kai/samba.git
23 years agofail a print start on a deleted auto printer
Andrew Tridgell [Wed, 10 May 2000 11:49:06 +0000 (11:49 +0000)]
fail a print start on a deleted auto printer

23 years agoI found a better way of handling deleted auto printers
Andrew Tridgell [Wed, 10 May 2000 11:48:30 +0000 (11:48 +0000)]
I found a better way of handling deleted auto printers

23 years agoformatting fix
Andrew Tridgell [Wed, 10 May 2000 11:45:16 +0000 (11:45 +0000)]
formatting fix

23 years agoremove autoloaded printers that are no longer in /etc/printcap when we
Andrew Tridgell [Wed, 10 May 2000 11:20:46 +0000 (11:20 +0000)]
remove autoloaded printers that are no longer in /etc/printcap when we
get a HUP

23 years agopatch from Dominik Kubla <dominik.kubla@uni-mainz.de>
Andrew Tridgell [Wed, 10 May 2000 11:08:08 +0000 (11:08 +0000)]
patch from Dominik Kubla <dominik.kubla@uni-mainz.de>

23 years agomore merging voodoo
Andrew Tridgell [Wed, 10 May 2000 10:41:59 +0000 (10:41 +0000)]
more merging voodoo

this adds "#define OLD_NTDOMAIN 1" in lots of places. Don't panic -
this isn't permanent, it should go after another few merge steps have
been done

23 years agomore merging
Andrew Tridgell [Wed, 10 May 2000 09:49:55 +0000 (09:49 +0000)]
more merging

it is now at the stage that winbindd can compile in the head branch,
but not link

23 years ago- add some reserved space to every tdb, this will be
Andrew Tridgell [Wed, 10 May 2000 06:02:37 +0000 (06:02 +0000)]
- add some reserved space to every tdb, this will be
  used for semaphores and future expansion
- update tdb version number to 2

23 years agoUse a prime for default tdb hash size - this makes a difference !
Herb Lewis [Wed, 10 May 2000 01:32:51 +0000 (01:32 +0000)]
Use a prime for default tdb hash size - this makes a difference !
Jeremy.

23 years agoFix for misunderstanding of fsync added when vfs layer
Herb Lewis [Wed, 10 May 2000 01:31:46 +0000 (01:31 +0000)]
Fix for misunderstanding of fsync added when vfs layer
was done. Samba was doing fsync's (bleagh).
Jeremy.

23 years agoUsing a structure for a tdb key can lead to insideous, hard
Herb Lewis [Wed, 10 May 2000 00:05:27 +0000 (00:05 +0000)]
Using a structure for a tdb key can lead to insideous, hard
to find bugs. On 64 bit IRIX, structure packing means that
a
struct {
SMB_DEV_T dev /* 4 bytes */
SMB_INO_T ino /* 8 bytes */
}

has 4 bytes of padding between the two members. If you
don't null the memory before using it as a tdb key,
you randomly can't find keys depending on what is in
the padding. This caused me immense pain and was hard
to track down.... :-)

Jeremy.

23 years agoFix for check_kernel_oplocks looking for the var directory
Herb Lewis [Tue, 9 May 2000 22:07:07 +0000 (22:07 +0000)]
Fix for check_kernel_oplocks looking for the var directory
before it may need to be created.
Jeremy.

23 years agoThis is reall Jeremy commiting as Herb at SGI labs. Fix
Herb Lewis [Tue, 9 May 2000 22:05:39 +0000 (22:05 +0000)]
This is reall Jeremy commiting as Herb at SGI labs. Fix
for "socket operation on non-socket" error in log.smb
on HEAD branch startup (server_fd not initialized to -1).
Jeremy.

