samba.git
25 years agoAdded the APPLEDOUBLE macro for the name of the directory containing
John Blair [Thu, 13 Aug 1998 18:56:09 +0000 (18:56 +0000)]
Added the APPLEDOUBLE macro for the name of the directory containing
a file's Macintosh resource fork (normally .AppleDouble under Netatalk).

25 years agoUh... I'm checking stuff in. Jeremy will by *so* pleased :)
John Blair [Thu, 13 Aug 1998 18:46:35 +0000 (18:46 +0000)]
Uh... I'm checking stuff in.  Jeremy will by *so* pleased :)

netatalk.c contains the meat of the code required to move Macintosh resource
forks around when Windows users manipulate files created by a mac.  This
allows for greater integration between netatalk and Samba, but is, of course,
potentially a recipe for disaster.  All the netatalk related code is insulated
by #ifdefs from the standard build.

25 years agomove smbd/print_svid.o from SMBD_OBJ1 to PARAM_OBJ so that nmbd, swat, etc.
Herb Lewis [Thu, 13 Aug 1998 17:38:26 +0000 (17:38 +0000)]
move smbd/print_svid.o from SMBD_OBJ1 to PARAM_OBJ so that nmbd, swat, etc.
will link when SYSV is defined. We may want to restructure this later and
move this to param instead of smbd ???

25 years agoreturn to old behaviour of dumping ALL parameters
Herb Lewis [Thu, 13 Aug 1998 16:32:20 +0000 (16:32 +0000)]
return to old behaviour of dumping ALL parameters

25 years agoHAVE_STATVFS should be STAT_STAVFS (caused printing to break on IRIX).
Jeremy Allison [Thu, 13 Aug 1998 16:20:37 +0000 (16:20 +0000)]
HAVE_STATVFS should be STAT_STAVFS (caused printing to break on IRIX).
Jeremy.

25 years agoMore blocking lock code - still #ifdef'ed out - under development.
Jeremy Allison [Thu, 13 Aug 1998 02:10:45 +0000 (02:10 +0000)]
More blocking lock code - still #ifdef'ed out - under development.
Jeremy.

25 years agoDebug output formatting change. I'm trying to combine log lines that
Christopher R. Hertel [Thu, 13 Aug 1998 00:01:02 +0000 (00:01 +0000)]
Debug output formatting change.  I'm trying to combine log lines that
really should be together.  Chris -)-----

25 years agoChanged a call to fflush(dbg) to dbgflush().
Christopher R. Hertel [Wed, 12 Aug 1998 15:27:13 +0000 (15:27 +0000)]
Changed a call to fflush(dbg) to dbgflush().

25 years agoI've added a dbgflush() function to debug.c. Calling this will cause the
Christopher R. Hertel [Wed, 12 Aug 1998 14:51:17 +0000 (14:51 +0000)]
I've added a dbgflush() function to debug.c.  Calling this will cause the
debug format buffer to be written out (and reset).  fflush() is also called
to force the issue.  I replaced the call to fflush() in client.c with a
call to dbgflush(), which seems to have fixed the problem that Andrew was
working on (i.e., that the prompt was not displayed when using smbclient).

Chris -)-----

25 years agofixed the nested comment - Jeremy, do you want that unbecome_user() or
Andrew Tridgell [Wed, 12 Aug 1998 10:09:41 +0000 (10:09 +0000)]
fixed the nested comment - Jeremy, do you want that unbecome_user() or
not?

25 years agoinclude/smb.h: Removed dir_ptr. Not needed.
Jeremy Allison [Tue, 11 Aug 1998 23:28:35 +0000 (23:28 +0000)]
include/smb.h: Removed dir_ptr. Not needed.
locking/locking.c: First cut a blocking lock code. #ifdef'ed out for now.
locking/locking_shm.c: Removed dir_ptr. Not needed.
smbd/nttrans.c: More work on ChangeNotify - return is not an error and needs
                to be handled as a nttrans with zero params. Removed dir_ptr. Not needed.
smbd/reply.c:
smbd/server.c:
smbd/trans2.c: Removed dir_ptr. Not needed.

Hmmm. At the moment smbclient is broken - doesn't issue prompt correctly.
This needs looking at.

Jeremy.

25 years agoQuick fix for a small problem. If you run 'nmbd -?' you'd get the usage
Christopher R. Hertel [Tue, 11 Aug 1998 21:25:55 +0000 (21:25 +0000)]
Quick fix for a small problem.  If you run 'nmbd -?' you'd get the usage
message *but the daemon would start anyway*.  I've added a call to exit()
in the default: case of the option processing loop.

