21 years agoadded a little description of useage and output format since this gives
Herb Lewis [Mon, 28 Aug 2000 18:58:46 +0000 (18:58 +0000)]
added a little description of useage and output format since this gives
more info than just what parameters are not supported by SWAT.

21 years agofixed typos spotted by Ron Alexander <rcalex@home.com>
Gerald Carter [Mon, 28 Aug 2000 13:41:46 +0000 (13:41 +0000)]
fixed typos spotted by Ron Alexander <rcalex@home.com>


21 years agoscript written by herb for reporting which smb.conf parameters
Gerald Carter [Mon, 28 Aug 2000 13:35:10 +0000 (13:35 +0000)]
script written by herb for reporting which smb.conf parameters
are in loarparm.c but not supported by SWAT.  I just thought it looked fun.


21 years agoMerge bug - still getting used to dirdiff.
Tim Potter [Mon, 28 Aug 2000 07:25:18 +0000 (07:25 +0000)]
Merge bug - still getting used to dirdiff.

21 years agoOops - missed a file.
Tim Potter [Mon, 28 Aug 2000 06:50:45 +0000 (06:50 +0000)]
Oops - missed a file.

21 years agoMerge from appliance branch.
Tim Potter [Mon, 28 Aug 2000 06:46:53 +0000 (06:46 +0000)]
Merge from appliance branch.

21 years agoyipee!
Andrew Tridgell [Mon, 28 Aug 2000 04:42:31 +0000 (04:42 +0000)]

The spoolss AddJob function has an [in,out] buffer not an [in] buffer
(despite the comment in the code to the contrary). Also, we must fail
this function - not just blindly reply "no problem" as AddJob should
always fail on non-local printers.

This fixes a bug where the "print test page" failed about half the
time. I suspect it will also fix a bunch of other intermittent spoolss
bugs where the client (incorrectly) tries to use the AddJob printing

21 years agomade reopen_logs() always re-open logs, not try and be smart about not
Andrew Tridgell [Mon, 28 Aug 2000 03:17:22 +0000 (03:17 +0000)]
made reopen_logs() always re-open logs, not try and be smart about not
re-opening in some circumstances. This fixes a problem where a HUP
does not re-open logs and leaves the log open on a unlinked file.

21 years agoarrgh.
Andrew Tridgell [Mon, 28 Aug 2000 02:06:25 +0000 (02:06 +0000)]

whoever added "auto" parameters in loadparam only supported "True", "False"
and "Auto", and did not support "Yes" and "No" like booleans. This
means anyone using "preferred master = yes" didn't get what they

21 years agovfs write data fix. Fix for name matching in addtosmbpass
Jeremy Allison [Fri, 25 Aug 2000 20:44:33 +0000 (20:44 +0000)]
vfs write data fix. Fix for name matching in addtosmbpass

21 years agoReplaced "\\MAILSLOT\\NET\\NETLOGON" with NET_LOGON_MAILSLOT constant.
Tim Potter [Fri, 25 Aug 2000 04:18:50 +0000 (04:18 +0000)]

21 years ago- fixed the %U macro so that the old (and documented) semantics work
Andrew Tridgell [Fri, 25 Aug 2000 03:58:02 +0000 (03:58 +0000)]
- fixed the %U macro so that the old (and documented) semantics work
again. This got broken with one of the substitute.c updates a couple
of months ago.

- also fixed %u to return the username from the current_user structure
  when called via a method that does not have direct access to the
  username. I cache the uidtoname() result to prevent thrashing nss.

21 years agoUpdated file with version in TNG.
Tim Potter [Fri, 25 Aug 2000 00:55:27 +0000 (00:55 +0000)]
Updated file with version in TNG.

21 years agoModified to use sid_to_uid and sid_to_gid. Now ready for POSIX ACL
Jeremy Allison [Thu, 24 Aug 2000 19:48:31 +0000 (19:48 +0000)]
Modified to use sid_to_uid and sid_to_gid. Now ready for POSIX ACL

21 years agoTidied up some error returns from printing calls. Still need to map UNIX
Jeremy Allison [Wed, 23 Aug 2000 23:05:49 +0000 (23:05 +0000)]
Tidied up some error returns from printing calls. Still need to map UNIX
errors to NT error for print job failure returns. Patch from John Reilly at HP.