23 years ago- use smb_gwtpwnam() in another couple of places
Andrew Tridgell [Tue, 9 May 2000 15:09:52 +0000 (15:09 +0000)]
- use smb_gwtpwnam() in another couple of places
- don't call add/del user if the scripts are empty

23 years agofixed some winbind cache bugs
Andrew Tridgell [Tue, 9 May 2000 15:08:46 +0000 (15:08 +0000)]
fixed some winbind cache bugs

23 years agothe beginnings of a new scheme I've working on to allow an easier
Andrew Tridgell [Tue, 9 May 2000 13:28:19 +0000 (13:28 +0000)]
the beginnings of a new scheme I've working on to allow an easier
head/tng merge.

It goes something like this:

- headers from tng get copied over one at a time

- the old headers get renamed to *_old.h

- server side code that used the old headers gets a
  #define OLD_NTDOMAIN 1
  #undef OLD_NTDOMAIN
  at the start and end of the code

- mkproto.awk recognises these special defines and does magic stuff so
  that each .c file sees the right headers

- we start moving the rpc client libraries from tng to head.

if this goes OK then, in theory, we should be able to move the client
side rpc code from tng to head without disturbing the existing head
server side code. Then when that works we can consider merging the
server side.

it remains to be seen if this scheme will work. So far I've moved
rpc_samr.h and don't seem to have broken anything.

Note this this is still a very delicate operation, as at every step of
the way I want to keep head fully functional. Please don't take part
unless you discuss it with me first.

23 years agodon't attempt to build rpcclient in the head branch
Andrew Tridgell [Tue, 9 May 2000 13:21:58 +0000 (13:21 +0000)]
don't attempt to build rpcclient in the head branch

23 years agobrought across some rpc header files from tng
Andrew Tridgell [Tue, 9 May 2000 11:44:09 +0000 (11:44 +0000)]
brought across some rpc header files from tng

23 years agobrought the winbindd code into head
Andrew Tridgell [Tue, 9 May 2000 11:43:00 +0000 (11:43 +0000)]
brought the winbindd code into head

this does not yet compile, but I'm working on that.

23 years agoadded some rules for winbindd and pam_winbind
Andrew Tridgell [Tue, 9 May 2000 11:41:38 +0000 (11:41 +0000)]
added some rules for winbindd and pam_winbind

23 years agoAdded tdb_store_by_string() and tdb_fetch_by_string() functions to store
Tim Potter [Tue, 9 May 2000 06:22:12 +0000 (06:22 +0000)]
Added tdb_store_by_string() and tdb_fetch_by_string() functions to store
data with null terminated string keys.

23 years agothe beginnings of a description of how to setup a Samba appliance
Andrew Tridgell [Tue, 9 May 2000 04:02:31 +0000 (04:02 +0000)]
the beginnings of a description of how to setup a Samba appliance

23 years agoFix for VMS platforms from "John E. Malmberg" <wb8tyw@qsl.net>
Jeremy Allison [Mon, 8 May 2000 18:24:26 +0000 (18:24 +0000)]
Fix for VMS platforms from "John E. Malmberg" <wb8tyw@qsl.net>
Jeremy.

23 years agoSomeone :-) forgot to add secrets.c to HEAD.
Jeremy Allison [Mon, 8 May 2000 18:14:25 +0000 (18:14 +0000)]
Someone :-) forgot to add secrets.c to HEAD.
Jeremy.

23 years agoadded secrets.tdb and changed storage of trust account password to use
Andrew Tridgell [Mon, 8 May 2000 10:42:21 +0000 (10:42 +0000)]
added secrets.tdb and changed storage of trust account password to use
it

23 years agoAdded winbindd parameters to smb.conf documentation.
Tim Potter [Mon, 8 May 2000 10:38:52 +0000 (10:38 +0000)]
Added winbindd parameters to smb.conf documentation.

23 years agoInitial version of winbindd manual page.
Tim Potter [Mon, 8 May 2000 10:00:02 +0000 (10:00 +0000)]
Initial version of winbindd manual page.