Chris -)-----

25 years agoI've come up with a different scheme for doing the DEBUG() and DEBUGADD()
Christopher R. Hertel [Tue, 11 Aug 1998 18:56:03 +0000 (18:56 +0000)]
I've come up with a different scheme for doing the DEBUG() and DEBUGADD()
macros that meets the RVALUE requirement and doesn't use the (a?b:c) format
that Andrew called "ugly".  I've added the new macros but kept the old
macros within a #if 0..#else..#endif block in case I've missed somthing.

Basically, I've used

  (void)( (a) && (b) )

or

  (void)( (a) && (b) && (c) )

instead of

  ( (a) ? (void)(b) : (void)(c) )

or similar.  I have this compiled and running now.

Here's the diff:

$ cvs diff smb.h
Enter passphrase for RSA key 'crh@Ruby':
Index: smb.h
===================================================================
RCS file: /data/cvs/samba/source/include/smb.h,v
retrieving revision 1.172
diff -r1.172 smb.h
143a144,145
> #if 0
>
150a153,164
>
> #else
>
> #define DEBUG( level, body ) \
>   (void)( (DEBUGLEVEL >= (level)) \
>        && (dbghdr( level, FILE_MACRO, FUNCTION_MACRO, (__LINE__) )) \
>        && (dbgtext body) )
>
> #define DEBUGADD( level, body ) \
>   (void)( (DEBUGLEVEL >= (level)) && (dbgtext body) )
>
> #endif

Chris -)-----

25 years agoCleaned up some lint warnings (functions with return values that were
Christopher R. Hertel [Tue, 11 Aug 1998 18:07:18 +0000 (18:07 +0000)]
Cleaned up some lint warnings (functions with return values that were
ignored).

25 years agostrip out extra spaces from smbclient output that were added recently
Herb Lewis [Tue, 11 Aug 1998 17:34:00 +0000 (17:34 +0000)]
strip out extra spaces from smbclient output that were added recently

25 years agoRemoved a stray semicolon in the DEBUGADD() macro.
Christopher R. Hertel [Tue, 11 Aug 1998 15:54:15 +0000 (15:54 +0000)]
Removed a stray semicolon in the DEBUGADD() macro.

25 years agoI've added a test for lp_loaded() so that timestamps *will* be output
Christopher R. Hertel [Tue, 11 Aug 1998 15:47:26 +0000 (15:47 +0000)]
I've added a test for lp_loaded() so that timestamps *will* be output
before the config file has been loaded.  Otherwise the default is no
timestamp.

Chris -)-----

25 years agofixed a nasty bug in debug.c
Andrew Tridgell [Tue, 11 Aug 1998 14:09:05 +0000 (14:09 +0000)]
fixed a nasty bug in debug.c

Debug1() was being called like this:

      Debug1( format_bufr );

but if format_bufr contains any %s or other % arguments (such as when
processing a smb.conf file containing % macros) then smbd dies a
horrible death.

The quick fix is to use:

Debug1( "%s", format_bufr);

25 years agomake sure that a fault cannot occur twice. The new Debug1() code is
Andrew Tridgell [Tue, 11 Aug 1998 13:55:16 +0000 (13:55 +0000)]
make sure that a fault cannot occur twice. The new Debug1() code is
causing seg faults on my machine and because the fault code calls
DEBUG() it is looping!

now to find the bug in the Debug1() code ...

25 years agoMakefile.in: Added CHECK target back in (at Herb's request). Added
Jeremy Allison [Tue, 11 Aug 1998 02:13:01 +0000 (02:13 +0000)]
Makefile.in: Added CHECK target back in (at Herb's request). Added
             manpath in (although we don't currently use it).
client/client.c: Added John Blair's fixes for "put -".
include/nterr.h: Added NT_STATUS_NOTIFY_ENUM_DIR error code.
smbd/nttrans.c: Added in devious fix for one-shot NT change notify bug.
Jeremy.

25 years agoupdate for new directory structure
Herb Lewis [Mon, 10 Aug 1998 22:36:11 +0000 (22:36 +0000)]
update for new directory structure

25 years agoMakefile.in: Fixed make clean bug with CVS directory in bin/.
Jeremy Allison [Mon, 10 Aug 1998 20:12:53 +0000 (20:12 +0000)]
Makefile.in: Fixed make clean bug with CVS directory in bin/.
include/smb.h: Fixed bugs in Debug macro's with SGI compiler.
smbd/trans2.c: Fixed bug reported by Zoltan Palmai <ZSPA@chevron.com>
               PR#8903 where get_lanman2_dir_entry() could return a
               mangled name as a resume key and call_trans2findnext()
               wasn't doing the same.
Jeremy.

