Jeremy Allison [Mon, 20 Dec 2004 22:01:42 +0000 (22:01 +0000)]
r4293: Fix inspired by debug trace from Rob Foehl <rwf@loonybin.net> - catch sendfile
errors correctly and return the correct values we want the caller to return (-1
meaning none in correct cases).
Jeremy.
Günther Deschner [Mon, 20 Dec 2004 18:42:58 +0000 (18:42 +0000)]
r4289: Close LDAP-Connection before retrying to open a new connection in the
retry-loop.
This fixes a deadlock-situation when ldapsam is used with the ldapi
interface: getpeername won't fail while trying to detect dead
connections on unix domain sockets. When the ldapi-connection was closed
server-side (due to OpenLDAP's idletimeout) we *never* got a new LDAP
connection.
Günther Deschner [Mon, 20 Dec 2004 11:36:39 +0000 (11:36 +0000)]
r4286: Give back 8 byte lm_session_key in Netrsamlogon-reply.
The old #ifdef JRATEST-block was copying 16 bytes and thus overwriting
acct_flags with bizarre values, breaking a lot of things.
This patch is successfully running in a production environment for quite
some time now and is required to finally allow Exchange 5.5 to access
another Exchange Server when both are running on NT4 in a
samba-controlled domain. This also allows Exchange Replication to take
place, Exchange Administrator to access other Servers in the network,
etc. Fixes Bugzilla #1136.
Volker Lendecke [Fri, 17 Dec 2004 10:20:53 +0000 (10:20 +0000)]
r4254: Add an undocumented hack. I had to delete a wrong mapping (a user that had
ended up as a gid in winbindd_idmap.tdb) from winbindd_idmap.tdb. Stopping
winbind was not an option on that machine....
Volker Lendecke [Fri, 17 Dec 2004 08:51:23 +0000 (08:51 +0000)]
r4248: Implement smbstatus -n, don't lookup users and groups. On heavily loaded
winbind systems, looking up hundreds of users can turn out to be a bit too
expensive if you just want to find out which smbd handles a particular IP
address.
Volker Lendecke [Fri, 10 Dec 2004 10:30:59 +0000 (10:30 +0000)]
r4128: Cron jobs etc seem to do an initgroups for root quite frequently. So
log.winbindd is spammed with 'user root does not exist'. Increase debug level.
Jeremy Allison [Thu, 9 Dec 2004 21:42:00 +0000 (21:42 +0000)]
r4120: Never, ever, doubt valgrind :-). Fix order of evaluation bug that's been in the
bitmap code for ever. Remove silly extra space in paranoid malloc.
Jeremy.
Jeremy Allison [Tue, 7 Dec 2004 18:25:53 +0000 (18:25 +0000)]
r4088: Get medieval on our ass about malloc.... :-). Take control of all our allocation
functions so we can funnel through some well known functions. Should help greatly with
malloc checking.
HEAD patch to follow.
Jeremy.
Gerald Carter [Wed, 1 Dec 2004 20:11:31 +0000 (20:11 +0000)]
r4028: * prevent the background LPQ daemon from updating
the print queue cache just because multiple smbd
processes sent a message that it was out of date.
* add some extra debug messages while trying to track down
down jobs being left in the queue after printing has
completed.
Gerald Carter [Tue, 30 Nov 2004 15:52:46 +0000 (15:52 +0000)]
r4018: * move claim_connection from the netbios session request
reply code to the negprot reply code to cope with
client connections on port 445. Fixes the spurious
"register_message_flags: tdb fetch failed" errors.
* don't run the backgroup LPQ daemon when we are running
in interactive mode.
Jeremy Allison [Tue, 30 Nov 2004 06:41:38 +0000 (06:41 +0000)]
r4016: Fix for bug found by Steve French client code (cifsfs) on
POSIX ACL set. You need to *get* a permset_t pointer from the entry before
any of the permset code will accept it as a valid value
Jeremy.
Jeremy Allison [Tue, 30 Nov 2004 01:01:43 +0000 (01:01 +0000)]
r4009: Fix from Timur Bakeyev <timur@com.bat.ru> for bugid #2100,
change the way we check for errors after a dlopen (which
may set internal warnings which get picked up by mistake
in dlsym).
Jeremy
Jeremy Allison [Wed, 24 Nov 2004 22:05:59 +0000 (22:05 +0000)]
r3944: Fix the problem we get on Linux where sendfile fails, but we've already sent the
header using send(). As our implementation of sendfile can't return EINTR (it
restarts in that case) use an errno of EINTR to signal the linux sendfile fail
after header case. When that happens send the rest of the data and then turn
off sendfile. Sendfile should be safe to enable on all systems now (even though
it may not help in all performance cases).
Jeremy.
Günther Deschner [Tue, 23 Nov 2004 01:05:31 +0000 (01:05 +0000)]
r3918: Allow to set OWNER- and GROUP-entries while setting security descriptors
with smbcacls and using with the -S or -M switch. Fixes #404 and #2076.
Jeremy Allison [Sat, 20 Nov 2004 21:24:52 +0000 (21:24 +0000)]
r3895: Fix for bug #2045. May also fix other timestamp bugs with Excel
(Volker please test). Setting a last write timestamp from Windows
overrides any subsequent write timestamp changes and must be immediately
seen by and findfirst/findnexts. This is a racy solution, but should
work most of the time. This may also fix #1061, not sure.
Jeremy.
Volker Lendecke [Fri, 19 Nov 2004 12:11:13 +0000 (12:11 +0000)]
r3873: The semantics of the parameter 'printcap name' are a bit tricky. I had seen
the effect that I could not list printers with smbclient -L. I have cups
libraries but no running cups server, so remove_stale_printers() removed all
my printer definitions from the share list. So I said 'printing = bsd' but it
still would not work.
This happened because init_globals() would initialize Globals.szPrintcapname
to "cups", and the explicit 'printing = bsd' did not reset it. 'printing=bsd'
can't reset it, as this might overwrite an explicit setting. Thus I separated
the lp_printcapname into a function of its own, looking at
Globals.szPrintcapname and subsequently at sDefault.iPrinting.
Please revisit, there are just too many cases to cover.
projects / thirdparty / samba.git / log
