IF YOU WOULD LIKE TO GET AN ACCOUNT, please write an
email to Administrator. User accounts are meant only to access repo
and report issues and/or generate pull requests.
This is a purpose-specific Git hosting for
BaseALT
projects. Thank you for your understanding!
Только зарегистрированные пользователи имеют доступ к сервису!
Для получения аккаунта, обратитесь к администратору.
this allows us to experiment with ensuring the tdb hash
size for our open files and locking db are appropriately
sized. Make the hash size larger by default (10007 instead
of 1049) and make the locking db hash size the same as the
open file db hash size.
Jeremy.
(This used to be commit e7225f7e81)
to be set in local.h. Change from the default (131) to
another prime (1049). Should this be an smb.conf tunable parameter
based on the number of open file descriptors available ?
If so what scaling factor ? More tests to follow.
Jeremy.
(This used to be commit 6a902ec49f)
"allocation roundup size", by default set as 1Mb. From
advice by BlueArc about Windows client behaviour. VC++
people can set this to zero to turn it off.
Jeremy.
(This used to be commit 833ca10177)
idea, and aparently improved performance in some circumstances, but it
breaks the VC++ compiler :-(. Not cool. Fix bug #2146.
Jeremy.
(This used to be commit b9f147634d)
the unix domain sockets used by winbindd (also solves FD_SETSIZE problem
in winbindd to boot !). Adds a "last_access" field to winbindd connections,
and will close the oldest idle connection once the number of open connections
goes over WINBINDD_MAX_SIMULTANEOUS_CLIENTS (defined in local.h as 200
currently).
Jeremy.
(This used to be commit 7a586552a3)
mangling implementation, selectable using "mangling method = " in smb.conf
It also tidies the interface a little, although it is still nasty.
(This used to be commit be23d87a17)
This allows embedded systems to compile out the higher debug
levels. It should gain speed as well as reducing the code
size. Setting it to 1 saves about 300k of code on my system.
(This used to be commit f34cac3b31)
the pam password change code to ensure that existing and working password
chat scripts don't break with 2.2.1. PAM password changing has to be explicitly
requested.
Allowed wildcards in pam password change matching (matches password chat
script matching). Had to add const (sorry Tim :-) to ms_fnmatch() to
stop warnings. Don't worry - the const changes are isolated and don't
cause any other warnings :-).
Jeremy.
(This used to be commit 47b4d82536)
- removes SMB_ALIGNMENT. That macro caused all sorts of problems with
getting unicode aligned right in sub-protocols (such as SMBtrans and
SMBtrans2). I believe the performance reasons for having
SMB_ALIGNMENT has gone away with the new variants of the SMB
protocol anyway, as newer commands tend to have their own internal
alignment.
- fix the locations where we set smb_flg2 to absolute values. We must
never do this if we want a hope of coping with unicode.
- add initial support for unicode on the wire in smbd. Currently
enabled using SMBD_USE_UNICODE environment variable.
(This used to be commit b98b1435e9)
this means "nmblookup -S" now always works, even with broken servers
the database stores all unexpected replies and these can be accessed
by any client.
while doing this I cleaned up a couple of functions, and put in place
a better trn_id generator. in most places the code got quite a bit
simpler due to the addition of simple helper functions.
I haven't yet put the code in to take advantage of this for pdc
replies - that will be next. Jeremys pdc finding code will then work :)
(This used to be commit 280e6359d3)
include/smb.h:
param/loadparm.c: Made GUEST_SESSSETUP run time selectable. Horror of horrors :-).
printing/printing.c: Added J.F.'s latest fix.
rpc_parse/parse_misc.c:
parse_reg.c:
rpcclient/cmd_reg.c:
rpcclient/display.c: SGI compiler signed/unsigned issues.
smbd/reply.c: Made GUEST_SESSSETUP run time selectable. Horror of horrors :-).
utils/testparm.c: Added extra test.
Jeremy.
(This used to be commit 9668a5ef50)
on by default on OSes that support it, given the network IO pattern
that Samba uses. He doesn't think it will cause problems.
(This used to be commit 15016aa024)
MS client doesn't respond to a oplock break request then we might try
resending the request.
This morning Sues Win95 machine had this problem (thus forcing me to
do something about it!). When starting winword her box refused to
answer an oplock break request. I have added code to resend the oplock
break request up to 3 times at 10 second intervals before killing the
socket.
of course, as soon as I did this her box started to behave again so I
haven't been able to tell if this fix actually works, but it can't be
worse than dropping the socket immediately.
(This used to be commit a1c4d8351b)
Sets up the files array correctly - limited by the smb.conf parameter
and by the max fd's per process as found by getrlimit().
Jeremy.
(This used to be commit eca24bd243)
I've finally changed this because I've found a real effect that this
has. Win95 shows long filenames in a directory listing in a dos box
only if this is set to NTFS or HPFS (only ones I've found so far
anyway).
If NTFS poses problems of some kind then we should try HPFS
I wonder if this change will trigger any attempts to use ACLs ?
(This used to be commit 410d54fa0d)
SOFF_T() macro for setting an SMB_OFF_T variable
also limited mmap based reads to MAX_MMAP_SIZE. We really can't mmap
2^50 bytes due to virtual address space problems.
(This used to be commit 4e784b1889)
removed most cases where a pnum is used and substituted a pipes_struct*.
in files.c I added a offset of 0x1000 to all file handles on the
wire. This makes it much less likely that bad parsing will give us the
wrong field.
(This used to be commit 8bc2627ff2)