23 years agoadded winbindd options in head branch, so it is possible to combine
Andrew Tridgell [Sat, 6 May 2000 15:06:45 +0000 (15:06 +0000)]
added winbindd options in head branch, so it is possible to combine
branches

23 years agofixed a vfs crash bug
Andrew Tridgell [Sat, 6 May 2000 15:00:49 +0000 (15:00 +0000)]
fixed a vfs crash bug

23 years agosigned/unsigned fixes so we can handle a lock base close to 2^32
Andrew Tridgell [Sat, 6 May 2000 01:36:59 +0000 (01:36 +0000)]
signed/unsigned fixes so we can handle a lock base close to 2^32

23 years agoFix for uninitialized memory read in brlock code. brl_locktest now needs
Jeremy Allison [Fri, 5 May 2000 20:50:22 +0000 (20:50 +0000)]
Fix for uninitialized memory read in brlock code. brl_locktest now needs
to correctly set the fnum, as the brl_conflict code looks at it.
Jeremy.

23 years agoFix for dual unix_to_dos conversion in readdir from A.V.Shutko <AVShutko@mail.khstu.ru>
Jeremy Allison [Fri, 5 May 2000 19:57:02 +0000 (19:57 +0000)]
Fix for dual unix_to_dos conversion in readdir from A.V.Shutko <AVShutko@mail.khstu.ru>
Jeremy.

23 years agoadd new release notes chapter to samba for irix package
Herb Lewis [Fri, 5 May 2000 14:31:36 +0000 (14:31 +0000)]
add new release notes chapter to samba for irix package

23 years agoFix for crash bug in OpenConfFile() if no smb.conf exists.
Tim Potter [Fri, 5 May 2000 04:18:22 +0000 (04:18 +0000)]
Fix for crash bug in OpenConfFile() if no smb.conf exists.

23 years agoFix for read_file() returning -1.
Jeremy Allison [Fri, 5 May 2000 02:39:02 +0000 (02:39 +0000)]
Fix for read_file() returning -1.
Jeremy.

23 years agoTwo fixes. Added missong logic & case in lock split code.
Jeremy Allison [Fri, 5 May 2000 02:17:39 +0000 (02:17 +0000)]
Two fixes. Added missong logic & case in lock split code.
Fixed range split into two, as DLIST_ADD has the wrong semantics...
Jeremy.

23 years agomake debug easier to read
Andrew Tridgell [Fri, 5 May 2000 01:26:52 +0000 (01:26 +0000)]
make debug easier to read

23 years agomake debug easier to read
Andrew Tridgell [Fri, 5 May 2000 01:20:25 +0000 (01:20 +0000)]
make debug easier to read

23 years agonasty hack to print posix locks
Andrew Tridgell [Fri, 5 May 2000 01:11:23 +0000 (01:11 +0000)]
nasty hack to print posix locks

23 years agoexample of broken posix lock behaviour
Andrew Tridgell [Fri, 5 May 2000 00:46:38 +0000 (00:46 +0000)]
example of broken posix lock behaviour

23 years agorpc_server/srv_lsa.c: Bring into sync with 2.0.x.
Jeremy Allison [Thu, 4 May 2000 21:57:28 +0000 (21:57 +0000)]
rpc_server/srv_lsa.c: Bring into sync with 2.0.x.
rpc_server/srv_pipe_hnd.c: Bring into sync with 2.0.x.
smbd/blocking.c: Improve blocking debug reporting.
utils/torture.c: Added check for NT locking bug.
Jeremy.

23 years agoUpdated to fix overlapping problem.
Jeremy Allison [Thu, 4 May 2000 19:01:14 +0000 (19:01 +0000)]
Updated to fix overlapping problem.
Jeremy.

23 years agoa minimal change to get appliance mode to work with winbindd
Andrew Tridgell [Thu, 4 May 2000 16:01:47 +0000 (16:01 +0000)]
a minimal change to get appliance mode to work with winbindd

we needed to accept usernames of the form DOMAIN/user, which means we
needed to pass the domain to a getpwnam() like routine in certain
critical spots.