25 years agomoved access.c into lib/ from smbd/ as it is needed by testparm. I
Andrew Tridgell [Mon, 10 Aug 1998 07:29:57 +0000 (07:29 +0000)]
moved access.c into lib/ from smbd/ as it is needed by testparm. I
modified it not to need any smbd specific structures and instead pass
things from smbd.

25 years agosplit the system password checking routines out of smbd/password.c and
Andrew Tridgell [Mon, 10 Aug 1998 07:04:53 +0000 (07:04 +0000)]
split the system password checking routines out of smbd/password.c and
into passdb/pass_check.c. This means SWAT no longer needs to link to
smbd/password.c

25 years agomoved username.c into lib/ so we no longer $(PASSDB_OBJ) in most
Andrew Tridgell [Mon, 10 Aug 1998 05:22:10 +0000 (05:22 +0000)]
moved username.c into lib/ so we no longer $(PASSDB_OBJ) in most
utilities

25 years agosmbtorture now compiles
Andrew Tridgell [Mon, 10 Aug 1998 02:01:18 +0000 (02:01 +0000)]
smbtorture now compiles

25 years agofixed the installation scripts for the new layout
Andrew Tridgell [Mon, 10 Aug 1998 01:55:42 +0000 (01:55 +0000)]
fixed the installation scripts for the new layout

25 years agothe rest of the binaries now compile and link
Andrew Tridgell [Mon, 10 Aug 1998 01:25:32 +0000 (01:25 +0000)]
the rest of the binaries now compile and link

25 years agoa few more things compile.
Andrew Tridgell [Sun, 9 Aug 1998 14:58:51 +0000 (14:58 +0000)]
a few more things compile.

I'll do the rest tomorrow.

25 years agoswat compiles, with some even uglier hacks. we really have to do
Andrew Tridgell [Sun, 9 Aug 1998 14:37:32 +0000 (14:37 +0000)]
swat compiles, with some even uglier hacks. we really have to do
something about our dependencies between seemingly unrelated code.

25 years agonmbd compiles and links. I had to do some ugly stuff, putting files in
Andrew Tridgell [Sun, 9 Aug 1998 14:20:02 +0000 (14:20 +0000)]
nmbd compiles and links. I had to do some ugly stuff, putting files in
illogical places. Once I get everything compiled these will need to be
revisited, but at least the new Makefile.in makes it really obvious
what is out of place.

25 years agoclose to having nmbd linking (just a bit of repository hacking to go)
Andrew Tridgell [Sun, 9 Aug 1998 14:08:42 +0000 (14:08 +0000)]
close to having nmbd linking (just a bit of repository hacking to go)

25 years agothis gets smbd compiling and linking correctly with the new layout.
Andrew Tridgell [Sun, 9 Aug 1998 13:25:34 +0000 (13:25 +0000)]
this gets smbd compiling and linking correctly with the new layout.

25 years agoremoved the if statements from the DEBUG() macro definitions.
Andrew Tridgell [Sun, 9 Aug 1998 13:12:09 +0000 (13:12 +0000)]
removed the if statements from the DEBUG() macro definitions.

Chris, you should never put if statements in macros, use the ugly
(a?b:c) form instead, otherwise you can produce incorrect code when
you have things like:

if (foo)
   DEBUG((blah));
else
   DEBUG((blooh));

25 years agoanother dummy file
Andrew Tridgell [Sun, 9 Aug 1998 12:09:48 +0000 (12:09 +0000)]
another dummy file

25 years agoadded ignore rules for the dummy files
Andrew Tridgell [Sun, 9 Aug 1998 11:25:49 +0000 (11:25 +0000)]
added ignore rules for the dummy files

25 years agothese dummy files are needed for autoconf processing
Andrew Tridgell [Sun, 9 Aug 1998 11:24:15 +0000 (11:24 +0000)]
these dummy files are needed for autoconf processing

25 years agothe autoconf scripts are now converted to the new directory structure
Andrew Tridgell [Sun, 9 Aug 1998 11:23:13 +0000 (11:23 +0000)]
the autoconf scripts are now converted to the new directory structure

now the Makefile .....

25 years agonew directory structure in configure.in
Andrew Tridgell [Sun, 9 Aug 1998 11:14:23 +0000 (11:14 +0000)]
new directory structure in configure.in

25 years agowe have to have the test for -lsocket before the test for some
Andrew Tridgell [Sat, 8 Aug 1998 05:25:04 +0000 (05:25 +0000)]
we have to have the test for -lsocket before the test for some
functions (notably innetgr) to ensure we don't replace them if they
are in -lsocket

25 years agoadded --with-nisplus-home option
Andrew Tridgell [Sat, 8 Aug 1998 01:15:14 +0000 (01:15 +0000)]
added --with-nisplus-home option

25 years agoforgot to remove -nostrip option for packaging. Makes inst package too big.
Herb Lewis [Fri, 7 Aug 1998 19:01:59 +0000 (19:01 +0000)]
forgot to remove -nostrip option for packaging. Makes inst package too big.