21 years agoDocumentation on NT_USER_TOKEN. Requested by Sander.
Jeremy Allison [Wed, 23 Aug 2000 17:38:52 +0000 (17:38 +0000)]
Documentation on NT_USER_TOKEN. Requested by Sander.

21 years agoAdded code to do SID to uid/gid conversion. Needed for ACL support.
Jeremy Allison [Wed, 23 Aug 2000 00:45:40 +0000 (00:45 +0000)]
Added code to do SID to uid/gid conversion. Needed for ACL support.

21 years agoPatch from John Reilly <jreilly@hp.com> for not processing duplicate files
Tim Potter [Tue, 22 Aug 2000 06:28:33 +0000 (06:28 +0000)]
Patch from John Reilly <jreilly@hp.com> for not processing duplicate files
when moving printer driver to download area.

21 years agoFix for safe_free freeing static strings. From Robert August Vincent, II.
Jeremy Allison [Mon, 21 Aug 2000 19:30:19 +0000 (19:30 +0000)]
Fix for safe_free freeing static strings. From Robert August Vincent, II.

21 years agooops. must return "" string and length zero when strlen(filebuf) == 0
Luke Leighton [Fri, 18 Aug 2000 06:27:24 +0000 (06:27 +0000)]
oops.  must return "" string and length zero when strlen(filebuf) == 0

21 years agofixed an off by one error where I was tromping on memory
Gerald Carter [Fri, 18 Aug 2000 06:03:13 +0000 (06:03 +0000)]
fixed an off by one error where I was tromping on memory
when initializing a driver_info_3 struct.  See spoolss_addprinterdriver()


21 years agofixed initialization errors (again!) I must have forgot to
Gerald Carter [Fri, 18 Aug 2000 05:57:09 +0000 (05:57 +0000)]
fixed initialization errors (again!)  I must have forgot to
check these in before.


21 years agogetfileline() - line with length of zero -> filebuf[strlen(filebuf)-1]
Luke Leighton [Fri, 18 Aug 2000 05:56:19 +0000 (05:56 +0000)]
getfileline() - line with length of zero -> filebuf[strlen(filebuf)-1]
is NOT ok.

21 years agochanging some debug message levels
Gerald Carter [Fri, 18 Aug 2000 05:55:37 +0000 (05:55 +0000)]
changing some debug message levels


21 years agoFix for Win9x driver downloads when information is
Gerald Carter [Thu, 17 Aug 2000 20:05:05 +0000 (20:05 +0000)]
Fix for Win9x driver downloads when information is
stored in the TDB.  See the message I posted about
this on samba-technical

* If the admin does not desire the new NT print driver
support, nothing needs to be done.  All existing
parameters work the same.

* If the admin wants to take advantage of NT printer
driver support but does not want to migrate the 9x drivers
to the new setup, the leave the existing printers.def
file.  When smbd attempts to locate a 9x driver for the
printer in the TDB and fails it will drop down to
using the printers.def (and all associated parameters).
The make_printerdef tool will also remain for backwards
compatibility but will be moved to the "this tool is the
old way of doing it" pile.

* If the admin installs a 9x driver for a printer on
a Samba host (in the printing TDB), this information will
take precedence and the three old printing parameters
will be ignored (including print driver location).

* If an admin wants to migrate the printers.def file into
the new setup, I can write a script to do this using
rpcclient.  The main problem is the driver namespace
(we need the NT driver name first in order to
be consistent).


21 years agoadded comment for my own memory purposes
Gerald Carter [Thu, 17 Aug 2000 20:03:53 +0000 (20:03 +0000)]
added comment for my own memory purposes


21 years agohack for unmarshalling unistr's on machines with BIG_ENDIAN arch
Gerald Carter [Wed, 16 Aug 2000 03:44:04 +0000 (03:44 +0000)]
hack for unmarshalling unistr's on machines with BIG_ENDIAN arch
I think we have a lot more of these bugs lurking (i'm fairly
confident of it).


21 years ago- fixed memcpy bug in copy_unistr2()
Gerald Carter [Wed, 16 Aug 2000 03:41:02 +0000 (03:41 +0000)]
- fixed memcpy bug in copy_unistr2()
- init_unistr2_from_unistr() does not NULL terminate
  the buffer


21 years agofixed unitialized struct bug in enum_users (whatever it is called)
Gerald Carter [Wed, 16 Aug 2000 03:38:52 +0000 (03:38 +0000)]
fixed unitialized struct bug in enum_users (whatever it is called)


21 years agofixed typo.
Gerald Carter [Wed, 16 Aug 2000 03:34:58 +0000 (03:34 +0000)]
fixed typo.