What I'd rather do is get rid of "char *user" everywhere and use the
new userdom_struct, but that will have to wait a few days.

23 years agoparameterize the lock timeout
Andrew Tridgell [Thu, 4 May 2000 09:32:21 +0000 (09:32 +0000)]
parameterize the lock timeout

23 years agoadded a test for the NT byte range lock into smbtorture
Andrew Tridgell [Thu, 4 May 2000 09:31:57 +0000 (09:31 +0000)]
added a test for the NT byte range lock into smbtorture

23 years agoan even simpler example of NT gettings its locking code wrong. This
Andrew Tridgell [Thu, 4 May 2000 08:58:07 +0000 (08:58 +0000)]
an even simpler example of NT gettings its locking code wrong. This
one doesn't even need two connections to the box, just two file handles.

it is a very simple case actually, and one I think will happen quite a
lot in real life. I wonder how they haven't noticed it? I checked and
W2K has the same bug.

23 years ago - use full_name instead of real_name
Andrew Tridgell [Thu, 4 May 2000 07:59:34 +0000 (07:59 +0000)]
 - use full_name instead of real_name
 - got rid of guest map code in lpq parser

23 years agoneed LOCKING_OBJ in locktest now
Andrew Tridgell [Thu, 4 May 2000 06:57:26 +0000 (06:57 +0000)]
need LOCKING_OBJ in locktest now

23 years agomore locktest improvements
Andrew Tridgell [Thu, 4 May 2000 06:30:17 +0000 (06:30 +0000)]
more locktest improvements

the current preset tests show a very interesting NT bug - it sometimes
refuses a lock that it shouldn't.

23 years agofixed message text
Andrew Tridgell [Thu, 4 May 2000 06:29:13 +0000 (06:29 +0000)]
fixed message text

23 years agoif the stat cache is off then don't initialise it
Andrew Tridgell [Thu, 4 May 2000 06:28:38 +0000 (06:28 +0000)]
if the stat cache is off then don't initialise it

23 years agoadded -A analyze mode to locktest - it can now automatically prune
Andrew Tridgell [Thu, 4 May 2000 04:14:23 +0000 (04:14 +0000)]
added -A analyze mode to locktest - it can now automatically prune
the list of locks that cause a failure

23 years agofound a much simpler case that kills the posix locking
Andrew Tridgell [Thu, 4 May 2000 02:12:33 +0000 (02:12 +0000)]
found a much simpler case that kills the posix locking

23 years agofixed a uninit memory read that insure found
Andrew Tridgell [Thu, 4 May 2000 02:01:48 +0000 (02:01 +0000)]
fixed a uninit memory read that insure found

23 years agoadded ability to present lock tests
Andrew Tridgell [Thu, 4 May 2000 01:40:51 +0000 (01:40 +0000)]
added ability to present lock tests

23 years agoFix for debug statement crash.
Jeremy Allison [Thu, 4 May 2000 00:29:50 +0000 (00:29 +0000)]
Fix for debug statement crash.
Jeremy.

23 years agounicode code pages are detected as compiler intermediate file so set to
Herb Lewis [Wed, 3 May 2000 22:29:54 +0000 (22:29 +0000)]
unicode code pages are detected as compiler intermediate file so set to
nostrip

23 years agoFix based on Elrond's code to reduce the store size on removing dead records.
Jeremy Allison [Wed, 3 May 2000 20:33:25 +0000 (20:33 +0000)]
Fix based on Elrond's code to reduce the store size on removing dead records.
Jeremy.

23 years agoFixed the bug locktest.c found, it was an off-by-one error in the non-overlap
Jeremy Allison [Wed, 3 May 2000 19:37:47 +0000 (19:37 +0000)]
Fixed the bug locktest.c found, it was an off-by-one error in the non-overlap
case.
Jeremy.