25 years agoadd startswat.sh to automate starting swat with inetd
Herb Lewis [Fri, 7 Aug 1998 18:56:12 +0000 (18:56 +0000)]
add startswat.sh to automate starting swat with inetd
update make scripts to use new autoconfigure setup

25 years agoFixed memory leak when freeing ChangeNotify structures.
Jeremy Allison [Wed, 5 Aug 1998 22:17:51 +0000 (22:17 +0000)]
Fixed memory leak when freeing ChangeNotify structures.
Jeremy.

25 years agonttrans.c: Fix change_notify. Queue processing code wasn't changing to
Jeremy Allison [Tue, 4 Aug 1998 23:45:34 +0000 (23:45 +0000)]
nttrans.c: Fix change_notify. Queue processing code wasn't changing to
           the correct directory before doing the stat. Doh !
uid.c: Fix for performance in security=share mode. Invalid vuid meant
       that in security=share mode Samba was always doing the become_user/
       undebome_user pairs for *every* smb. This code fixes it, but tridge
       should review for security implications.
Jeremy.

25 years agoloadparm.c: Added "debug timestamp" synonym for Chris.
Jeremy Allison [Tue, 4 Aug 1998 01:01:26 +0000 (01:01 +0000)]
loadparm.c: Added "debug timestamp" synonym for Chris.
nttrans.c: Moved common code into function in server.c
           Removed left over debug level 0.
server.c: Moved common code into function in server.c
Jeremy.

25 years agoMore formatting changes. Mostly converted some DEBUG() calls to DEBUGADD()
Christopher R. Hertel [Mon, 3 Aug 1998 22:10:53 +0000 (22:10 +0000)]
More formatting changes.  Mostly converted some DEBUG() calls to DEBUGADD()
so that we wouldn't get too many timestamps.

Chris -)-----

25 years agoFirst implementation of ChangeNotify - this version only checks
Jeremy Allison [Mon, 3 Aug 1998 19:07:55 +0000 (19:07 +0000)]
First implementation of ChangeNotify - this version only checks
for changes in the directory modify timestamps. A better version
will look at the requested client flags, and create a hash that
represents the current state of the directory, and check against
this instead.
debug.c: Added lp_timestamp_logs() function.
loadparm.c: Added "change notify timeout" in seconds (default 60)
- this is the scan rate for a directory.
            Added ""timestamp logs" boolean - default True. Turns
off log timestamps (so I can read them :-).
nttrans.c: ChangeNotify implementation.
server.c: ChangeNotify implementation.
shmem_sysv.c: Added exits on shmem errors (without them smbd can
              core dump if some calls fail).
smb.h: Added ChangeNotify flags for future use.
util.c: Tidied up typedef.
Jeremy.

25 years agoOne more minor change to the format of a DEBUG message. I broke up a
Christopher R. Hertel [Mon, 3 Aug 1998 18:33:49 +0000 (18:33 +0000)]
One more minor change to the format of a DEBUG message.  I broke up a
very long output line in become_domain_master_query_success().
Chris -)-----

25 years agoI finished removing timestring() calls from DEBUG() messages. Also went
Christopher R. Hertel [Mon, 3 Aug 1998 18:13:13 +0000 (18:13 +0000)]
I finished removing timestring() calls from DEBUG() messages.  Also went
through and changed some DEBUG() calls to DEBUGADD() to combine output
under a single timestamp.  There were too many timestamps.

Note that Jeremy has told me that he's working on adding a config parameter
to turn timestamps off.  Cool.

Chris -)-----

25 years agoFixing clitar.c so that tar to stdout works correctly.
Richard Sharpe [Mon, 3 Aug 1998 10:35:00 +0000 (10:35 +0000)]
Fixing clitar.c so that tar to stdout works correctly.

Replaced printfs with DEBUG(0. Tested ... Works.

Hope I didn't disturb the autoconf code ... It feels
great to be able to run configure for Samba at long last!

Regards
Richard Sharpe

25 years agoThis is the remaining set of changes needed to replace the changes lost
Christopher R. Hertel [Mon, 3 Aug 1998 03:50:07 +0000 (03:50 +0000)]
This is the remaining set of changes needed to replace the changes lost
when Andrew and I were both working with util.c.  I really don't know
how I lost the autoconfigure changes (honest, I *did* run frequent
updates).