21 years agofixed error in [public] spotted by Ron Alexander <rcalex@home.com>
Gerald Carter [Tue, 15 Aug 2000 23:36:05 +0000 (23:36 +0000)]
fixed error in [public] spotted by Ron Alexander <rcalex@home.com>
changed to read only = yes


21 years agoMerge of uninitialised variable fix from APPLIANCE_HEAD.
Tim Potter [Tue, 15 Aug 2000 21:09:54 +0000 (21:09 +0000)]
Merge ... uninitialised variable fix from APPLIANCE_HEAD.

21 years agoTidied up extra ()'s.
Jeremy Allison [Tue, 15 Aug 2000 02:33:27 +0000 (02:33 +0000)]
Tidied up extra ()'s.

21 years agoReverted the change Luke made. Removed IS_BITS_SET_XX macros.
Jeremy Allison [Mon, 14 Aug 2000 23:48:07 +0000 (23:48 +0000)]
Reverted the change Luke made. Removed IS_BITS_SET_XX macros.
Do not re-add them. These macros are unsafe as they are not understood.
Change all TNG code using them to correct '&' and '|' please.
IS_BITS_SET_ALL was being used in cmd_interp.c when IS_BITS_SET_SOME
should have been used.

21 years agoMerge.
Tim Potter [Mon, 14 Aug 2000 21:27:54 +0000 (21:27 +0000)]

21 years agofixed a few bugs in new_smb_io_relarraystr().
Gerald Carter [Mon, 14 Aug 2000 21:18:02 +0000 (21:18 +0000)]
fixed a few bugs in new_smb_io_relarraystr().
Marshalling code path now uses spoolss_smb_io_unistr()
also caught a NULL pointer case.

i'll update TNG later.

21 years agoNT printer permissions test program.
Tim Potter [Mon, 14 Aug 2000 18:39:51 +0000 (18:39 +0000)]
NT printer permissions test program.

21 years agorestored IS_BITS_SET_xxx macros.
Luke Leighton [Mon, 14 Aug 2000 11:03:30 +0000 (11:03 +0000)]
restored IS_BITS_SET_xxx macros.

21 years agohey, jerry: turns out you _have_ converted to get_safe_nt_errmsg!!!
Luke Leighton [Mon, 14 Aug 2000 07:37:12 +0000 (07:37 +0000)]
hey, jerry: turns out you _have_ converted to get_safe_nt_errmsg!!!
[got confused with which way dirdiff was going :)].

... but you forgot to update TNG's lib/cmd_interp.c...

21 years agojeremy,
Luke Leighton [Mon, 14 Aug 2000 07:33:34 +0000 (07:33 +0000)]

please do not remove IS_BITS_SET_XXX macros just because you happen
"not to like them".  whilst still working on merging, you are not
responsible for this code, therefore do not hinder its development
whilst people are attempting to do merges.  after the merge from
TNG to cvs main is completed, you can do what the hell you like.

thank you.


please could you consider adding underlying routines (in this case
get_safe_smb_errstr()) instead of reverting to a [less secure]
previous coding method.  thanks.

21 years agominor formatting fixes.
Gerald Carter [Mon, 14 Aug 2000 03:18:17 +0000 (03:18 +0000)]
minor formatting fixes.


21 years agofix for new_smb_io_relarraystr() and new_smb_io_relstr()
Gerald Carter [Mon, 14 Aug 2000 03:17:17 +0000 (03:17 +0000)]
fix for new_smb_io_relarraystr() and new_smb_io_relstr()
to use spool_smb_io_unistr() as this does not call prs_align()
befrore parsing the UNISTR.

Parsing a void* buffer from an RPC should not be aligned
on 4 byte boundaries.  Don't think this change affects
any marshalling code, only unmarshalling...


21 years agoremoved the prs_align in spoolss_smb_io_unistr() (JF, please check)
Gerald Carter [Sat, 12 Aug 2000 14:51:36 +0000 (14:51 +0000)]
removed the prs_align in spoolss_smb_io_unistr()  (JF, please check)
as it was throwing things off when parsing buffers passed in SPOOLSS

fixed the unmarshalling of UNISTR arrays to be double NULL terminated


21 years agofixes to prs_unistr UNMARSHALLING only. Problem was due to talloc's
Gerald Carter [Sat, 12 Aug 2000 14:31:29 +0000 (14:31 +0000)]
fixes to prs_unistr UNMARSHALLING only.  Problem was due to talloc's
of 0 bytes.