23 years agoFixed silly bug where I wasn't checking for matching fd's in closing a file.
Jeremy Allison [Wed, 3 May 2000 18:57:21 +0000 (18:57 +0000)]
Fixed silly bug where I wasn't checking for matching fd's in closing a file.
This caused smbd crashes on SIGKILL.
Jeremy.

23 years agoAdded range info to the panic message to track down the bug with
Jeremy Allison [Wed, 3 May 2000 17:50:42 +0000 (17:50 +0000)]
Added range info to the panic message to track down the bug with
locktest.
Jeremy.

23 years agoimproved the lock test program
Andrew Tridgell [Wed, 3 May 2000 14:49:15 +0000 (14:49 +0000)]
improved the lock test program

23 years ago- clear dead locks at startup
Andrew Tridgell [Wed, 3 May 2000 14:29:05 +0000 (14:29 +0000)]
- clear dead locks at startup

- fixed a bug which caused lock records to not decrease in size

23 years agoadded a nasty lock testing program
Andrew Tridgell [Wed, 3 May 2000 13:54:28 +0000 (13:54 +0000)]
added a nasty lock testing program

it opens 2 connections to each of 2 servers, and opens 2 fnums on the
same file on each connection (a total of 8 file descriptors)

then it does random lock/unlock/reopen requests in a 100 byte range on
the file and compares the results from the 2 servers.

strangely enough, NT fails this test against itself right now - I'm
still trying to figure that out.

23 years agoFix for stacking locks in brlock and POSIX. Windows only allows a read lock
Jeremy Allison [Wed, 3 May 2000 02:27:41 +0000 (02:27 +0000)]
Fix for stacking locks in brlock and POSIX. Windows only allows a read lock
to overlay a write lock on the same fnum. When overlaying read locks onto
a write lock, the number of locks is counted, and the first unlock removes
the write lock and downgrades this to a read lock. Do the same when mapping
to POSIX.
Jeremy.

23 years agoAdded read overlay write lock on same fnum test to locktest2.
Jeremy Allison [Wed, 3 May 2000 02:24:56 +0000 (02:24 +0000)]
Added read overlay write lock on same fnum test to locktest2.
Added locktest5 for locking stacks.
Jeremy.

23 years agoInsure uninitialized memory read fixes.
Jeremy Allison [Wed, 3 May 2000 02:24:01 +0000 (02:24 +0000)]
Insure uninitialized memory read fixes.
Jeremy.

23 years agomerge 2.0 changes
Herb Lewis [Wed, 3 May 2000 01:07:33 +0000 (01:07 +0000)]
merge 2.0 changes

23 years agoFixed wild pointer diff found by insure.
Jeremy Allison [Tue, 2 May 2000 23:56:24 +0000 (23:56 +0000)]
Fixed wild pointer diff found by insure.
Jeremy.

23 years agoImplemented the last (I hope:-) part of the locking puzzle, the reference
Jeremy Allison [Tue, 2 May 2000 21:07:25 +0000 (21:07 +0000)]
Implemented the last (I hope:-) part of the locking puzzle, the reference
counting when Windows downgrades a write lock to a read lock, then reference
counts the unlocks to match the locks.
With this code the POSIX unlock isn't done until the final Windows unlock.
Jeremy.

23 years agoadded support for deleting printers into the spoolss system
Andrew Tridgell [Tue, 2 May 2000 15:31:55 +0000 (15:31 +0000)]
added support for deleting printers into the spoolss system

23 years agosplit the username in the vuser structure into a separate
Andrew Tridgell [Tue, 2 May 2000 13:55:42 +0000 (13:55 +0000)]
split the username in the vuser structure into a separate
userdom_struct. As the name implies this also contains a domain
(unused at the moment).

This will be important shortly, as operation in appliance mode needs
the domain to be always carried with the username.

