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 attempts to fix the build on Solaris
Autobuild-User: Volker Lendecke <vlendec@samba.org>
Autobuild-Date: Sun Mar 6 10:45:16 CET 2011 on sn-devel-104
Abartlet, this commit makes check_sam_security_info3 use talloc_tos() and also
cleans up the temporary talloc stackframe.
The old code created a temporary talloc context off "mem_ctx" but failed to
clean up the tmp_ctx in all but one return paths.
talloc_stackframe()/talloc_tos() is designed as a defense against exactly this
error: Even if we failed to free the frame when returning from the routine, it
would be cleaned up very soon, in our main event loop.
Please check this patch!
Thanks,
Volker
Autobuild-User: Volker Lendecke <vlendec@samba.org>
Autobuild-Date: Sat Mar 5 14:08:37 CET 2011 on sn-devel-104
For a non-preallocated dest-string and sourcestring of len < 2,
(one or both of the) final two two zero-bytes would be written
after the end of the allocated dest-string. The sourcelen did
not include the source string terminator. For longer strings,
this was not a problem because the dest-string would have been
reallocated in the convert-loop. This is fixed now by allocating
two extra bytes for the terminating 0-bytes that are needed anyways
in the initial allocation.
Pair-Programmed-With: Gregor Beck <gbeck@sernet.de>
If we immediately afterwards perform an LDB base operation then we don't
need an explicit "ldb_dn_validate" check anymore (only OOM makes sense).
Reviewed by: Tridge
This is a much better solution than we had before - so all important DN
checks are enforced for each type of LDB database (and not limited to DSDB).
Many "ldb_dn_validate" checks will now become obsolete.
Reviewed by: Tridge
building smbtorture4 with configure --disable-shared failed
with an error that ldb.h could not be found
Signed-off-by: Matthias Dieter Wallnöfer <mdw@samba.org>
Reviewed by: Tridge
This should test the non proxy rodc kdc path.
metze
Autobuild-User: Stefan Metzmacher <metze@samba.org>
Autobuild-Date: Fri Mar 4 22:06:10 CET 2011 on sn-devel-104
By having kdc_tcp_proxy_send/recv(), which just asks any writeable
dc for a reponse blob, we simplify the interaction between
client-local and local-writeable sockets.
This allows us to make kdc_socket, kdc_process_fn_t, kdc_tcp_call and kdc_tcp_socket
private to kdc.c again.
metze
By having kdc_udp_proxy_send/recv(), which just asks any writeable
dc for a reponse blob, we simplify the interaction between
client-local and local-writeable sockets.
This allows us to make kdc_udp_call and kdc_udp_socket private to
kdc.c again.
metze
Bit shifting is non-trivial in C:-)
This
int32_t a = 0x12340000;
uint32_t b = (a >> 16);
results in 0x00001234, but this
int32_t a = 0xEDCB0000;
uint32_t b = (a >> 16);
results in 0xFFFFEDCB, while we expected 0x0000EDCB.
metze
The fix for bug 7976 added a vfs connect call to create_conn_struct.
This change adds two further disconnect calls that were missed by the
original change.
Signed-off-by: Andreas Schneider <asn@samba.org>
Autobuild-User: Andreas Schneider <asn@cryptomilk.org>
Autobuild-Date: Fri Mar 4 16:33:53 CET 2011 on sn-devel-104
cups_async_callback() is called to receive new printcap data from a
child process which requests the information from cupsd.
Newly received printcap information is stored in a temporary printcap
cache (tmp_pcap_cache). Once the child process closes the printcap IPC
file descriptor, the system printcap cache is replaced with the newly
populated tmp_pcap_cache, however this only occurs if tmp_pcap_cache is
non null (has at least one printer).
If the printcap cache is empty, which is the case when cups is not
exporting any printers, the printcap cache is not replaced resulting in
stale data.
https://bugzilla.samba.org/show_bug.cgi?id=7915
Signed-off-by: Andreas Schneider <asn@samba.org>
The cups local_pcap_copy global served as a temporary buffer during
asynchronous cups printcap cache updates, as well as indicating when
the printcap cache had not yet been filled and printcap cache update
should block.
As smbd printcap reads are now triggered by the parent smbd following
printcap cache update, the variable and blocking mechanism are no longer
needed.
Signed-off-by: Andreas Schneider <asn@samba.org>
Configuring force user/group settings for the print$ share currently has
unexpected results, this is explained by how the driver upload/add
process takes place. Consider the following example:
[print$]
path = /print-drv
write list = $DRIVER_UPLOAD_USER
force group = ntadmin
- the client connects to the [print$] share and uploads all driver
files to the /print-drv/W32X86 directory.
- This is permitted, as /print-drv/W32X86 is owned by group ntadmin, and
the "force group = ntadmin" takes effect for the [print$] session.
- Once all files are uploaded, the client connects to the [ipc$]
share and issues an AddPrinterDriverEx spoolss request.
- In handling this request move_driver_to_download_area() is called,
which attempts to create the directory /print-drv/W32X86/3
- The create directory fails, as it is done as the user connected to
the [ipc$] share which does not have permission to write to the driver
directory. The [print$] "force group = ntadmin" has no effect.
This is a regression from previous behaviour prior to the commit:
783ab04 Convert move_driver_to_download_area to use create_conn_struct.
https://bugzilla.samba.org/show_bug.cgi?id=7921
Signed-off-by: Andreas Schneider <asn@samba.org>
