 2000 06:23:31 +0000 (06:23 +0000)]
preparing for release of alpha.2.1
Luke Leighton [Sun, 9 Apr 2000 06:11:35 +0000 (06:11 +0000)]
gets really interesting. it's dog-slow (at debug level 100), but
i think i now have security=domain working.
i also think i fixed the bug where NT5 wks fails with an RPC error
when the user password is wrong.
Luke Leighton [Sun, 9 Apr 2000 05:49:34 +0000 (05:49 +0000)]
- fixing cli_pipe.c asserts
- default client NETLOGON Schannel to false, server Schannel to Auto
- netlogond now sub-refers authentication requests if it's not in
its own SAM database. (yes, sounds horrible, doesn't it!)
Luke Leighton [Sun, 9 Apr 2000 04:43:26 +0000 (04:43 +0000)]
i hate this code. (understatement of the year).
Luke Leighton [Sun, 9 Apr 2000 04:13:54 +0000 (04:13 +0000)]
changed ROLE_DOMAIN_NONE to ROLE_STANDALONE.
Luke Leighton [Sun, 9 Apr 2000 04:04:42 +0000 (04:04 +0000)]
lp_server_role() == SERVER_ROLE_NONE was being refused. hmm. i think
i am going to call this SERVER_ROLE_STANDALONE, or maybe SERVER_ROLE_WGRP.
Luke Leighton [Sun, 9 Apr 2000 03:49:03 +0000 (03:49 +0000)]
good grief. sometimes, i can be very stupid. ok, mike, that mode you
wanted (lp_server_role() = SERVER_ROLE_NONE, which is stand-alone
workgroup member) is working:
security = user
domain master = no
domain logons = no
you should check this with lsaquery command, it should give:
Domain: WORKGROUP SID: S-0-0
Domain: WORKSTATIONNAME SID: S-1-5-21-xxx-yyy-zzz
this just leaves security = domain, which is broke, too.
Luke Leighton [Sun, 9 Apr 2000 03:10:13 +0000 (03:10 +0000)]
update version string.
Samba Release Account [Sun, 9 Apr 2000 03:09:45 +0000 (03:09 +0000)]
preparing for release of alpha.2.0
Luke Leighton [Sun, 9 Apr 2000 03:06:53 +0000 (03:06 +0000)]
add cvsignore
Luke Leighton [Sun, 9 Apr 2000 02:41:16 +0000 (02:41 +0000)]
GETDC request short-detection was wrong, explains why the poor win9x people
couldn't do anything (one, two, three, aaahhh).
Luke Leighton [Sun, 9 Apr 2000 02:34:35 +0000 (02:34 +0000)]
added warning about if a cli_net_use_add fails.
Luke Leighton [Sun, 9 Apr 2000 02:34:15 +0000 (02:34 +0000)]
adding debug warning if cli_ulogoff attempted to be called with a fd of -1
Luke Leighton [Sun, 9 Apr 2000 02:04:31 +0000 (02:04 +0000)]
added pam_ntdom. amazingly, it actually works.
Elrond [Sat, 8 Apr 2000 15:09:05 +0000 (15:09 +0000)]
Removed the rpc_client-protos from proto.h
Seems to compile fine for me.
added LC_ALL=C, LC_COLLATE=C to mkproto.sh
Elrond [Fri, 7 Apr 2000 18:26:58 +0000 (18:26 +0000)]
using sambaconfdir for smb.conf and lmhosts.
Which other files in ${LIBDIR} are config-files?
Samba Release Account [Fri, 7 Apr 2000 08:27:50 +0000 (08:27 +0000)]
preparing for release of alpha.1.9
Tim Potter [Fri, 7 Apr 2000 07:41:01 +0000 (07:41 +0000)]
Replaced a whole bunch of 0x02000000's with SEC_RIGHTS_MAXIMUM_ALLOWED
Luke Leighton [Fri, 7 Apr 2000 06:50:16 +0000 (06:50 +0000)]
jf, looks like you put some extra error codes in...
Luke Leighton [Fri, 7 Apr 2000 06:33:10 +0000 (06:33 +0000)]
cut too much across.
Luke Leighton [Fri, 7 Apr 2000 05:48:55 +0000 (05:48 +0000)]
spoolss from cvs main
Luke Leighton [Fri, 7 Apr 2000 05:47:12 +0000 (05:47 +0000)]
open alias permissions were wrong, in msrpc_samr_query_aliasmem.
Luke Leighton [Fri, 7 Apr 2000 05:42:15 +0000 (05:42 +0000)]
domain name not obtained correctly.
Luke Leighton [Fri, 7 Apr 2000 05:34:21 +0000 (05:34 +0000)]
samaliasmem modified to take DOMAINNAME\name.
Tim Potter [Fri, 7 Apr 2000 02:18:04 +0000 (02:18 +0000)]
Same for cmd_sam_create_dom_alias.