21 years agoFIxed some unmarshalling problems with relstr and unistr arrays
Gerald Carter [Sat, 12 Aug 2000 14:26:01 +0000 (14:26 +0000)]
FIxed some unmarshalling problems with relstr and unistr arrays


21 years agofixed unistr_to_ascii to deal with NULL src strings
Gerald Carter [Sat, 12 Aug 2000 14:20:40 +0000 (14:20 +0000)]
fixed unistr_to_ascii to deal with NULL src strings


21 years agoreworked talloc() to in order to help track down invalid
Gerald Carter [Sat, 12 Aug 2000 14:14:40 +0000 (14:14 +0000)]
reworked talloc() to in order to help track down invalid
memory access.

Tridge, I don't think using 0xdeadbeef for size allocations
of 0 is going to work.  I ended up having to use NULL
as much code that works on UNISTR checks to see if the buffer ptr
is NULL.  So valid code ends up with a seg fault.

Rather than rewriting it all, I added a DEBUG_TALLOC
#ifdef in talloc.h that sets a macro BAD_PTR.
This is the value assigned to ptr for an allocation of 0 bytes.


21 years agoremoved (void) typecast from SMB_ASSERT
Gerald Carter [Sat, 12 Aug 2000 14:10:27 +0000 (14:10 +0000)]
removed (void) typecast from SMB_ASSERT


21 years agoUpdated by Patrick Powell <papowell@lprng.org>
John Terpstra [Sat, 12 Aug 2000 02:41:42 +0000 (02:41 +0000)]
Updated by Patrick Powell <papowell@lprng.org>

21 years agoFixed the problem with UNISTR marshalling in a buffer5 struct.
Jeremy Allison [Fri, 11 Aug 2000 23:13:09 +0000 (23:13 +0000)]
Fixed the problem with UNISTR marshalling in a buffer5 struct.
The smb_io_unistr() code called a prs_align, this was not being
counted in the size or taken account of in the offset calculation.
Fixed size_ calculation to always return a size a multiple of 4
and also set the offset correctly. This fixes the problem I saw
and will hopefully fix the problem HP reported. JF please check
this change.

21 years agoFirst shot at actually *doing* WINS failover.
Christopher R. Hertel [Fri, 11 Aug 2000 22:29:44 +0000 (22:29 +0000)]
First shot at actually *doing* WINS failover.
If libsmb/namequery.c:name_query() times out while doing a non-broadcast
query, I mark that WINS server 'dead'.  Note that I don't try the new
WINS server.  I think I can get that working too.

This is only for queries, not registrations.  The biggest problem is that
I may have to fiddle with the UNICAST SUBNET, but I need to check talk
that over with someone (Jeremy?) before I hack at it.

I can't actually test the above change, I'm 'fraid.  I'm getting:

4 errors detected in the compilation of "rpc_server/srv_spoolss_nt.c".

in head branch.

Chris -)-----

21 years agoFix case where volumename could be returned as UNIX charset or DOS codepage
Jeremy Allison [Fri, 11 Aug 2000 00:53:13 +0000 (00:53 +0000)]
Fix case where volumename could be returned as UNIX charset or DOS codepage
depending on where it came from.

21 years agoAdded "add_printer_hook" call to update printer for Win9x clients.
Jeremy Allison [Thu, 10 Aug 2000 23:41:16 +0000 (23:41 +0000)]
Added "add_printer_hook" call to update printer for Win9x clients.
Patch from John Reilly <jreilly@hp.com>.

21 years agoTidied up security rights definitions.
Jeremy Allison [Thu, 10 Aug 2000 19:51:45 +0000 (19:51 +0000)]
Tidied up security rights definitions.

21 years agoRemoved requirement that sid have an owner before being interpreted.
Jeremy Allison [Thu, 10 Aug 2000 17:48:15 +0000 (17:48 +0000)]
Removed requirement that sid have an owner before being interpreted.
Thanks to Elrond for pointing this out.