23 years agoand yet another memory leak - this one in the client
Andrew Tridgell [Tue, 2 May 2000 07:50:33 +0000 (07:50 +0000)]
and yet another memory leak - this one in the client

23 years agofixed a memory leak (calling hash_table_init twice)
Andrew Tridgell [Tue, 2 May 2000 07:10:54 +0000 (07:10 +0000)]
fixed a memory leak (calling hash_table_init twice)

23 years agofixed a memory leak
Andrew Tridgell [Tue, 2 May 2000 07:10:26 +0000 (07:10 +0000)]
fixed a memory leak

23 years agoMoved uglyness needed in fcntl locking (64->32 bit mapping, NFS
Jeremy Allison [Tue, 2 May 2000 03:20:47 +0000 (03:20 +0000)]
Moved uglyness needed in fcntl locking (64->32 bit mapping, NFS
errors etc.) into locking/posix.c, where it is needed. fcntl_lock in lib/util.c
is now very small and clean.
Added (*lock) op to vfs layer.
Jeremy.

23 years agoAdded sys_fork() and sys_getpid() functions to stop the overhead
Jeremy Allison [Tue, 2 May 2000 02:23:41 +0000 (02:23 +0000)]
Added sys_fork() and sys_getpid() functions to stop the overhead
of doing a system call every time we want to just get our pid.
Jeremy.

23 years agolocking/posix.c: Fixed double-free nasty crash bug found by insure.
Jeremy Allison [Mon, 1 May 2000 19:43:50 +0000 (19:43 +0000)]
locking/posix.c: Fixed double-free nasty crash bug found by insure.
utils/make_smbcodepage.c:
utils/make_unicodemap.c: Insure 'make install' fixes.
Jeremy.

23 years agohandle tabs in printcap files
Andrew Tridgell [Mon, 1 May 2000 05:42:39 +0000 (05:42 +0000)]
handle tabs in printcap files

this is needed for lprng

23 years agoallow a DEFAULT_PRINTING to be specified in CFLAGS
Andrew Tridgell [Mon, 1 May 2000 05:42:09 +0000 (05:42 +0000)]
allow a DEFAULT_PRINTING to be specified in CFLAGS

23 years agofixed typo
Andrew Tridgell [Mon, 1 May 2000 00:48:41 +0000 (00:48 +0000)]
fixed typo

23 years agoadded TDB_INTERNAL, TDB_NOLOCK and TDB_NOMMAP flags.
Andrew Tridgell [Mon, 1 May 2000 00:41:47 +0000 (00:41 +0000)]
added TDB_INTERNAL, TDB_NOLOCK and TDB_NOMMAP flags.

TDB_INTERNAL replaces the old method of passing a null filename

23 years ago - added some error checking
Andrew Tridgell [Sun, 30 Apr 2000 15:13:15 +0000 (15:13 +0000)]
 - added some error checking
 - removed the VTP hook in smbd

23 years agofixed dptr_wcard handling (need to use strdup)
Andrew Tridgell [Sun, 30 Apr 2000 14:59:00 +0000 (14:59 +0000)]
fixed dptr_wcard handling (need to use strdup)

23 years agofixed parsing of broken NT short name
Andrew Tridgell [Sun, 30 Apr 2000 14:58:13 +0000 (14:58 +0000)]
fixed parsing of broken NT short name

23 years agohandle the special rule of *.* for old style listings when old_list is
Andrew Tridgell [Sun, 30 Apr 2000 14:37:57 +0000 (14:37 +0000)]
handle the special rule of *.* for old style listings when old_list is
true

23 years agoremoved more cruft from our old wildcard matching code
Andrew Tridgell [Sun, 30 Apr 2000 14:34:31 +0000 (14:34 +0000)]
removed more cruft from our old wildcard matching code

23 years agogo back to ascii in SMB_FIND_FILE_BOTH_DIRECTORY_INFO
Andrew Tridgell [Sun, 30 Apr 2000 14:30:47 +0000 (14:30 +0000)]
go back to ascii in SMB_FIND_FILE_BOTH_DIRECTORY_INFO

NT gives out unicode, but NT clients can't handle a unicode response!