Chris -)-----

25 years agoI have fixed some of the autoconfigure problems. I'm studying the diffs
Christopher R. Hertel [Mon, 3 Aug 1998 03:22:42 +0000 (03:22 +0000)]
I have fixed some of the autoconfigure problems.  I'm studying the diffs
for the rest.  I've found that only debug.h seems to be out of sync (i.e.,
util.c itself appears to be okay).

Chris -)-----

25 years agoAs per a Andrew's message, I went through and removed the timestring()
Christopher R. Hertel [Fri, 31 Jul 1998 22:39:15 +0000 (22:39 +0000)]
As per a Andrew's message, I went through and removed the timestring()
timestamps from several DEBUG messages.  The timestamps are redundant now
that DEBUG() provides them automatically.

There are still a few more files to do, but I've got to get home for dinner.

Chris -)-----

25 years agoDebugging functions are now in their own module.
Christopher R. Hertel [Fri, 31 Jul 1998 20:17:36 +0000 (20:17 +0000)]
Debugging functions are now in their own module.

Chris -)-----

25 years agoThis is the checkin of the debug changes.
Christopher R. Hertel [Fri, 31 Jul 1998 20:16:35 +0000 (20:16 +0000)]
This is the checkin of the debug changes.

  Makefile.in:  I've added debug.o.
  proto.h    :  Rebuilt, as is standard for these sorts of things.
  smb.h      :  New macros, etc.
  util.c     :  Debug code removed.

I'll check in debug.c in the next step.

Chris -)-----

25 years agoadded test for getpwanam().
Andrew Tridgell [Fri, 31 Jul 1998 03:33:25 +0000 (03:33 +0000)]
added test for getpwanam().

25 years agoMakefile.in: Moved UBIQX stuff into UTILOBJ.
Jeremy Allison [Thu, 30 Jul 1998 21:18:57 +0000 (21:18 +0000)]
Makefile.in: Moved UBIQX stuff into UTILOBJ.
loadparm.c: Added "ole locking compatibility" option (default "true").
locking.c: Changes to implement union in files_struct.
locking_shm.c: Changes to implement union in files_struct.
nttrans.c: Made opening a directory explicit (we have to).
           Added create directory code for nttrans.
reply.c: Changes to implement union in files_struct.
server.c: Changes to implement union in files_struct. Added create directory code.
trans2.c: Changes to implement union in files_struct.
smb.h: Changes to implement union in files_struct.
util.c: Changed linked list code to UNIQX linked list. This will make
        the other lists I need to implement for ChangeNotify and blocking
        locks easier.
Jeremy.

25 years agoIn addition to the HAVE_FUNCTION_MACRO test that Andrew used as an example,
Christopher R. Hertel [Thu, 30 Jul 1998 20:39:30 +0000 (20:39 +0000)]
In addition to the HAVE_FUNCTION_MACRO test that Andrew used as an example,
I also need to know if the __FILE__ macro exists.  I followed the example,
but ran into two problems:

1) I don't have the autoconf tools installed.
2) The instructions did not say which files should be checked in after
   running autoheader and autoconf.

So I'm checking in the modified acconfig.h and configure.in on the
assumption that the next time someone runs autoheader and autoconf my
new tests will be included.

Hope that's not a problem (and I'll try to grab a copy of autoconf from
somewhere).

Chris -)-----

25 years agoadded HAVE_FUNCTION_MACRO test
Andrew Tridgell [Thu, 30 Jul 1998 07:05:15 +0000 (07:05 +0000)]
added HAVE_FUNCTION_MACRO test

25 years agofixed sin_len test for FreeBSD
Andrew Tridgell [Thu, 30 Jul 1998 04:54:14 +0000 (04:54 +0000)]
fixed sin_len test for FreeBSD

25 years agofixed a make proto bug pointed out by Chris.
Andrew Tridgell [Thu, 30 Jul 1998 01:49:58 +0000 (01:49 +0000)]
fixed a make proto bug pointed out by Chris.