21 years agoWorking on spoolss_getprinterdriver() as it does not display all the
Gerald Carter [Thu, 10 Aug 2000 14:05:19 +0000 (14:05 +0000)]
Working on spoolss_getprinterdriver() as it does not display all the


21 years agoworking on the unmarshalling of dependentfiles to a DRIVER_INFO_3
Gerald Carter [Thu, 10 Aug 2000 14:03:13 +0000 (14:03 +0000)]
working on the unmarshalling of dependentfiles to a DRIVER_INFO_3
struct.  Not wuite there.


21 years agocode formatting cleanup
Gerald Carter [Thu, 10 Aug 2000 14:02:07 +0000 (14:02 +0000)]
code formatting cleanup


21 years agodeal with allocation size of 0 in prs_unistr when UNMARSHALLING
Gerald Carter [Thu, 10 Aug 2000 14:00:40 +0000 (14:00 +0000)]
deal with allocation size of 0 in prs_unistr when UNMARSHALLING


21 years agoDeal will NULL UNISTR in unistr_to_ascii
Gerald Carter [Thu, 10 Aug 2000 14:00:05 +0000 (14:00 +0000)]
Deal will NULL UNISTR in unistr_to_ascii


21 years agotalloc returns 0xdeadbeef when asked to allocate 0 bytes
Gerald Carter [Thu, 10 Aug 2000 13:58:15 +0000 (13:58 +0000)]
talloc returns 0xdeadbeef when asked to allocate 0 bytes


21 years agoNeeded to add the SHARED attribute bit when creating a printer
Gerald Carter [Thu, 10 Aug 2000 04:10:29 +0000 (04:10 +0000)]
Needed to add the SHARED attribute bit when creating a printer
on a Samba host.  Also needed to add an option to pass the share
name (printer name) on the command line.

And fixed the checking of the return code for spoolss_r_addprinterex()


21 years agoFixed bug in new printer access check code. If PRINTER_ALL_ACCESS is changed
Jeremy Allison [Wed, 9 Aug 2000 23:19:50 +0000 (23:19 +0000)]
Fixed bug in new printer access check code. If PRINTER_ALL_ACCESS is changed
in the ACE's the same thing must be done in the desired access or they will
never match. This took me a while to find... :-).

21 years agomake proto
Gerald Carter [Wed, 9 Aug 2000 20:19:37 +0000 (20:19 +0000)]
make proto

21 years agoMore work on AddPrinterDriver() and AddPrinterEx() client RPC's
Gerald Carter [Wed, 9 Aug 2000 20:14:29 +0000 (20:14 +0000)]
More work on AddPrinterDriver() and AddPrinterEx() client RPC's

Also fixed init_unistr() to deal with a NULL source character string.


21 years agoFixed memory leak with NT tokens.
Jeremy Allison [Wed, 9 Aug 2000 18:40:48 +0000 (18:40 +0000)]
Fixed memory leak with NT tokens.
Added debug messages to se_access_check().
Added FULL_ACCESS acl to default acl on printers.

21 years agogot error code right for printer update/add failure
Andrew Tridgell [Wed, 9 Aug 2000 07:34:35 +0000 (07:34 +0000)]
got error code right for printer update/add failure

21 years agoadd printer admin docs
Andrew Tridgell [Wed, 9 Aug 2000 04:25:01 +0000 (04:25 +0000)]
add printer admin docs

21 years agoadded printer admin option
Andrew Tridgell [Wed, 9 Aug 2000 04:19:18 +0000 (04:19 +0000)]
added printer admin option
any user in that list can do anything to a printer

21 years agoFixed AddPrinterDriver()
Gerald Carter [Wed, 9 Aug 2000 02:04:25 +0000 (02:04 +0000)]
Fixed AddPrinterDriver()


21 years agofixed help string
Andrew Tridgell [Wed, 9 Aug 2000 00:11:01 +0000 (00:11 +0000)]
fixed help string

21 years agoFound the sec_ctx_stack overflow - a become_root() should have been an
Jeremy Allison [Tue, 8 Aug 2000 23:49:26 +0000 (23:49 +0000)]
Found the sec_ctx_stack overflow - a become_root() should have been an
unbecome_root() - typo.

21 years agoAdded SID "Everyone" S-1-1-0 as always matching if present in an ACE.
Jeremy Allison [Tue, 8 Aug 2000 21:51:22 +0000 (21:51 +0000)]
Added SID "Everyone" S-1-1-0 as always matching if present in an ACE.