23 years agofixed our smbsearch code. We now store the mask with the dptr, this
Andrew Tridgell [Sun, 30 Apr 2000 14:29:45 +0000 (14:29 +0000)]
fixed our smbsearch code. We now store the mask with the dptr, this
turns out to be essential for a correct implementation (there ins't
enough room to store all possible masks in the status return
structure!)

23 years agofix handing of ascii_to_unistr
Andrew Tridgell [Sun, 30 Apr 2000 14:27:39 +0000 (14:27 +0000)]
fix handing of ascii_to_unistr

23 years ago- get the findclose code right
Andrew Tridgell [Sun, 30 Apr 2000 14:26:59 +0000 (14:26 +0000)]
- get the findclose code right
- handle broken NT response to trans2 findfirst

23 years agoadded cli_list_old() to allow for old style directory listing from
Andrew Tridgell [Sun, 30 Apr 2000 12:34:26 +0000 (12:34 +0000)]
added cli_list_old() to allow for old style directory listing from
masktest

23 years agosome cleanups
Andrew Tridgell [Sun, 30 Apr 2000 11:25:09 +0000 (11:25 +0000)]
some cleanups

23 years agowe don't need fnmatch.c any more
Andrew Tridgell [Sun, 30 Apr 2000 11:15:27 +0000 (11:15 +0000)]
we don't need fnmatch.c any more

23 years agochanged masktest to test the internal algorithm against one server,
Andrew Tridgell [Sun, 30 Apr 2000 11:11:19 +0000 (11:11 +0000)]
changed masktest to test the internal algorithm against one server,
rather than 2 servers

23 years ago- removed all our old wildcard matching code and replaced it with a
Andrew Tridgell [Sun, 30 Apr 2000 11:04:28 +0000 (11:04 +0000)]
- removed all our old wildcard matching code and replaced it with a
call to ms_fnmatch(). This also removes all the Win9X semantics stuff
and a bunch of other associated cruft.

- moved the stat cache code into statcache.c

- fixed the uint16 alignment requirements of ascii_to_unistr() and
  unistr_to_ascii()

- trans2 SMB_FIND_FILE_BOTH_DIRECTORY_INFO returns the short name as
  unicode always (at least thats what NT4 does)

- fixed some errors in the in-memory tdb code. Still ugly, but doesn't
  crash as much

23 years agoYIPEE!!!!!
Andrew Tridgell [Sun, 30 Apr 2000 04:45:16 +0000 (04:45 +0000)]
YIPEE!!!!!

We finally have a perfect emulation of Microsoft wildcard
matching. The routine ms_fnmatch() does wildcard matching with all MS
wildcards (including the unicode wildcards), and masktest against a
NT4 workstation with hundreds of thousands of random exmaples has not
found a single error.

amazingly it is only about 60 lines of code, but it has taken us years
to get it right. I didn't sleep much last night :)

23 years agoput tdb utility functions in a separate file
Andrew Tridgell [Sat, 29 Apr 2000 04:05:38 +0000 (04:05 +0000)]
put tdb utility functions in a separate file

23 years agoRemoved tdb locking calls (picky Andrew...:-)
Jeremy Allison [Sat, 29 Apr 2000 01:09:31 +0000 (01:09 +0000)]
Removed tdb locking calls (picky Andrew...:-)

23 years agoMoved deletion of vfs handle into smbd/conn.c as it was being done too
Jeremy Allison [Fri, 28 Apr 2000 21:09:26 +0000 (21:09 +0000)]
Moved deletion of vfs handle into smbd/conn.c as it was being done too
soon in smbd/service.c (file operations were being done after the
handle was closed).

It looks cleaner in smbd/conn.c as it is part of the closing of a conn
struct anyway.

Jeremy.