Tim Potter [Fri, 7 Apr 2000 02:16:12 +0000 (02:16 +0000)]
Return from cmd_sam_create_dom_group() if not enough or too many arguments
passed.
Jeremy Allison [Thu, 6 Apr 2000 22:48:53 +0000 (22:48 +0000)]
Modified interfaces and added checks around *all* *alloc calls
so that errors are returned on memory allocation failure.
Jeremy.
Elrond [Thu, 6 Apr 2000 17:21:42 +0000 (17:21 +0000)]
Added --with-sambaconfdir, that currently defaults to
${LIBDIR} for backward-compatibility.
Hope to change this soon.
(@sambaconfdir@ not yet used)
Jean-François Micouleau [Thu, 6 Apr 2000 16:23:04 +0000 (16:23 +0000)]
filled the architecture table (useless for the moment).
fixed GetJob.
J.F.
Luke Leighton [Thu, 6 Apr 2000 07:54:43 +0000 (07:54 +0000)]
tdbsec_store takes the TDB_xxx flags arg.
Tim Potter [Thu, 6 Apr 2000 05:42:45 +0000 (05:42 +0000)]
General configuration file for DejaGnu.
Luke Leighton [Thu, 6 Apr 2000 02:36:57 +0000 (02:36 +0000)]
From johanh@fusion.kth.se Thu Apr 6 12:34:13 2000
Date: Wed, 5 Apr 2000 12:03:11 +0200 (MET DST)
From: Johan Hedin <johanh@fusion.kth.se>
To: samba-patches@samba.org
Cc: lkcl@samba.org
Subject: More Patch for configure.in in Samba TNG (PATCH#78)
I found another error I made. As long as my old config.cache was used, it
worked, but not with a fresh checkout... This patch will fix it. Again,
patch towards Samba TNG.
Tim Potter [Wed, 5 Apr 2000 22:44:47 +0000 (22:44 +0000)]
fsync returns int rather than void.
Tim Potter [Wed, 5 Apr 2000 22:44:33 +0000 (22:44 +0000)]
Improved clean target.
Tim Potter [Wed, 5 Apr 2000 22:44:24 +0000 (22:44 +0000)]
Added *.lo
Tim Potter [Wed, 5 Apr 2000 22:42:26 +0000 (22:42 +0000)]
fsync() returns an int rather than void.
Tim Potter [Wed, 5 Apr 2000 22:42:06 +0000 (22:42 +0000)]
Added *.so
Tim Potter [Wed, 5 Apr 2000 22:41:42 +0000 (22:41 +0000)]
Improved clean target.
Tim Potter [Wed, 5 Apr 2000 22:20:34 +0000 (22:20 +0000)]
Move across example VFS files from HEAD branch.
Jeremy Allison [Wed, 5 Apr 2000 19:52:57 +0000 (19:52 +0000)]
Paranioa purify fixes from HEAD.
Jeremy.
Jeremy Allison [Wed, 5 Apr 2000 19:48:06 +0000 (19:48 +0000)]
JF - you might want to look at this patch.
Went through and fixed up all failure possibilities. If something (malloc,
file open etc.) fails, then this code goes through and cleans up all allocated
resources on exit. This is the equivalent of a C++ destuctor for the allocated
structures. We need code to be written like this to pass purify tests.
Jeremy.
Elrond [Wed, 5 Apr 2000 18:02:05 +0000 (18:02 +0000)]
Another SMB_ASSERT in cli_pipe.c.
This time it got hit by an anonymous enumusers in
rpcclient.
Ohh, don't ask me, what's special about enumusers.
Jean-François Micouleau [Wed, 5 Apr 2000 17:53:45 +0000 (17:53 +0000)]
off by one error ...
J.F.
Jean-François Micouleau [Wed, 5 Apr 2000 10:05:32 +0000 (10:05 +0000)]
changed all the status code to ERROR_xxx instead of NT_STATUS_xx which are
wrong in the spoolss case.
fxed a bug in the job notify code (that's the polite answer), the truth is
different: there is a bug in the NT spooler service, including SP6a and
NT2K.
changed the default lpcommand in the LPRNG case.
J.F.
Luke Leighton [Wed, 5 Apr 2000 07:52:31 +0000 (07:52 +0000)]
patch from case for openssl
Luke Leighton [Wed, 5 Apr 2000 07:06:59 +0000 (07:06 +0000)]
case@seattle.gii.net spotted that msdfs.c needs nterr.h because of
NT_STATUS_PATH_NOT_COVERED.
Luke Leighton [Wed, 5 Apr 2000 07:02:12 +0000 (07:02 +0000)]
optimisation of TDB_MODIFY: use tdb_update() to tell us if the attempted
in-place replacement of the record didn't exist (and so failed), and if it
didn't exist, we don't need to check again (don't need to make _another_
call to tdb_exists(), just fail).