21 years agoChanged the sec desc access checks to match the spec. Needs testing.
Jeremy Allison [Tue, 8 Aug 2000 19:34:34 +0000 (19:34 +0000)]
Changed the sec desc access checks to match the spec. Needs testing.

21 years agomake proto
Gerald Carter [Tue, 8 Aug 2000 07:23:20 +0000 (07:23 +0000)]
make proto

21 years agocleanup in init_unistr2_from_unistr()
Gerald Carter [Tue, 8 Aug 2000 06:59:35 +0000 (06:59 +0000)]
cleanup in init_unistr2_from_unistr()


21 years agoAll changes related to rpcclient...
Gerald Carter [Tue, 8 Aug 2000 06:57:48 +0000 (06:57 +0000)]
All changes related to rpcclient...

- cleaned up some code
- Fixed a few memory leaks of my own making
- Add AddPrinterDriver(); I'm missing some of the semantics
  here as the call is done correctly, but I'm not getting all
  the information right in the DRIVER_INFO_3 struct I think.
  Will work on it tomorrow some more...


21 years agoit is not my day it seems. :-(
Gerald Carter [Sat, 5 Aug 2000 19:18:25 +0000 (19:18 +0000)]
it is not my day it seems.  :-(
Fixed missing )


21 years agoremoved the for() loop to copy the buffer in init_unistr2_from_unistr()
Gerald Carter [Sat, 5 Aug 2000 19:03:00 +0000 (19:03 +0000)]
removed the for() loop to copy the buffer in init_unistr2_from_unistr()
Replaced with a memcpy()

Forgot to commit this a few moments ago


21 years agoFixed bug in init_unistr2_from_unistr() found by Elrond.
Gerald Carter [Sat, 5 Aug 2000 18:58:45 +0000 (18:58 +0000)]
Fixed bug in init_unistr2_from_unistr() found by Elrond.
Thanks :-)


21 years agoAdding build file.
John Terpstra [Fri, 4 Aug 2000 21:03:10 +0000 (21:03 +0000)]
Adding build file.

21 years agoAdding build system files for Caldera OpenLinux.
John Terpstra [Fri, 4 Aug 2000 20:53:28 +0000 (20:53 +0000)]
Adding build system files for Caldera OpenLinux.

21 years agomake proto
Gerald Carter [Fri, 4 Aug 2000 20:47:06 +0000 (20:47 +0000)]
make proto

21 years agospoolss_addprinterex() was adding the printer and returning the
Gerald Carter [Fri, 4 Aug 2000 20:32:34 +0000 (20:32 +0000)]
spoolss_addprinterex() was adding the printer and returning the
correct handle, but was deleting the connection to the server.


21 years agoFixed up se_access_check() to use the token list from the user struct
Jeremy Allison [Fri, 4 Aug 2000 19:56:58 +0000 (19:56 +0000)]
Fixed up se_access_check() to use the token list from the user struct
as the SID list. Now to go through and tidy up the algorithm.

21 years agoAdded my course notes. Had asked Jeremy some weeks ago.
Volker Lendecke [Fri, 4 Aug 2000 16:08:04 +0000 (16:08 +0000)]
Added my course notes. Had asked Jeremy some weeks ago.
It's german language, feel free to remove it again.


21 years agomake proto
Gerald Carter [Fri, 4 Aug 2000 13:30:52 +0000 (13:30 +0000)]
make proto

21 years agoclunky support for calling AddPrinterEx(). The code currently reports
Gerald Carter [Fri, 4 Aug 2000 12:46:36 +0000 (12:46 +0000)]
clunky support for calling AddPrinterEx().  The code currently reports
that the call failed, but the printer shows up on the remote NT client.
(note this is the client side call).  I've botched the return value
somewhere and will fix that today.


21 years agoAfter talking with Jeremy and JF (and staring at packet traces between
Gerald Carter [Fri, 4 Aug 2000 12:42:19 +0000 (12:42 +0000)]
After talking with Jeremy and JF (and staring at packet traces between
NT <-> NT), I've come to realize that UNISTR2 strings should be NULL


21 years agoIn cli_error() return ENOENT when an ERROR_INVALID_NAME is received instead
Tim Potter [Fri, 4 Aug 2000 06:13:05 +0000 (06:13 +0000)]
In cli_error() return ENOENT when an ERROR_INVALID_NAME is received instead
of the generic EINVAL.