25 years agoupdated prototypes
Andrew Tridgell [Wed, 29 Jul 1998 07:11:33 +0000 (07:11 +0000)]
updated prototypes

25 years agofixed a bug in the replacement inet_ntoa
Andrew Tridgell [Wed, 29 Jul 1998 07:07:08 +0000 (07:07 +0000)]
fixed a bug in the replacement inet_ntoa

25 years agotest for a broken inet_ntoa and replace it if necessary (for
Andrew Tridgell [Wed, 29 Jul 1998 07:02:06 +0000 (07:02 +0000)]
test for a broken inet_ntoa and replace it if necessary (for
IRIX+gcc-2.8.1)

25 years agoignore the auto-generated dummy file
Andrew Tridgell [Wed, 29 Jul 1998 05:09:10 +0000 (05:09 +0000)]
ignore the auto-generated dummy file

25 years agoadded Makefile to .cvsignore in the hope that this will prevent people
Andrew Tridgell [Wed, 29 Jul 1998 05:07:29 +0000 (05:07 +0000)]
added Makefile to .cvsignore in the hope that this will prevent people
from committing changes to it.

25 years agoget rid of the runtime test for broken getgroups() and add a compile
Andrew Tridgell [Wed, 29 Jul 1998 05:05:36 +0000 (05:05 +0000)]
get rid of the runtime test for broken getgroups() and add a compile
time test instead. This also allows us to get rid of the igroups
element of a couple of structures.

25 years agoremoved some feedback request comments - we get enough email :)
Andrew Tridgell [Wed, 29 Jul 1998 03:34:43 +0000 (03:34 +0000)]
removed some feedback request comments - we get enough email :)

25 years agoupdated the UNIX install instructions to include ./configure
Andrew Tridgell [Wed, 29 Jul 1998 03:34:08 +0000 (03:34 +0000)]
updated the UNIX install instructions to include ./configure

25 years agohopefully permissions are right now
Andrew Tridgell [Wed, 29 Jul 1998 03:20:19 +0000 (03:20 +0000)]
hopefully permissions are right now

25 years agostill trying to get permissions right - CVS doesn't commit permissions changes it...
Andrew Tridgell [Wed, 29 Jul 1998 03:19:10 +0000 (03:19 +0000)]
still trying to get permissions right - CVS doesn't commit permissions changes it seems

25 years agocheckin configure again in order to get permissions right
Andrew Tridgell [Wed, 29 Jul 1998 03:16:09 +0000 (03:16 +0000)]
checkin configure again in order to get permissions right

25 years agosome merge cleanups
Andrew Tridgell [Wed, 29 Jul 1998 03:15:02 +0000 (03:15 +0000)]
some merge cleanups

25 years agomerge from the autoconf2 branch to the main branch
Andrew Tridgell [Wed, 29 Jul 1998 03:08:05 +0000 (03:08 +0000)]
merge from the autoconf2 branch to the main branch

25 years agolocking.c: Print messages when we downgrade a lock.
Jeremy Allison [Wed, 29 Jul 1998 00:27:23 +0000 (00:27 +0000)]
locking.c: Print messages when we downgrade a lock.
reply.c: Do the same mask expansion we do in trans2.c - needed for Win98.
trans2.c: Make the mask expansion into a function call now we have to
          do it twice.
Jeremy.

25 years agolocking.c: Don't need to do map_lock_type on *testing* a lock, just on setting.
Jeremy Allison [Tue, 28 Jul 1998 18:26:47 +0000 (18:26 +0000)]
locking.c: Don't need to do map_lock_type on *testing* a lock, just on setting.
Jeremy.

25 years agoloadparm.c: Added strict sync parameter.
Jeremy Allison [Tue, 28 Jul 1998 18:15:31 +0000 (18:15 +0000)]
loadparm.c: Added strict sync parameter.
locking.c: Added code to deal with real open mode of file.
reply.c: Added strict sync parameter.
server.c: Added strict sync parameter. Fixed open modes.
Jeremy.

25 years agochgpasswd.c: Fixed up debug calls to stop crashes if ptsname failed.
Jeremy Allison [Mon, 27 Jul 1998 18:50:45 +0000 (18:50 +0000)]
chgpasswd.c: Fixed up debug calls to stop crashes if ptsname failed.
local.h: Kept FSTYPE_STRING as Samba for now.
nmbd_browsesync.c: Added bugfix from Matt Chapman mattyc@cyberdude.com
                   - lmb_browserlist is now a struct ubi_dlList not a
                   struct browse_cache_record *.
server.c:
smb.h:
uid.c:
password.c: Removed attrs code - it is not used anywhere.
Jeremy

25 years agoCleaned up some testing code and made it more "permanent" looking. The
Christopher R. Hertel [Sat, 25 Jul 1998 15:45:42 +0000 (15:45 +0000)]
Cleaned up some testing code and made it more "permanent" looking.  The
NetBIOS name lists attached to the subnet records are now managed by
the splay tree code.  I am still working on the WINS database as a
separate issue.  Code is written, it's just a matter of incorporating it.
CRH

25 years agoNoticed that I was using the strlen() of a string that I had strdup()'d
Christopher R. Hertel [Sat, 25 Jul 1998 15:18:07 +0000 (15:18 +0000)]
Noticed that I was using the strlen() of a string that I had strdup()'d
before testing that the strdup() worked.  Fixed.

