Age | Commit message (Collapse) | Author | Files | Lines |
|
smbd/quotas.c:
Added quotas patch for autoconf from Dejan Ilic <svedja@lysator.liu.se>.
printing/printing.c: Filenames with spaces patch from Allan Bjorklund <allan@umich.edu>
utils/nmblookup.c: Fix usage() function.
smbd/reply.c: Split out the security=server and security=domain checks into
check_server_security() and check_domain_security() to aid the
writing of the 'hack' appliance mode invented by John Schimmel.
Jeremy.
(This used to be commit f09ab9b52251087a58af92ec753537ca34a970fc)
|
|
(This used to be commit 4e6b548b04b2ff982baf49ebd08e22180cd521b1)
|
|
nmbd/nmbd.c: Someone (a "yank" no doubt :-) Changed instances of "initialise" to
"initialize". Someone get that man an *English* dictionary.... :-) :-).
Jeremy.
(This used to be commit 6279be7f96802bb132e5e18dd8c6912652296e70)
|
|
(This used to be commit 5e2fc8b7ccf4a9c8b0a2acf4bd752531ac6fa775)
|
|
(This used to be commit ef9749798ab72a5e26f9cb5d0904266d6df70f8c)
|
|
(ie. it breaks Sues machine).
Luke, your comment was:
put unicode strings after SAMLOGON query regardless of whether it's
an NT mailslot or a non-NT mailslot, after having observed this behaviour
out of NT machines.
perhaps you could post the relevant tcpdump or netmon capture so we
can see what is going on? I suspect that what you saw isn't exactly
what the win95 boxes are generating. Maybe you saw a GETDCxxx instead
of a GETDC000 ? Maybe we need a switch based on the GETDC request
type?
We won't know unless we see sniffs.
(This used to be commit a2bacc08955ba61aac4b45b63a54b279e5776261)
|
|
(This used to be commit d398ae487a3edbfefb980fe0abddca1e95f6f3d5)
|
|
password change server as anonymous - you don't need to have
logged on as the user to do the password change. Doh ! (Why
didn't *I* think of that one.... :-).
Jeremy.
(This used to be commit dd3bb250a6fb6020d32ea35d7a3d3301d7f9764b)
|
|
(This used to be commit 7f3c96c5dae19b203bf79a88225768003d2f7516)
|
|
(This used to be commit 93b195d1e9490fb0e4310532fd15ab376f08aba5)
|
|
that caused NetBench to fail as the wrong error code was being
returned from a getatr call. Change was not to set the
unix_ERR_class global when you are not returning an error. This
is a big of Samba internal folklore that is *important*.
Jeremy.
(This used to be commit a1923d465be4c63a40d55abb800cc9b16a4b87d0)
|
|
1) inserts the pathname of the object file in the dependency file
(currently, gcc will only generate the basename)
2) the directory tree within .deps mirrors the source tree, so that we use
shorter filenames
problem 1 was reported to me by Tom Angert <angert@bellsouth.net>
(This used to be commit 274a2cff30f7325d1bc1647075a6c4f887e42e4f)
|
|
reflects the real options. Removed '-S' option - couldn't see what
this was doing (seemed to have something to do with nt_domain_logon ?).
Luke please check if you needed this code (code is only #ifdef'ed out
and can be re-instated).
Did this to make writing of yodl smbclient man page possible :-).
Jeremy.
(This used to be commit 97a14db75634b09dd272e9f59bc39c2b1eb52544)
|
|
test for both before including ACL support.
(This used to be commit 86743dfda446fddaffff7cbb57ca6b5901948ad5)
|
|
Again, to compile debug2html, you need to do a 'make debug2html'.
Chris -)-----
(This used to be commit bd0c5ee8e796bc624bf51fbca1c894ca6190b4ea)
|
|
know that 'make proto' will ignore them there. The debugparse.h header
file is included in includes.h, and includes.h is included in debugparse.c,
so all of the pieces "see" each other. I've compiled and tested this,
and it does seem to work. It's the same compromise model I used when
adding the ubiqx modules into the system, which is why I put it all into
the same directory.
Chris -)-----
(This used to be commit b6888faacdba035e1b608a404a71d93791de2d52)
|
|
directory in the next batch. Chris -)-----
(This used to be commit 81f64d89924523279a2dcbe53d5355f3dd56b5ff)
|
|
Chris -)-----
(This used to be commit a1aa808d5b85d150ec21d09e301aa0211658e021)
|
|
Jeremy.
(This used to be commit 2bce2cf6d83d9889d8b116f9d2ee254b7ceb4828)
|
|
annoying exception I'm still looking at :-).
Jeremy.
(This used to be commit 9cade656e2db743cb6ee67bbb8b0f50e8e4c39f9)
|
|
Chris - feel free to fix these things differently if these fixes don't
work for you.
Jeremy.
(This used to be commit 1f225abfbe6b021b589711c2e4c599a63050c5b2)
|
|
fix username overwriting problem in security=share.
Jeremy.
(This used to be commit 17621964344336c6303386c3c9e8a3e23b9e46e1)
|
|
(This used to be commit 99430ce899c02b95ab3c577afff7f28c6d5c6c45)
|
|
the reason we don't return() here is that smb_panic() is used not just
for segv but also for detected errors (such as buffer overflows)
(This used to be commit 7a50ca81a5842f7139e67c0ff9de79518599f445)
|
|
(This used to be commit 177b7b6b190c78ab59e19187d3221d23bb280911)
|
|
current status is "it works". I need to add some syntax error recovery
and a usage message. Basic stuff.