21 years agoFixed up the user/group contexts when using authenticated pipes.
Jeremy Allison [Fri, 4 Aug 2000 00:59:09 +0000 (00:59 +0000)]
Fixed up the user/group contexts when using authenticated pipes.
Added a become_root()/unbecome_root() (push/pop security context)
around the initgroups() call to ensure it would succeed. Hmmm - I
wonder if this call being done as non-root might explain any "group access"
bugs we've had in the past....

21 years agoAdded an NT_USER_TOKEN structure that is copied/passed around associated
Jeremy Allison [Thu, 3 Aug 2000 22:38:43 +0000 (22:38 +0000)]
Added an NT_USER_TOKEN structure that is copied/passed around associated
with the current user. This will allow se_access_check() to quickly do
a SD check without having to translate uid/gid's to SIDs.
Still needs work on pipe calls.

21 years agoMuch though I hate to admit it - JF was completely correct. I cannot
Jeremy Allison [Wed, 2 Aug 2000 18:49:36 +0000 (18:49 +0000)]
Much though I hate to admit it - JF was completely correct. I cannot
now reproduce the bug I had with adding a printer driver, and PrintMig.exe
crashes if I use my SD code but works with his.
I stand completely corrected :-). - So I'm reverting to his code as it works :-).

21 years agoStarted to canonicalize our handling of uid -> sid code in order to
Jeremy Allison [Wed, 2 Aug 2000 02:11:55 +0000 (02:11 +0000)]
Started to canonicalize our handling of uid -> sid code in order to
get ready and fix se_access_check().
Added cannonical lookup_name(), lookup_sid(), uid_to_sid(), gid_to_sid()
functions that look via winbind first the fall back on local lookup.

All Samba should use these rather than trying to call winbindd code

Added NT_USER_TOKEN struct in user_struct, contains list of NT sids
associated with this user.

se_access_check() should use this (cached) value rather than attempting
to do the same thing itself when given a uid/gid pair.

More work needs to be done to preserve these things accross security
context changes (especially with the tricky pipe problem) but I'm
beginning to see how this will be done..... probably by registering
a new vuid for an authenticated RPC pipe and not treating the
pipe calls specially.

More thoughts needed - but we're almost there...


21 years agoTidyup removing many of the 0xC0000000 | NT_STATUS_XXX stuff (only need NT_STATUS_XXX).
Jeremy Allison [Tue, 1 Aug 2000 18:32:34 +0000 (18:32 +0000)]
Tidyup removing many of the 0xC0000000 | NT_STATUS_XXX stuff (only need NT_STATUS_XXX).
Removed IS_BITS_xxx macros as they were just reproducing "C" syntax in a more
obscure way.

21 years agoAdded print job substitutions for %{printername}, %{sharename} and %{portname}
Jeremy Allison [Tue, 1 Aug 2000 00:41:19 +0000 (00:41 +0000)]
Added print job substitutions for %{printername}, %{sharename} and %{portname}
from the NT printer tdb.
Also added checks for time restrictions before allowing a job to print.

21 years agoSave & restore the port name correctly.
Jeremy Allison [Mon, 31 Jul 2000 21:41:03 +0000 (21:41 +0000)]
Save & restore the port name correctly.

21 years agoAdded John Reilly's enumports/addprinter/delprinter scripting code plus the
Jeremy Allison [Mon, 31 Jul 2000 20:41:51 +0000 (20:41 +0000)]
Added John Reilly's enumports/addprinter/delprinter scripting code plus the
fix for the Win9x printer drivers.
Changed command names to add "command" string on the end for some consistancy
with the other scripting commands.
Added '%P' option to tdbpack/unpack to store long comment string.
Made port name be "Samba Printer Port" if no enum port script given.
Fixed prs_uint32_pre code to cope with null args.

21 years agomake proto
Gerald Carter [Mon, 31 Jul 2000 15:17:29 +0000 (15:17 +0000)]
make proto

21 years agoMore work on rpcclient...
Gerald Carter [Mon, 31 Jul 2000 14:56:01 +0000 (14:56 +0000)]
More work on rpcclient...

  * Fixed to work with Jeremy's recent changes re: dunamic
    memory allocation when unmarshalling unistr[2]
  * included EnumPorts level 1
  * more work on AddPrinterEx