25 years agoConverted the browser database to a ubi_dLinkList. This should reduce code
Christopher R. Hertel [Fri, 24 Jul 1998 19:03:11 +0000 (19:03 +0000)]
Converted the browser database to a ubi_dLinkList.  This should reduce code
size, etc.  Also did a bit of work to add comments.
Chris -)-----

25 years agoChanged the definition of the linked list header used in the directory
Christopher R. Hertel [Fri, 24 Jul 1998 07:43:13 +0000 (07:43 +0000)]
Changed the definition of the linked list header used in the directory
cache so that it uses the new ubi_dlNewList() macro in ubi_dLinkList.h.

25 years agoTo each list type, I added a macro that makes it easier to define and
Christopher R. Hertel [Fri, 24 Jul 1998 07:39:14 +0000 (07:39 +0000)]
To each list type, I added a macro that makes it easier to define and
initialize a list header.

25 years agoUpdates smbstatus to have a -S and -L flag.
Richard Sharpe [Fri, 24 Jul 1998 01:08:31 +0000 (01:08 +0000)]
Updates smbstatus to have a -S and -L flag.

Also updated the documentation :-)

The code is ugly, we need to rewrite smbstatus someday.

25 years agolocking.c: Added lock type to is_locked() and do_lock()
Jeremy Allison [Thu, 23 Jul 1998 00:10:26 +0000 (00:10 +0000)]
locking.c: Added lock type to is_locked() and do_lock()
           as the code in reply_lockingX wasn't taking account of
           the difference between read and write locks ! How did this
           ever work :-) !
reply.c:
server.c: Add lock type to is_locked() and do_lock().
util.c: Also added code from klausr@ITAP.Physik.Uni-Stuttgart.De
        to fix problem with log files growing too large if an
        smbd writes less than 100 debug messages.
Jeremy.

25 years agoFixed bug found by Richard Sharpe. After increasing files_struct size by
Jeremy Allison [Wed, 22 Jul 1998 13:59:19 +0000 (13:59 +0000)]
Fixed bug found by Richard Sharpe. After increasing files_struct size by
MAX_OPEN_DIRECTORIES for nttrans I forgot to update the code that
enumerates the array.
Created new MAX_FNUMS in local.h, changed all code that iterates
through the files_struct array to use this.
(sorry Richard).
Jeremy.

25 years agoincludes.h: Added feature type USE_GRANTPT for pty code.
Jeremy Allison [Wed, 22 Jul 1998 01:31:59 +0000 (01:31 +0000)]
includes.h: Added feature type USE_GRANTPT for pty code.
chgpasswd.c: Updated to use USE_GRANTPT feature definition.
Jeremy.

25 years agoForgot to increase size of files_struct by MAX_OPEN_DIRECTORIES in
Jeremy Allison [Fri, 17 Jul 1998 22:22:48 +0000 (22:22 +0000)]
Forgot to increase size of files_struct by MAX_OPEN_DIRECTORIES in
previous checkin.
Jeremy.

25 years agoCode added to fix the renaming of a directory under NT SMB calls.
Jeremy Allison [Fri, 17 Jul 1998 22:21:24 +0000 (22:21 +0000)]
Code added to fix the renaming of a directory under NT SMB calls.
local.h: Changed MAXDIR to MAX_OPEN_DIRECTORIES - shmem size also tuned by this.
dir.c: Use MAX_OPEN_DIRECTORIES.
nttrans.c: Allow opening of a directory to succeed. Doesn't actually open
           a file descriptor but takes a files_struct slot marked as an
           fd.
reply.c: Changed to close any outstanding is_directory files.
         reply_close changed to understand directory files.
server.c: Added open_directory(), close_directory() calls.
smb.h: Added is_directory to files_struct.
       Changed OPEN_FNUM to check that target is !is_directory (this
       prevents the normal file calls from processing a directory
       files_struct.
Jeremy.

25 years agonttrans.c: Added FILE_EXECUTE to required permissions checked.
Jeremy Allison [Fri, 17 Jul 1998 01:38:08 +0000 (01:38 +0000)]
nttrans.c: Added FILE_EXECUTE to required permissions checked.
trans2.c: Fixed up the SMB_QUERY_FILE_ALT_NAME_INFO code, now I've
          got an NT client to generate it. It uses unicode without
          bothering to check the unicode bit of course, but now we
          can execute 16 bit exe's on a WinNT client from a long
          filename directory (hurrah).
Jeremy.