I've also modified Makefile.in. If you want to compile it you'll have to
do a 'make debug2html', as I used smbtorture as a model. We can decide
later if this tool is useful enough to be compiled always.
BTW, a 'make realclean' fails because the bin directory isn't empty.
That's because it doesn't delete optionally compiled files such as
smbtorture and debug2html (and because of the CVS subdirectory, but I
think that's only a problem for developers).
Chris -)-----
(This used to be commit e86fd87a3806f4c161a8f887bbdd4d400ae56ac2)
|
|
Chris -)-----
(This used to be commit 26fcdfd082e84df675ad06c7ba8c0cdbf07eed98)
|
|
(This used to be commit 22721b027139ffb7ff2625741b68701ba15fbb98)
|
|
now as "RPC fault" if the UNIHDR structure lengths do not exactly
match up to the length of the data stream.
so, all versions of samba prior to this one have an off-by-one bug
in unicode string lengths.
all versions of NT prior to NT 5 beta 2 could possibly have buffer
problems when receiving badly formatted UNICODE strings.
(This used to be commit 161eb6f511e161b63c1fa90a08c562fcf208344a)
|
|
(This used to be commit ae1888529b00f1a98d2b76fec7f19684efca90c7)
|
|
license at the top of lib/snprintf.c
I've always been slightly uneasy about our half-baked vslprintf()
implementation and the risks on platforms that don't have vsnprintf()
so when I saw this code in another GPLd package I wanted it for Samba.
(This used to be commit e2cb50af45be6683d02ab48a6648816ee3d52ab1)
|
|
* Debug message filtering.
* I need to add all this to Makefile.in
(If it looks at all strange I'll ask for help.)
If you want to compile debug2html, you'll need to do it by hand until I
make the changes to Makefile.in. Sorry.
Chris -)-----
(This used to be commit fae161b9051c3ea8aa129bdb3df11d215007723d)
|
|
Ensured offset was being reported correctly for 64 bit file access,
ensured delete on close bit was being reported correctly.
Jeremy.
(This used to be commit 4332bc06c6604a3e14007eda4f40e6183a1f8526)
|
|
(This used to be commit b2210614e810c8e84c9a14a8e32e05f95d92479b)
|
|
(This used to be commit b2a7f85d597d4d2a71fd38d76aac0464d53df626)
|
|
don't seek unless absolutely necessary on files in do_put()
(This used to be commit 9d33d405cdbd9ae61fb68b6f5ce5f56f86cc54bb)
|
|
(for example /usr/bin/man because it is setgid).
(This used to be commit 85f639ab2ba72dda5a00cf11c8ee705f5a51ec54)
|
|
changes uses the unique index number from a SMB_QUERY_FILE_ALL_INFO to
try to provide inode numbers. If it is 0 then use the hash of the
filename as before.
(This used to be commit 2565ccf9de9d5e80fdb5bcadbc7130faba386d95)
|
|
descriptors by catching close attempts on those fds and returning
EBADF.
(This used to be commit 9d863fb1681a5b03696552e1d93fe339b4bae455)
|
|
(This used to be commit 552175fdcb78c79833464648b5ccdc91b935e69c)
|
|
do_get() call.
(This used to be commit 98ba0b3a0a5074cf2b7adba407c5fd46f63c3f27)
|
|
long not 128 (fstring) length.
(This used to be commit 1e96e14e790e733bbd182a828b0db8727fead2a2)
|
|
don't have the ALLOW_DELETE_ON_CLOSE share flag enabled. Told you
this was fiddly code :-).
Jeremy.
(This used to be commit 222d686a6cdbc2947e7201636ae80ef4c650a65d)
|
|
this is 0x4000 01ff on NT4 SP4, and 0x0000 01ff on pre-NT4 SP4.
(This used to be commit 59cf9c00d91f9706d58c3d3000842eccb9d272fc)
|
|
with a call to nt_lm_owf_gen. if this still doesn't get the NT hash
generated correctly then there may instead be a bug in mod_smbpwd_entry().
(This used to be commit 1c8c644210870fef7ea3eddf411b01a98f991fcc)
|
|
open_file_shared().
Jeremy.
(This used to be commit 8de02620ea8f45f6a700617911947467e7be46e1)
|
|
Jeremy.
(This used to be commit d815e2d7a65c95ae71372b4fc8fe82f0202386dc)
|
|
one fsp with that of another. Not good :-).
Jeremy.
(This used to be commit 269bca2b3bbf5cae9fdf05d0405bfd80272da5d1)
|
|
semantics.
This was trickier than it looks :-). Check out the new DELETE_ON_CLOSE
flag in the share modes and the new code that iterates through all open
files on the same device and inode in files.c and trans2.c
Also changed the code that modifies share mode entries to take
generic function pointers rather than doing a specific thing so
this sort of change should be easier in the future.
Jeremy.
(This used to be commit 5e6a7cd99d29d1cf068fc517272559c1cf47ea3a)
|
|
to generate PIC code
(This used to be commit a66c56d70e7e663f3a19cbfcc2e704653d8fc9b2)
|