25 years agolocal.h: Added NTFS define if HAVE_NT_SMBS is defined.
Jeremy Allison [Fri, 17 Jul 1998 00:47:16 +0000 (00:47 +0000)]
local.h: Added NTFS define if HAVE_NT_SMBS is defined.
nttrans.c: Fixed issue with access DELETE on renaming files.
Jeremy.

25 years agoMakefile: Added nttrans.o
Jeremy Allison [Thu, 16 Jul 1998 22:46:06 +0000 (22:46 +0000)]
Makefile: Added nttrans.o
includes.h: Added termios.h for AIX.
nttrans.c: Working NT SMB calls !
pipes.c: Use strequal instead of strcmp.
server.c: Use #defines rather than numbers.
smb.h: Updated NT SMB #defines.
Jeremy.

25 years agoMakefile: Added CC=gcc to DGUX on Intel. Comment from ross@filmworks.com.
Jeremy Allison [Thu, 16 Jul 1998 00:06:29 +0000 (00:06 +0000)]
Makefile: Added CC=gcc to DGUX on Intel. Comment from ross@filmworks.com.
ipc.c:
loadparm.c:
printing.c:
           Added code from <Dirk.DeWachter@rug.ac.be> to implement print
           queue pausing. New parameters are "queuepause command" and
           "queueresume command".
util.c: Added fix for mount options in autmount map.
lib/rpc/include/rpc_misc.h: Removed duplicate pipe names for Jean-Francois.
Jeremy.

25 years agoutil.c: I've added a function called mem_dup(). Similar to strdup(),
Christopher R. Hertel [Wed, 15 Jul 1998 20:15:25 +0000 (20:15 +0000)]
util.c:  I've added a function called mem_dup().  Similar to strdup(),
         mem_dup() allocates the required memory before copying the
         source data.  It returns NULL if memory could not be allcoated,
         else a pointer to the newly allocated memory.
proto.h: Rebuilt to add the prototype for mem_dup().

25 years agoloadparm.c:
Jeremy Allison [Tue, 14 Jul 1998 21:23:59 +0000 (21:23 +0000)]
loadparm.c:

    With apologies to Charlton Heston and Pierre Boule.

    "You damn fools, you finally did it".

    Changed default security mode to be security=user.
    Yes this is a big (although small in code) change.
    It's something we've been discussing for a while, to
    finally wean people off the legacy security=share mode
    which is *never* what you want.

    Jeremy.

nmbd_incomingrequests.c: Bug fix for nmbd core dumps caused by overrun.
                         Found by <samuel@public.szonline.net>.
nttrans.c: More NT smb stuff.
reply.c: Unlink will overwrite an existing file. Well you learn
         something new about POSIX every day. :-).
server.c: Tidyup unreadable code.
smbpasswd.c: Code to allow -U remote_username to allow ordinary
             users to change remote passwords if their NT username
             is different from their UNIX username.
             Patch from <torbjorn.lindh@allgon.se>.

Jeremy.

25 years agoAdded strupper() function call to up-case the scope field in the
Christopher R. Hertel [Tue, 14 Jul 1998 01:52:09 +0000 (01:52 +0000)]
Added strupper() function call to up-case the scope field in the
make_nmb_name() function.  Database lookups (eg. gdbm) will often use
byte-by-byte comparisons, so it is important that the case and padding
are correct.

Chris -)-----

25 years agonntrans.c: Fully implemented transact rename.
Jeremy Allison [Sat, 11 Jul 1998 01:25:02 +0000 (01:25 +0000)]
nntrans.c: Fully implemented transact rename.
reply.c: Added NT specific rename if exists flag to rename_internals().
smb.h: Added NT rename flag.
Jeremy.

25 years agonttrans.c: More NT SMB stuff.
Jeremy Allison [Sat, 11 Jul 1998 00:28:34 +0000 (00:28 +0000)]
nttrans.c: More NT SMB stuff.
reply.c: Broke out the internals of reply_mv so that they may be called
         externally from the NT transact rename.
server.c: Changed stat calls to sys_stat - found in code review
          of bugfix.
Jeremy.

25 years agoFix for bug PR#8294 reported by <detlef.lammermann@er.materna.de>
Jeremy Allison [Fri, 10 Jul 1998 20:03:09 +0000 (20:03 +0000)]
Fix for bug PR#8294 reported by <detlef.lammermann@er.materna.de>
- the Files array was storing the 'dos_to_unix' translated name,
rather than the untranslated name. This could case problems when
the name was run through dos_to_unix again.
Jeremy.